Professional Documents
Culture Documents
Contents
The paper by Nganji, Brayshaw and Tompsett, suggests a semantic approach to achieving reasonable
adjustments to the presentation of e-learning resources to meet the needs of disabled students, drawing on an
ontology of various disabilities encountered in higher education. The process aims to present disabled students
with tailored learning resources relevant and suitable for their individual and specific needs. The paper proposes
a viable assemblage of established architectural components which overall could deliver such a customised
Learning Environment.
The paper by Crane, Benachour and Coulton identifies and discusses infrastructural and sociological barriers
limiting access to learning environments via mobile devices. A survey of network coverage data, mobile widgets
and students’ experiences suggests that a lack of a dedicated mobile application or mobile website coupled with
inept network access impose significant constraints at this time.
The paper by Thomas, Waugh and Smith describes an approach to the generalisation of tools for teaching and
learning the skills associated with modelling with diagrams. The paper briefly describes the existing tools and one
approach to the automatic marking of diagrams. The authors report on their work to generalise both a marking
algorithm and a drawing editor in such a way that revision tools can easily be generated for new domains. They
also report on how they have incorporated their tools into their own institution’s Moodle-based Virtual Learning
Environment.
The paper by Whitworth, McIndoe and Whitworth details an open educational resource designed to develop
media and information literacy skills in postgraduate students. The authors describe the motivations for
undertaking the project, the model of media and information literacy education which they used to create the
materials, how the resource was created, and its evaluation using eight students. The outcome is the provision of
a stand-alone resource that can be completed in 7 to 10 hours of independent study by postgraduate students.
The resource is to be used by Manchester University in early 2011 but the authors are keen for other institutions
to adapt and use the resource and to receive feedback. Of particular interest is the use of the resource across
different disciplines. Links to the live resource are provided in the paper.
Laramee offers a template and guidelines underpinning the mechanics of writing a final year student project
report in Computer Science. These may serve as a useful starting point for students and new project supervisors
alike and are offered openly for adoption and use.
Kerins provides a reflective analysis of the utilisation of Higher Education Innovation funding to establish a
software development team within a department of computer science and information systems to grow industrial
links, and to effect synergy between students, academic staff and external clients in problem solving and in
forging solutions to real-world problems. The author outlines numerous benefits derived from the initiative
including the honing of the employability skills of participating students.
Solly and Matthews presents an overview of a real-world interaction design assignment where the focus was on
encouraging students to explore the conceptual design space to find creative solutions. By providing students
with background information extracted from case studies and with access to a domain expert, the students
developed a deeper understanding of the context. This assignment allowed for the integration of research and
teaching in a way that created an exciting environment where both students and teachers became co-learners.
The paper also discusses the benefits to the students of working with experts drawn from a range of non-cognate
disciplines.
The paper by Coleman and Nichols presents a case study on significant improvements in student attendance and
in assessment mark profiles obtained by pairing students within a first year algorithmic programming module.
Students worked together in pairs and were also assessed in pairs. The authors also captured the students’
views on working and being assessed as a paired unit and noted how the collaborative relationship and peer
support continued beyond the module of study and contributed positively to the overall student experience.
The paper by Coull and Duncan reflect upon the problems associated with learning and teaching first year
University Computer Science programming classes and review the various support tools and techniques which
have been developed to improve student success in this area. The authors propose ten requirements that a
support tool should satisfy and offer these to those interested in selecting among the support tools available or
moving to build a dedicated system.
Jayal, Lauria, Tucker and Swift presents an experimental study wherein the understanding and proficiency of
students studying introductory programming via Java is contrasted with an approach which taught basic
programming concepts and constructs via Python before moving to Java. The study reported a marked
improvement in the performance of those students initially exposed to Python over those using Java throughout
the module. The results as set out are an interesting contribution to the ongoing introductory programming
language debate.
The paper by Salt, Lallie and Lawson presents the experiences of students studying Computer Forensics at the
University of Derby, together with the views of the teaching and technical support staff. The authors strongly
suggests that there is a growing mismatch between students expectations and reality, an underdeveloped ability
to problem solve and communicate findings and unrealistic expectations of the job market. The authors suggest
that Universities need to appraise their activities honestly and within an ethical framework. In introducing digital
forensics courses they suggest that there is a need to rapidly assess the needs of our students and to ensure that
our practices and their expectations align with the reality of the employment market.
ONTOLOGY-BASED E-LEARNING PERSONALISATION FOR
DISABLED STUDENTS IN HIGHER EDUCATION
ABSTRACT
The number of students with disabilities in UK higher education institutions increases every year. Delivering
education online is becoming increasingly challenging as institutions encounter some disabilities requiring
adjustments of learning environments. The law requires that people with disabilities be given equivalent
learning experiences to their non-disabled peers through “reasonable adjustments”. Educational institutions
have thus utilised assistive technologies to assist disabled students in their learning, but some of these
technologies are incompatible with some learning environments, hence excluding some disabled students and
resulting in a disability divide. To solve this problem, amongst other solutions, e-learning personalisation has
been used and more recently, this is also achieved using Semantic Web technologies such as ontologies.
Nevertheless, as ontologies are incorporated into learning environments little seems to be done to personalise
learning for some disabled students. This study, in order to bridge the gap, proposes a personalisation
approach based on a disability ontology containing information on various disabilities encountered in higher
education, which can be used to present disabled students with learning resources relevant and suitable for
their specific needs.
Keywords
E-learning, learning environments, Semantic Web, ontology, disability, disability divide, disabled students,
accessibility, personalisation, higher education.
1. INTRODUCTION
Today’s fast advancing information and communication technology (ICT) has changed the way in which we
live and work. Many activities in our daily lives are steadily being driven by such technology, including
distance and online learning. This is a global phenomenon as ICTs pervade many societies worldwide, with
poor and middle income countries also adopting and diffusing these technologies much faster than in high-
income countries (World Bank, 2008). No one wants to be left out of this information revolution. Coupled with
the Web, computers greatly enable learning and teaching and increase output. The benefits are potentially
enormous and hence, many educational institutions are employing these technologies which have greatly
driven the learning and teaching process. Teaching in higher education has thus progressed from the
traditional classroom method (e.g. chalk and talk) to embracing e-learning which involves the use of modern
technology in teaching and learning as the Web is increasingly being used by educational institutions
(Alzahrahni et al, 2008; Gonzalez, 2007).
Although these technologies have facilitated many tasks especially for people without disabilities, they have
also created an unequal access to such services for disabled people. Steyaert (2005) observes that this
digital divide for students with functional impairments is because:
“Web-based higher education is failing to transpose the basic accessibility notions from the physical to the
digital environment, hence, the threat of increasing exclusion”.
The fact that disabled students may be disadvantaged in education due to their disability could account for
poor performance, which necessitates “reasonable adjustments”. However, according to a 2009 research
report by the UK Department for Innovation, Universities and Skills (DIUS, 2009):
“Disabled students are slightly less likely to attain a good degree (first or upper second class) than those who
do not report a disability, and there is evidence to suggest that this persists even after controlling for a range
of other factors.”
“The Semantic Web is not a separate Web but an extension of the current one, in which information is given
well-defined meaning, better enabling computers and people to work in cooperation”
The Semantic Web will therefore have more meaning, enhancing the current Web, and will also enhance
study for disabled students given the ability to facilitate information retrieval as earlier discussed. It is the fact
that “online content of websites, databases and other resources is accompanied by machine-readable
descriptions that add meaning to the content, and describe the structure and status of the knowledge of that
content” (Carmichael, 2008). Anderson and Whitelock (2004) earlier observed that:
“The capacity of the Semantic Web to add meaning to information, stored such that it can be searched and
processed, provides greatly expanded opportunities for education, simulation and real-time action anywhere
on the distributed network”.
Semantic Web technologies could revolutionize the way people with disabilities access learning
environments, by increasing accessibility and relevance of presented content and hence e-learning
personalisation greatly benefits from this technology. Intelligent agents on the Semantic Web could facilitate
reading for some dyslexic students or could transform information and present it in a format that could be
more suitable for their needs. The Semantic Web has a layered structure as shown in Table 1.
Semantic Web technologies add more meaning to the Web especially with the use of ontologies (Horrocks,
2008). The word ontology has its origin from Philosophy where it refers to the subject of being or existence.
In Computer Science however, Gruber (2009) defines ontology as “a specification of a conceptualisation”.
Also noted Ibid is the fact that “ontologies are used for integrating heterogeneous databases, enabling
interoperability among disparate systems, and specifying interfaces to independent, knowledge-based
services”. Ontologies also play a vital role in personalising web content for both disabled and non-disabled
people. The Web Ontology Language (OWL) which superseded DAML+OIL is widely used today and is a
language for the Semantic Web developed by the W3C Web Ontology Working Group (McGuinness &
Harmelen, 2004). OWL is a revision of the DAML+OIL ontology language and has three sub languages: OWL
Lite, OWL DL and OWL Full. OWL Lite supports users needing a classification hierarchy and simple
constraints. OWL DL supports those users who want the maximum expressiveness while retaining
computational completeness and decidability. OWL Full is meant for users who want maximum
expressiveness and the syntactic freedom of RDF with no computational guarantees.
Table 1. Layers of the semantic web
Logic and Proof: This is an automatic reasoning system on top of the ontology layer to make new
inferences.
Ontology vocabulary: The ontology layer provides description for properties and relation between these
properties and resources.
RDF + rdfschema: The Resource Description Framework (RDF) layer enables the representation of
metadata about web resources. It uses URIs to identify web-based resources but also has a graph model
for describing relationships between resources.
XML + NS + xmlschema: The XML layer with its namespace and schema ensures the use of a common
syntax on the Semantic Web. This layer allows structuring of data on the web but does not communicate
the meaning of the data.
Unicode and URI: This is responsible for encoding characters and uniquely identifying resources on the
internet. Unicode is the standard for computer character representation while URIs provide the standard for
identifying and locating resources.
Figure 1: ADOOLES ontology class hierarchy
Ontology editors like Protégé, developed at Stanford University, have been designed to support ontology
modelling via the protégé-frames or protégé-OWL editors. This Java-based platform supports several plugins
amongst which is OWL Viz which makes use of GraphViz for graphically displaying ontologies as seen in
Figure 2. We used Protégé to develop an ontology which could serve as a knowledge base for personalising
4. E-LEARNING PERSONALISATION
With the exponential growth of the World Wide Web (WWW), it is increasingly becoming difficult to access
information online. One approach to solve this problem is through web personalisation, which makes it a lot
easier for individuals to access relevant information. Over the years, personalisation of learning environments
Figure 4: The application architecture used by Goudos et al (2006)
Nevertheless, in our system whose architecture is presented in Figure 5, content specific to education is used.
Additionally, we have made provision for various components that are vital to the functioning of the system as
follows:
Information translation and presentation component: this component enables the disabled student to
obtain content suitable to their specific needs. When the student interacts with the learning environment
through the user interface, they are initially accessing general content. The system identifies the student as
one with special needs and retrieves learning resources specific and suitable for their disability. This is
enhanced by the translation module which translates and presents the resources to the student in an
appropriate format. The translator for instance could collect text and translate them to speech for blind
students or can semantically annotate images, facilitating understanding and description of the visual
resource.
Knowledge representation component: this component represents domain knowledge in a format that
could be easily understood by the system. Also, the reasoner can easily query the knowledge base and obtain
results which could be used to personalise the learning environment. The component also provides a means
for assistive technologies to interact with the learning environment, facilitating compatibility.
Information retrieval component: this component is responsible for interacting with the knowledge
representation component, collecting information and transmitting to the information translation and
presentation component. This is mainly made up of the reasoner which queries the knowledge base,
obtaining some results and the web server which invokes the reasoner, obtaining web pages with the results
and returning them to the student.
Management component: this component has a visual interface which enables administrators such as
lecturers, the VLE manager and systems developer to manage the student’s learning and to solve some
problems they may encounter.
Alternative formats of learning materials: with this being an online system, learning resources can be
provided to students in various formats (audio, video, slides, text, etc.), which could be suitable to their needs
as well as their preferred learning styles.
Accommodation of different learning styles: the system allows for materials to be presented in various
formats which therefore accommodates different learning styles. Auditory and visual learners could benefit
from audio, video, as well as textual materials. Additionally, the system could translate information into
appropriate formats to suit the learner. For instance, a learner may use the system to convert textual
information into an audio format which might be more suitable to their needs.
Accessible learning resources: given that the learning resources are accompanied by machine-readable
descriptions implies they will be more meaningful and also adhering to accessibility and usability standards
increases the accessibility of the resources. They could also be easily retrieved through search. This system
will also enable students to access learning materials at anytime and anywhere, which would be convenient
for them to study at their pace.
This personalised system also makes it easier for students to learn, cutting down on potential confusion (Kim,
2007).
6. REFERENCES
Alonso, F., Lopez, G., Manrique, D. & Vines, J. M. (2005). An instructional model for web-based e-learning
education with a blended learning process approach. British Journal of Educational Technology, Vol. 36,
No. 2, pp. 217-235.
Alzahrani, S., Ayesh, A., & Zedan, H. (2008). Regionally Distributed Architecture for Dynamic e-Learning
Environment (RDADeLE). Proceedings of the 2008 Conference on Human Systems Interactions. Krakow,
Poland, pp. 579-584.
Anderson, T. and Whitelock, D. (2004). The Educational Semantic Web: Visioning and Practicing the Future
Education. Journal of Interactive Media in Education, 2004(1).
Berners-Lee, T., Hendler, J., & Lassila, O. (2001). The Semantic Web. Scientific American, Vol. 284, No.5, pp.
34-43.
Castells, P., Fernández, M and Vallet, D. (2007). An Adaptation of the Vector-Space Model for Ontology-
Based Information Retrieval. IEEE Transactions on Knowledge and Data Engineering, Vol. 19, No.. 2,
February 2007, pp. 261-272
Carmichael, P. (2008). The semantic web and ‘web 3.0’. In: Selwyn, N. (ed.) Education 2.0? Designing the
web for teaching and learning. [Online] Teaching and Learning Research Programme, Available:
http://www.tlrp.org/pub/documents/TELcomm.pdf [Accessed: 7 January 2011].
DCLG. (2008). Delivering Digital Inclusion: An Action Plan for Consultation. Department for Communities and
Local Government (UK). [Online] Available:
http://www.communities.gov.uk/documents/communities/pdf/1001077.pdf, [Accessed: 10 March 2009].
Dellschaft, K. and Staab, S. (2008). Strategies for the Evaluation of Ontology Learning. In Buitelaar, P. and
Cimiano, P., editors, Ontology Learning and Population: Bridging the Gap Between Text and Knowledge,
pages 253{272. IOS Press.
Dolog, P., Henze, N., Nejdl, W. & Sintek, M. (2004). Personalisation in distributed e-learning environments.
Proceedings of the 13th International World Wide Web conference on Alternate track papers & posters.
New York, USA, pp. 170-179.
DIUS, (2009). Disabled Students and Higher Education. Department for Innovation, Universities and Skills.
[Online] Available: http://www.dius.gov.uk/~/media/publications/D/DIUS_RR_09_06, [Accessed: 5 October
2009].
Gao, M., Liu, K. and Wu, Z. (2010). Personalisation in web computing and informatics: Theories, techniques,
applications, and future research. Inf Syst Front, 12:607-629.
Ghaleb, F., Daoud, S., Hasna, A., ALJa’am, J.M., El-Seoud, S.A. and El-Sofany, H. (2006). E-Learning Model
Based on Semantic Web Technology. International Journal of Computing & Information Sciences, Vol. 4,
No. 2, pp. 63-71.
Ghauth, K.I. and Abdullah, N.A. (2010). Learning materials recommendation using good learners’ ratings and
content-based filtering. Education Tech Research Dev, 58:711–727
ABSTRACT
This paper describes and discusses two issues which limit the delivery of mobile learning from assisting in
disseminating course and module information to higher education students. The concept of delivering
information to support learning is designed to augment their engagement with their subject areas and
ultimately enhance their learning experience by allowing for increased flexibility in their access to learning
materials. The paper concentrates on both the infrastructural and sociological issues associated with providing
VLE access from a mobile or flexible position. Using mobile widgets, network coverage data and student’s
responses to understand the potential benefits and limitations of using mobile devices to access information
from the VLE. Discovering the majority of students are without a dedicated application or mobile website,
coupled with inept network access this paper investigates the apparent constraints of a promising method of
disseminating information to learners.
Keywords
Mobile Learning, Virtual Learning Environments, Student Engagement, Flexible Learning
1. INTRODUCTION
Mobile learning or ‘M-Learning’ holds the promise to enhance and optimize learning by utilizing pre-existing
information systems and deliver them in a contextual, more meaningful manner. Mobile devices have become
increasingly popular and a number of applications have been developed to aid teaching and learning
strategies (Cheng et al, 2008).The modern or nomadic learner travels interminably between locations and
utilizes mobile technology to access information and learning materials from anywhere and at any time. This
revolution of pre-existing concepts of space, time and educational content into personalized, interminably
connected, and dynamic learning environments is currently in rapid development. Yet, it is the transmission
and delivery of these contemporary learning environments which are imperative to the successful uptake and
adoption of mobile learning. Subsequently, educators and technologists must design delivery systems which
are both effective and appropriate for the students.
With mobile phone penetration in Western Europe currently measured at 130% (Ofcom, 2008) the potential of
mobile devices lies in not only the popularity of the technology, but in the unique properties possessed by the
mobile phone or device. Although originally expressed as an outgrowth and consequence of e-learning, it was
soon defined as being ‘spontaneous, private, portable, situated, informal, bite-sized, light-weight and context
aware’ (Traxler, 2010).Possessing such inherent properties means that education can be delivered and
realized for digital native and nomadic students.
Unfortunately, mobile learning does have its limitations. Deliberations over these shortcomings (Shudong &
Higgins, 2005) cite a number of inherent weaknesses of why mobile learning disappoints on its massive
potential. Most notably, access to internet connections and personal boundaries are mentioned as potential
problems in the rapid adoption of the technology. With the delivery of learning objects being as imperative as
the content itself, discovering the next major development in technology enhanced learning is essential to the
Although it has been more difficult to create a fully comprehensive definition of mobile learning; several
attempts have already been proposed. Originally defined as its differences from traditional and ‘tethered’ e-
learning (Traxler, 2005) the ever developing technology with high mobile device penetration within society has
made the definition increasingly difficult. Although O’Malley (2003) encompasses mobile learning as “Any sort
of learning that happens when the learner is not at a fixed, predetermined location, or learning that happens
when the learner takes advantage of learning opportunities offered by mobile technologies”. As a result,
mobile learning can be considered a part of flexible learning. Perhaps mobile learning can offer the context
dimension to flexible learning. Further, mobile learning can be considered an application example of a
disruptive technology where users are interrupted regularly, where a user will receive information on the move
and within a context. This suggests that mobile learning has more control over the learner, whereas for flexible
learning, the user (learner) has more control in choosing study modes and methods.
The figure below demonstrates the characteristics or features for both flexible learning (Gunn, 2000) and
mobile learning (winters, 2006). As demonstrated the obvious parallel is of the allowance for personalization in
both time and special dimensions for accessing and adhering to learners varying learning schedules and
styles. For mobile learning to become a legitimate solution to the requirements posed by flexible learning, the
limitations which could hamper the delivery of information must be investigated and understood.
In addition, such a platform encourages higher education students to think about their future careers, and
supports personal development. Even with a blogging feature directly available for the students, many still
utilize social networks and blogging sites away from the university itself. Social software is perceived
differently within an educational context, compared with ordinary personal usage and this discourages
student’s adoption (Small & Morgan, 2008). The use of blogging in an educational context or as part of an
application built within an educational establishment is something that students have not really adopted. Even
the use of RSS or Twitter as technologies where students can get updates or share educational experiences
as part of or built-in within virtual learning environments is not something that students prefer to be engaged
with. This issue is one of the major motivations for this study and raises subsequent hypotheses, in which the
need for the pervasive nature of a ubiquitous environment is necessary in education. Traxler (2010) argues
that students no longer need to connect and consume information at the expense of real life, but can do so as
part of real life as they freely move around the world, using their own personal devices to engage with people,
ideas, and information, perhaps using these same mobile devices to produce, consume, and store content
and conversation. This can be achieved in a mobile network environment or within a wireless campus
network. A wireless network can be based on different technologies such as the work presented by Wang, Ci
Zhan, and Xu (2007) Here, a mobile learning environment based on applying wireless sensor networks to
context-aware learning enabling people to learn at any time and at any place was proposed and described .
What is important is that a cost- effective wireless infrastructure is available to support a learning environment
especially within a university campus. This is one of the main promises for mobile and wireless technology to
support and assist learners, allowing students to efficiently engage with information; and subsequently the
4. INFRASTRUCTURAL CONSTRAINTS
The communication and data transfer possibilities created by mobile technologies can significantly reduce
dependence on fixed locations for work and study, and thus have the potential to revolutionize the way we
work and learn (Peters, 2007). Of course; this entirely depends upon the accessibility and range of wireless
access available to both the staff and students. Higher education institutions have identified the student’s
requirement for mobility by providing wireless LAN access on campus, using such initiatives as ‘Eduroam’
which provides a free, secure, and internationally available roaming access service for the education
community. Unfortunately, with the developments in mobilizing access to education, nomadic learners have
now been transformed into mobile learners, and’ Eduroam’ is effective, but does not entirely encompass the
needs of mobile or flexible learners.
.
Currently, wireless access is configured for nomadic learning practises, with ‘hotspots’ being favoured within
busy, communal buildings or lecture halls. Figure 3 represents the data collected from a focus group of
students regarding their routes around the campus and the network access which is readily available for all
members of the university.
From the three separate maps of campus above, it can be seen that the wireless LAN access is limited to
buildings. This also conveys a lack of wireless access around the open areas which are used for communal
use by the students, as well as the housing area (accommodation) and therefore a large area which students
use as part of their daily routes on campus.
Image ‘a’ is the university’s own publication which denotes the availability of wireless access across the
campus. These are mostly buildings in which most educational and research activities take place, such as
lecture theatres, laboratories or seminar rooms. Image (b) is a representative simulation which provides a
The most obvious alternative for access is using the mobile internet itself, although cost can be especially
large for the considerable proportion of overseas students. With an average rate of £1 a megabyte for both
uploading and the downloading data, strategies must be implemented to allow cost-effective and continual
retrieving methods for a fully pervasive learning framework. Of course cost efficiency is an issue, but so is
quality of access. Table 1 demonstrates the quality of access from five separate university campuses
according to each of the major mobile phone service providers in the UK for accessing email, browsing
internet pages and viewing video content. The mean values represent the calculated average value of
coverage provided by each of the mobile network operator’s freely available coverage data.
Campus 1 0.84
Campus 2 0.92
Campus 3 0.84
Campus 4 0.92
Campus 5 0.92
Campus 6 0.96
Campus 7 0.88
Campus 8 0.92
Campus 9 0.88
Campus 10 1
Campus
0.908
Mean
____________________________________________
Table 1 – The average 3G network coverage from ten UK campus based higher education institutions
Although each of the network providers possessed a relatively high level of coverage for each of the ten
university campuses, unfortunately only one campus environment had a full level of coverage for 3G access.
One apparent answer to the problem would lie in higher education institutions purchasing network capacity
from a mobile network operator to provide custom and higher quality mobile services to end-users. Again, the
downside would be significant cost, both in the setup and maintenance of an information distribution channel
which subsists without any expense. But this must be considered and costs for such a system should be
calculated.
Apart from the obvious infrastructure of communications access, the other issue is that of the access of the
content for the virtual learning environments. The major challenge to the utilization of such mobile learning
channels as RSS feeds for mobile learning is that of the paradox which arrives when using feeds directly from
a secure virtual learning environment. With RSS feeds being inherently mechanisms intended to freely inform
the user of recent content modifications, and authorization being the procedure to restrict access to
information (Crane & Benachour, 2010). This pre-existing information infrastructure poses an obstruction in
the access to mobile learning due to security policies of each institution. Although the wireless access is a
major factor in constraining mobile learning, the virtual learning environments structure is also important in
5. SOCIOLOGICAL BOUNDARIES
Although current network access can limit the usage of a mobile virtual learning environment; if the students
feel it is an infringement upon their personal space, the concept of a mobile VLE for supporting learning can
be somewhat invalidated due to not understanding personal contexts of time and space.
For many students, a mobile phone is the ultimate personal device, serving a multitude of functions in their
day to day lives and holding important information. Being the definitive personal device means there is a
strong sense of personal space and limitations to the extent academic practises should intrude. Certain
sociologists even argue that with the evolution of a mobile connected society, “the distinction between public
and private domains should be dispensed with since nothing much of contemporary social life remains on one
side or the other of the divide” (Sheller & Urry, 2003).
In order to attain more information surrounding the sociological aspects which affect the necessity for mobile
learning; a survey was conducted from a collective of 66 students, from a total of twenty different higher
education institutions; with the entirety answering positively to both using a university maintained virtual
learning environment and owning a mobile phone.
From this investigation, the majority of students do not have a dedicated mobile VLE either by a dedicated
application or by a mobile website. Further to this, most of those who answered believed they would use a
mobile VLE; yet the appropriateness of has a larger, more comprehensive and apprehensive range of
responses.
One other boundary which has arisen during the last few years is the use of social networks and Web 2.0 by
higher education institutions to informally improve student engagement away from the VLE. Figure 5 shows
the responses to the integration of VLE and social networks in the delivery of learning content and information.
Interestingly, there is a lack of positivity in the incorporation of external social networks, with the majority of
those who answered finding it less comfortable being connected in this domain. Although a student having
their own phones being transformed into a learning form seems acceptable; the concept of social networks
doing the same seems relatively unacceptable.
7. STUDENT EXPERIENCES
In order to measure the student’s experiences of the mobile learning trials, the participating users were asked
to attend a short, informal debriefing to induce some insights into the mobilized virtual learning environment
content. The students were provided with a blank SWOT analysis and asked to fill in the strengths,
weaknesses, opportunities and threats to supporting mobile learning using the VLE. Figure 7 is the
aggregated results of both the RSS and Twitter group’s thoughts towards the strengths, weaknesses,
opportunities and threats of the system they had experienced.
The students stated a range of issues, benefits and limitations of the mobile supporting technology. Most
notably, the wireless and network connectivity, teamed with the economic costs which they cited as a threat to
their adoption of the mobile delivery of VLE information. None of the students questioned touched on, nor
discussed the intrusion or disturbance of their personal boundaries. Alternatively, the students mentioned the
aggregation of social network into their mobile delivery channel as an opportunity for the technology. As seen,
there were some differences in the responses between both the official RSS channel and that of the twitter,
social networking channel, which include all four sectors of the SWOT analysis. Ultimately, the problem of cost
efficient wireless access was the most apparent consistent problem.
Investigating the limiting factors in mobile learning will be of interest to both educational technologists and
educators themselves. From the widgets constructed and the subsequent investigation into information
delivery in a mobile learning domain; it can be easily argued that a mobile VLE could be efficiently utilized.
Overall all the students responded positively about the mobile widgets position and the appropriation of their
use during the trials. Although there is no clear leader, students generally responded more positively to the
RSS widget, rather than the Twitter application. As far as limitations and drawbacks of the mobile study, the
data sets recruited and tested were from one department and may have had existing knowledge or
background of a technical nature. Conceivably, using another set of students from an arts or humanities
subject field may provide differing results on the strengths and weaknesses of a mobile VLE information
system. Again, the focus groups used were not interviewed or chosen on any existing or previous mobile
learning background, and may have been exposed to similar systems prior to the study itself. Once more, all
the students were provided with identical mobile phones, although some may have been familiar with the
handset or manufacturers interaction menu system previous to the study. Several directions and research
areas can be considered and investigated from the results of this paper. One major issue for further review is
the coverage and infrastructure of the universities wireless network. The need to create cost effective mobile
information systems directly for students could be considered inane without the necessary internet access
available. Furthermore, the constrains which appear to exist in utilizing externally available Web 2.0 should be
discouraged or discarded due to the student’s own personal space. Ultimately, this paper is only a foundation
to the discussion and discovery of the apparent limitations of mobile learning channelling VLE information.
Identifying and answering the issues raised are imperative for both student engagement and future education
and technological advancements.
9. ACKNOWLEDGMENT
The authors wish to acknowledge the support of Nokia for the real-time hardware which was used for the
development of this project. Again, appreciation must be directed towards Lancaster University’s Information
System Services and Learning Technology Group for their on-going support.
10. REFERENCES
Cheng, G. D., Chang, C. K., & Wang, C. Y. (2008). Ubiquitous learning website: scaffold learners by mobile
devices with information aware techniques. Computers & Education,50 (1), 77-90.
Choy, Sarojni C. and McNickle, Cathy and Clayton, Berwyn (2002) Learner expectations and experiences.
Student views of support in online learning. In: Guthrie, H. (Ed) Online learning. Research readings. National
Centre for Vocational Education Research, Adelaide, pp. 106-122.
Collis, B., and Moonen, J. (2001). Flexible learning in a digital world. London: Kogan Page.
Crane, L. & Benachour, P. (2010). Dissemination of learning services: using RSS for M-Learning. IADIS
Mobile Learning Conference. Porto, Portgual, (pp. 269-273).
Gunn, C. (2000). A guide to flexible learning principles and practices. Auckland, New Zealand: The University
of Auckland, Centre for Professional Development.
Jones, N., Blackey, H., Fitzgibbon, K, & Chew, E. (2010).Get Out of MySpace!. Computers & Education, 54
(3), 776-782.
Pettenati, M, Abou Khaled, O, & Giuli, D. (2000). Technology and pedagogy; developing a webagogy.
Submitted to Journal of Interactive Learning Research.
Traxler, J. (2010). Will student devices deliver innocation, inclusion, and transformation?. Journal of the
Research Center for Educational Technology, 31 (2) 129-138.
Traxler, J. (2005). Defining Mobile Learning. Proceedings IADIS International Conference Mobile Learning
2005, Malta, pp 261-266.
Sheller, M. and J. Urry (2003) ‘Mobile Transformations of “Public” and “Private” Life’, Theory, Culture &
Society 20(3): 107–25.
Shudong W., & Higgins, M. (2005). Limitations of mobile learning. The JALT CALL Journal. (2), 3-14.
Pachler, N., Bachmair, B. & Cook, J. (2010). Mobile learning: structures, agency, practices. New York:
Springer.
M.C. Pettenati, O. Abou Khaled, D. Giuli Technology and Pedagogy; developing a Webagogy, submitted to
Journal of Interactive Learning Research, AACE, March 2000.
Peters, K. (2007). M-Learning: Positioning educators for a mobile, connected future. International Review of
Research in Open and Distance Learning, 8(2).
Traxler, J. (2010). Will student devices deliver innocation, inclusion, and transformation?. Journal of the
Research Center for Educational Technology, 31 (2) 129-138.
Wang, M., Ci, Li., Zhan, P., & Xu, Y. (2007). Applying wireless sensor networks to context-awareness in
ubiquitous learning. Third International Conference on Natural Computation.
Winters, N. (2006) 'What is mobile learning?', In M. Sharples (ed.) Big Issues in Mobile Learning: Report of a
workshop by the Kaleidoscope Network of Excellence Mobile Learning Initiative Nottingham: University of
Nottingham.
ABSTRACT
We describe an approach to the generalisation of tools for teaching and learning the skills associated with
modelling with diagrams. The paper briefly describes the existing revision tools and our approach to automatic
marking of diagrams. We report on our work to generalise both the marking algorithm and the drawing editor
in such a way that revision tools can be easily generated for new domains. We also report how we have
incorporated our tools into our institution’s Moodle-based Virtual Learning Environment (VLE).
Keywords
automatic marking, computer-assisted instruction, computer science education, diagrams.
1. INTRODUCTION
Diagrams are used in many disciplines for communication, recording decisions and as a mechanism for
exploring alternatives. In some fields, formal syntax and semantics for diagramming have been defined. For
example, the Unified Modeling Language, UML, (Jacobson et al, 1998) in Computing is a well-defined notation
for recording design decisions in software development. The UML defines a range of diagram types for
different purposes. Similarly, there are also agreed diagramming standards in Electronics for circuit diagrams
and in Chemistry for molecular diagrams.
Diagrams are ubiquitous in teaching and learning and there are often associated modelling skills that students
should master. Often, students use off-the-shelf diagramming software that would typically be used by
practitioners. While they need such skills in practice, such tools can be daunting to learn. Additionally, they
often constrain the student to (a subset of) the syntax and semantics of the standard language and are not
specifically designed for teaching and learning. We would prefer to allow students to construct diagrams with
little constraint on what they draw and have a tool provide feedback on errors made in the syntax.
Integrating such tools into e-learning and e-assessment systems is problematic. E-assessment systems are
almost exclusively text-based with little support for diagramming. Quiz engines, for example, expect
responses to be textual, particularly if the response is to be automatically marked. While quiz engines can
present the student with a diagram embedded within a question and ask for a response based on hot spots,
for example, there are no facilities for entering a diagram created by the user. Even if there were such a
facility, it is likely that responses would be routed to a human for grading.
In an attempt to address these problems, we have been developing computer-based tools for teaching and
learning graph-based diagram modelling skills (Thomas et al, 2007a). We refer to these tools as revision tools
as they are typically used by students as formative learning aids to revise or practice skills already taught in a
module. These revision tools provide the student with a series of graded questions each of which requires a
diagram to be drawn. The tools incorporate a diagram marking engine that compares a student-generated
diagram with one or more model solutions and grades the student’s attempt (Thomas et al, 2008a; Thomas et
al, 2008b). The tools also provide graphical as well as textual feedback at a variety of levels of detail which
the student can use to improve their answers and hence their understanding. Our intention has been that the
revision tools should be used as formative learning aids for practicing the modelling skills taught in a module.
In our work, we seek a fully automatically generated grade rather than provide support to the human marker
seen in semi-automatic approaches to diagram assessment (Batmaz and Hinde, 2007; Higgins and Bligh,
2006) which are the most common approaches in this field. In this paper, we concentrate more on describing
the learning tools rather than the automatic marker whose high degree of effectiveness, when compared with
human markers, has been addressed elsewhere (Thomas et al, 2008b).
Borrows Owns *
*
checkIn(s)
g := identifyGuest( s)
HYPOTHALAMUS r := findAFreeRoom()
accept(g, r)
PITUITARY setOccupied(g)
LH
Oestrogens setRoom(r)
FSH
OVARY
links have adornments of a variety of types. Boxes can also have adornments in the form of attributes and
labels, for example.
ITALICS Volume 10 Issue 1 February 2011 - ISSN: 1473-7507 23
At its simplest, an entity-relationship diagram (ERD) is a set of entities (represented as boxes) with
relationships between them. The relationships are either explicit in the diagram, being represented by lines
joining boxes, or implicit, being represented by the juxtaposition of entities. In some notations, attributes of
entities are drawn using other shapes that are connected to the entities by lines and constitute another type of
relationship in the diagram.
When one examines UML sequence diagrams (SDs), there is a great deal of similarity to ERDs in that there
are entities (now called objects) which are related by arrowed lines (representing the sending and receiving of
messages between objects). Sequence diagrams also have activations and different types of message. In
other words, a sequence diagram has a small set of types of things (objects and activations) related in
different ways (indicated by different types of arrows and lines).
However, sequence diagrams, as their name suggests, have one very different feature from ERDs: the
messages and their associated activations are in sequences (indicating the passage of time). Sequences are
represented on a diagram by the relative position of activations and messages, with time conventionally
progressing down a diagram from top to bottom.
In general, diagrams may contain several types of objects (see the sequence diagram in Figure 1, for
example) and several types of relationship (as found in UML class diagrams, for example). Relationships do
not have to be represented by drawn links; it is also possible to represent relationships spatially. Figure 2
shows an entity-relationship diagram (ERD) in which some objects are located wholly within another object. In
this case the notation represents the idea of subtyping (Parent and Child are subtypes of Person). We want to
allow assessment of any of these diagram types.
A further aim is to produce a new question type for an online quiz in which the student would be presented
with a question that requires them to draw a free-form diagram as the answer. The student would be
presented with a drawing tool in the form of an applet with which to create a diagram. The student’s diagram
would be submitted to a marking engine which would compare it with one or more model solutions and apply
a marking scheme (a description of how marks are to be allocated). The marking scheme applicator would
also produce feedback that would be fed back to the student’s browser. This mechanism is illustrated in
Figure 3.
Applet running in
student’s browser Marking Engine running on server
Thus, there are three tools that have to be generalised: the drawing tool, the diagram comparator and the
mark scheme applicator. In our initial investigations we produced standalone applications which integrated the
three tools into a single revision tool.
However, there is another dimension to this work if the tools are to be effectively used. Suitable questions,
model solutions and marking schemes have to be devised. Therefore, we have built a prototype authoring tool
for this purpose (Thomas et al, 2008a).
4. DIAGRAM MARKING
The similarities between ERDs and sequence diagrams have encouraged us to look at other types of
diagram. We have examined biological flow diagrams, UML class diagrams and influence diagrams and found
considerable similarities. These were sufficient to convince us that our approach to diagram marking, in which
we analyse a diagram into what we have called its minimal meaningful units (MMUs) and then aggregate
these into more complex meaningful units (MUs) (Smith et al, 2004; Thomas et al, 2008b) before finding
similar structures in two diagrams to determine their overall similarity, is a viable approach in a range of
diverse domains. For example, in ERDs, the types of MMU are entities, relationships and subtypes.
The relationships between the objects in a diagram are the most significant features of the types of diagram
we are investigating. Our approach compares all MMUs of one type in one diagram with all MMUs of the
same type in the other diagram and determines a similarity value for each pair. An optimisation step
determines, for each type of MMU, the most likely overall match between the instances of that type of MMU in
the two diagrams. Using a similarity value allows for imprecise matching, for example when a student makes
a small error in their diagram.
If any MMU in the student diagram remains unmatched with an MMU in the model solution, the context of that
MMU is examined to see whether it could possibly be paired with an MMU in the model solution. Once the
student’s diagram has been matched with a model solution, the marking scheme is applied.
We normally specify marking schemes in terms of the MMUs in the model solution with marks allocated to
each MMU. Marks are awarded to each MMU in the student solution in proportion to the degree of similarity
between it and the MMU in the model solution with which it is paired. If a similarity is below a specified
threshold, no marks are awarded to that MMU.
We explored the quality of our marking algorithm in a series of experiments (Thomas et al, 2007b). These
have provided positive evidence that our approach determines marks which compare well with those awarded
by expert human markers. After these initial experiments we have continued to improve the marking algorithm
and Table 1 shows the comparison between the current version of the automatic marker and human markers
for a corpus of 394 ERDs using our latest algorithm.
The table shows that there was exact agreement in over 72% of cases. However, given that there is variability
in human marking, we feel that a better measure of agreement is where the difference between the automatic
and human-generated mark is no more than 0.5, namely 93% (the diagrams were marked out of 7 and both
the human markers and the auto marker marked to the nearest half mark.). No diagrams differed by more
ITALICS Volume 10 Issue 1 February 2011 - ISSN: 1473-7507 26
than one mark.
Figure 6 shows good agreement between the automatic and human markers across the range of possible
marks.
50
Hum an
Auto
45
40
35
30
Num ber
25
20
15
10
0
0 0.5 1 1.5 2 2.5 3 3. 5 4 4.5 5 5. 5 6 6.5 7
Mar k
We also looked at the agreement between markers using the AC1 (Gwet 2001) and Kappa (Fleiss 1971) tests
for inter-rater reliability. We use Gwet’s measure as it accounts for chance agreement between markers.
Table 2 (which first appeared in Thomas et al, 2008b) contains three columns comparing the original marks
from the markers (unmod), the marks following moderation (mod - the gold standard) and the marks produced
by the automatic marker (auto).
Both statistical measures show a high degree of conformity between markers and, since the measures are all
above 0.15, we can reject the null hypothesis that the marks were awarded randomly. The results show that
the automatic marker agrees with the moderated marks far more strongly than the unmoderated human
markers agree with the moderated human marks. We conclude that the automatic marker marks better than
most human markers. Further analysis of the results shows that the automatic marker awards marks that are
neither consistently above nor below the humans’ marks. A fuller description of these results appears in
Thomas et al (2008b).
The advantage of this approach is that the algorithm can match broken or erroneous MUs in the student
diagram with MUs in the model solution to determine the most likely match. Then, having determined
Through this mechanism, the teacher selects those values that each adornment can take. The lists of values
will be displayed in the drawing tool so that the student can select a particular value to appear in a particular
place on the diagram. It is possible, if required, to include distracter values in these lists as a further test of
student knowledge.
Figure 10 The student interface for a link Figure 11 Choosing the properties of a box
fixed) and whether attributes, if any, should be displayed using a diagram element (shape) or recorded in a
table.
The specification of a spatial relationship requires just the type(s) of objects involved and the types of their
relationship together with a precision value to be entered. For example, if the relationship exists when one box
is contained in another box the user chooses the relationship Contains to exist between the objects of type
box and enters a value (as a percentage) indicating how much of the inner box has to be within the outer box
for the relationship to be deemed to hold. This is an example of how the tools will allow for imprecision in
drawing a diagram.
Using the specification tool we have successfully generated specifications for UML Class diagrams, UML
Sequence diagrams, top-level context diagrams and biological flow diagrams.
6. ONLINE QUIZZES
We have successfully incorporated our technology into the online quiz component of our institution’s VLE
(Moodle). The Open University has invested heavily in the development of the e-assessment features of
Moodle (being the lead developer). In particular, we have interfaced our own quiz module, OpenMark, to
Moodle (Openmark, 2010). This has meant that, in our implementation, a Moodle quiz can consist of question
types taken from both Moodle and OpenMark. OpenMark supports a much wider range of question types than
the base Moodle quiz and allows us to develop new question types with relative ease. To the user, there is no
apparent difference between questions that originate in Moodle or in OpenMark.
Following the scheme outlined in Figure 3, we have been prototyping a new diagram quiz question type for
OpenMark which presents the user with a question requiring a free-form diagram to be created. Clicking on a
button within the question (see Figure 14), the user launches a drawing tool (appropriately tailored to the
question via the specification tool) in the form of an applet. The drawing tool looks very much like the ones in
the ERD and SD revision tools (Figures 4 and 5).
Once the user is satisfied with their diagram, they press a button to submit the diagram to the marking engine.
In a formative setting, once the marking of the diagram is complete, the score for the diagram plus feedback is
returned to the user and displayed in their browser. See: Figure 15.
To support the use of the new diagram quiz question type our authoring tool has been developed to enable
the teacher to construct questions (text), model solutions (diagrams) and marking schemes in a form that can
be plugged into the OpenMark system. Figure 16 illustrates the question development process.
Define
Data Model File Define Question,
Diagram Type
Model Solution,
and Mark
Figure 16 Question development process using OpenMark and the specification and question authoring
tools
The first step is to define the type of diagram required for the question using the specification tool (assuming
such a specification does not already exist). This produces a data model file which is essentially a set of
parameters. The data model file is input to the authoring tool in order to specialise the authoring tool for that
diagram type. The teacher then builds one or more questions using the authoring tool whose output consists
of two files for each question: a question file containing the text of the question in OpenMark format, and a file
containing the model solution and marking scheme used by the marking engine.
The data model file also plugs into the drawing applet to specialise it for the type of diagram required by the
question. In this way, the output from the specification and authoring tools specialise the drawing tool and
marking engine.
8. REFERENCES
Batmaz and Hinde 2007 Batmaz, F. and C.J. Hinde. 2007. “A Web-Based Semi-Automatic Assessment Tool for
Conceptual Database Diagrams”. Proceedings of the Sixth IASTED International
Conference on Web-Based Education, Chamonix, France, March 14-17, 2007, 427—432.
Fleiss 1971 Fleiss J.L. 1971. “Measuring nominal scale agreement among many raters.” Psychological
Bulletin 76(5):378-382
Gwet 2001 Gwet, K. 2001. Handbook of Inter-Rater Reliability, Gaithersburg: StatAxis Publishing.
Higgins and Bligh 2006 Higgins, C. A. and B. Bligh. 2006. “Formative Computer Based Assessment in Diagram
Based Domains”. Proceedings of the 11th Annual Conference on Innovation and
Technology in Computer Science Education (ITiCSE, Bologna, Italy, June 26-28, 2006),
98—102.
Iizuka et al 2001 Iizuka, K., Tanaka, J. & Shizuki, B. 2001. “Describing A Drawing Editor By Using Constraint
Multiset Grammars”. In Proceedings of the Sixth International Symposium on the Future of
Software Technology (ISFST 2001), Zhengzhou, China. November, 2001.
www.iplab.is.tsukuba.ac.jp/paper/international/iizukia-isfst2001.pdf
Jacobson et al 1998 Jacobson, Ivar; Grady Booch; James Rumbaugh (1998). The Unified Software Development
Process. Addison Wesley Longman.
Jarsen et al 2004 Jansen, A.R., Marriott, K. and Meyer, B. 2004. “Cider: A Component-Based Toolkit for
Creating Smart Diagram Environments”. Springer Lecture Notes in Computer Science. 2980,
pp415-419.
Jayal and Shepperd 2008 Jayal, A. and Shepperd, M. 2008. "The Problem Of Labels In E-Assessment Of
Diagrams," ACM J. of Educational Resources in Computing, 8(4), 2008. Available from
BURA http://hdl.handle.net/2438/2441.
Openmark 2010 “OpenMark examples”, last accessed 28th October 2010
http://www.open.ac.uk/openmarkexamples
Smith et al 2004 Smith, N, P.G. Thomas and K. Waugh. 2004. “Interpreting Imprecise Diagrams”.
Proceedings of the Third International Conference in the Theory and Application of
Diagrams. March 22-24, Cambridge, UK. Springer Lecture Notes in Computer Science, eds:
Alan Blackwell, Kim Marriott, Atsushi Shimojima, 2980, 239—241.
ABSTRACT
This paper reports on a project funded by the HEA-ICS, which sought to create an open educational resource
to help develop media and information literacy skills in postgraduate research students. The resource uses a
distinctive and holistic approach to the teaching of these subjects, which brings together Bruce et al’s ‘Six
Frames of Information Literacy’ model (2007), Egan’s division of literacy into conventional, emergent and
comprehensive (1990), and Whitworth’s objective, subjective and intersubjective domains of value (2009b).
The paper describes the structure and methods of the course, and some results from the summer 2010
evaluation of the draft version. It explains how this evaluation has resulted in the final version of the materials
which are in place on http://madigitaltechnologies.wordpress.com/infoliteracy.
Keywords
media literacy, information literacy, postgraduate research, open educational resource, holistic.
Note
This is a slightly revised and extended version of a paper presented at the HEA-ICS annual conference in
Durham, August 2010. We would like to acknowledge the help that Anne Morris and our (anonymous)
academic and student reviewers gave us in undertaking this project.
1. INTRODUCTION
This paper reports on a project which took place at the University of Manchester and was funded by the HEA-
ICS. The objective was to create a resource which would help develop media and information literacy skills in
postgraduate research (PGR) students. We first describe the project background and the motivations which
lay behind it. We then explore the distinctive model of media and information literacy (hereafter, M&IL)
education in use in the materials, before reviewing how this model has been translated into practice within the
constraints of available resources.
Piloting of the course took place throughout May, June and July 2010, and we report on some results from
this evaluation, before explaining how these informed the final version of the materials. No attempt is being
made here to generalise about the design of IL materials (or online learning generally), but we do try to
establish whether the project met its objectives, which were to create an effective, stand-alone resource that
could be completed in approximately 7-10 hours of independent study by PGR students.
5.1 Syllabus
Applying the holistic model to the needs of PGR students has led us away from a focus on technology and
information searching skills. Certainly these form part of the M&IL resource, but in accordance with the model
illustrated in Figure 1, the skills are located in a wider setting. The general theme is the role of research, and
the researcher, in a media- and ICT-saturated environment.
The structure and content of the resource are as follows:
• Introduction: why six frames of IL?
• Content frame: What is research? The role of information. Where to look. Citation guidance.
• Competency frame: Information literacy procedures and standards. Credibility of information.
• Learning to learn frame: How information becomes knowledge. How our tendency to seek patterns
and support for existing beliefs can prevent us learning.
• Personal relevance frame: Reflecting on information searching. Scientific method and
counterknowledge.
• Social impact frame: Bad science (Goldacre 2009), and other examples of how the media use and
abuse information and research. Media literacy. The information commons, intellectual property and
enclosure (Hess and Ostrom 2007).
• Conclusion: objective, subjective and intersubjective value, and why each one matters.
Remember that the sixth, relational frame represents the course viewed as a whole; the aim being to have
students move between the other frames as appropriate.
In designing the resource we have been mindful of the need to address employability as much as students’
academic skills. Employability is part of institutions’ strategic vision for student and researcher development,
but nevertheless is rarely addressed in most IL schemes (Beetham et al p. 42). Yet only a limited number of
PGR students actually enter research as a profession following graduation, and by no means all of these do
ITALICS Volume 10 Issue 1 February 2011 - ISSN: 1473-7507 38
so in HE (Jeffrey 2010) - and those that stay in HE may not all become researchers. Hence, the content of the
resource has more general application: being relevant to IL skills needed in ‘everyday life’ as well as in the
lab, classroom or research centre.
5.2 Form and methods
The intention was to design the resource such that it could be completed by students in around 7-10 hours of
work, whether in one go, or smaller chunks corresponding to the seven topics listed above, spread over a few
weeks. Each topic contains two audio files, averaging 5-6 minutes in length, which are the principal content
delivery method: there are also some optional readings, additional web-based resources, and three activities
which test participants’ understanding. Note that at this point, the question of how students’ work on this
resource will be accredited has not been resolved, even at Manchester: it is almost certain this will vary from
institution to institution.
Some justification is required. The project brief required the resource to stand alone, that is, not be part of any
wider programme of teaching, and which students could follow independently. Hence there are at present no
collaborative spaces such as discussion boards or wikis included in the resource. Instead, the emphasis is on
content, which may seem to give it a didactic feel. Yet if adopted by other institutions it may be that students
are guided by an instructor, supervisor, or a community of researchers: it may be that students will continue to
work on the resource independently. The resource should be viewed not as a finished course of teaching but
as a base on which may be overlaid other activities, forms of interaction, and so on. Ultimately it could be
embedded in other research support and development activities, such as group seminars or research
methods courses which are discipline-specific and more oriented towards developing communities of
researchers, though note that student reviewers (see §5.3) generally agreed that the resource could stand
alone without such support.
The resource was piloted as a set of Moodle materials, as this was the system in use on the original Media
and Information Literacy resource. However, this proved unsatisfactory when it came to releasing the
resource into the public domain. This is not due to problems with Moodle as such, even if VLEs like this have
been criticised as an “undead” technology (University of Wolverhampton 2009). Instead, the choice was made
for practical reasons, such as removing the need to download very large files (the ‘wrapped up’ Moodle
course coming in at some 140 Mb in size due to all the audio files). The final version of the resource was
created as a stand-alone web site using Wordpress. It can be found at
http://madigitaltechnologies.wordpress.com/infoliteracy. Guidance, for those interested in using the resource,
is available on the site.
There is certainly scope for adopting institutions to incorporate other, more personalisable and/or collaborative
technologies, such as e-portfolios or blogs, in developing research skills. The audio files could be
disseminated as podcasts via iTunes, with supporting downloads for non-audio materials. Again it is stressed
that these are decisions that can eventually be made by adopting institutions. The project specifications
allowed neither the time nor other resources to innovate pedagogically.
5.3 Observations from the pilot
Eight postgraduate students were used as ‘alpha testers’ on the pilot course. These were found by
advertisement and were offered a small fee (£30) as an incentive to follow the course and participate in a
focus group at the end, which was conducted by the development assistant (CW) and recorded. The students
came from several different countries (Uganda, Pakistan, Cyprus and China as well as the UK) and almost as
full a range of disciplines as could be expected considering the number (students came from three of
Manchester’s four faculties, arts, humanities and natural sciences: only the medical school was missing).
They were also diverse in terms of the stage of study: one student was a Masters’ (postgraduate taught)
student, one was about to start on a PhD in September 2010, and others ranged from first to third year PhD
students.
The focus of the evaluation, where the student reviewers were concerned, was on whether the resource met
their needs, and whether its pedagogical methods were effective. Although students were asked to keep
notes on their interactions if they wished, students were given no prior introduction to the resource except in
the most general sense. This was so, as far as possible, their interaction with it would reflect how future
students would first engage with it: that is, knowing it was about Media and Information Literacy, but nothing
else.
General observations made by the student reviewers included the following, which all reflected positively on
the value of IL training generally, and the specific usefulness of the resource:
I was a bit shocked to be honest to have gone through 5 years of university and [prior to completing
course] not actually have a good definition or know what information literacy was. (2nd year PhD
student)
6. CONCLUSION
Information literacy is important; that much is broadly accepted. But there remains a risk that it will turn into
another functional skill, with students’ success measured only against quantitative criteria such as how quickly
they can find information and from what (approved?) sources. Such a view will turn students into mere
information-processing machines. But the research student must be creative, autonomous and individualistic.
IL for PGR students must therefore pay respect to the particular nature of research work.
This is the approach we have taken in developing this resource with the help of the HEA-ICS Subject Centre.
The evaluation suggests that the resource can at least potentially meet needs in this area: only through
adoption and use of the resource can the impact of the materials, and of the holistic model of IL more
generally, be tested. Such follow-up work will be ongoing from this point, with the course coming into use in
Manchester in early 2011.
7. REFERENCES
Beetham, H., L. McGill and A.Littlejohn (2009). Thriving in the 21st century: Learning Literacies for the Digital
Age. Glasgow Caledonian University, Glasgow.
Bruce, C., S. Edwards and M. Lupton (2007). Six frames for information literacy education: a conceptual
framework for interpreting the relationship between theory and practice. In Andretta, S., ed., Change and
Challenge: Information literacy for the 21st century, Adelaide: Auslib (pp. 37-58).
Catts, R. and J. Lau (2008). Towards Information Literacy Indicators, UNESCO, Paris.
Egan, K. (1990). Romantic Understanding: The Development of Rationality and Imagination, Ages 8 – 15.
London: Routledge.
Goldacre, B. (2009). Bad Science. Harper, London.
Hess, C. and Ostrom, E., eds. (2007). Understanding Knowledge as a Commons. Boston, MA: MIT Press.
Janis, I. (1972): Victims of Groupthink, Boston, Houghton Mifflin.
Jeffrey, L. (2010). Generic versus subject-specific researcher training, paper presented at LILAC 2010,
Limerick, Ireland.
JISC/British Library (2008), Google Generation
http://www.jisc.ac.uk/whatwedo/programmes/resourcediscovery/googlegen.aspx
Lau, J., ed. Information Literacy: An international state-of-the-art report. UNESCO/IFLA, Veracruz, Mexico.
OfCom (2009), Report of the Digital Britain Media Literacy Working Group,
http://www.ofcom.org.uk/advice/media_literacy/media_lit_digital_britain/
Streatfield, D., D. Allen and T. Wilson (2010). Information literacy training for postgraduate and postdoctoral
researchers: a national survey and its implications, Libri 60, 230-240.
Thompson, D. (2008). Counterknowledge. London: Atlantic.
University of Wolverhampton (2009), The VLE is Undead, at
http://www.learninglab.org.uk/eventdetails.php?Plv=&Pid1=&Pid2=&id=15.
Vedvik Tonning, A. S., T. Rullestad and T. Skagen. (2010). Integrating information literacy within the university
curriculum: Cooperation between University of Bergen Library and the Centre for University Pedagogy. Paper
presented at LILAC 2010. Limerick, Ireland: Librarians’ Information Literacy Annual Conference.
Whitworth, A. (2009a). Information Obesity. Chandos, Oxford.
Keywords
project guidelines, computer science
1. INTRODUCTION (MOTIVATION)
[included in the Initial Document, Interim Document (if updated), and Final Dissertation]
Although the title of the starting section is “Introduction” it should really be Motivation. In one or two
paragraphs, introduce the topic. This is followed immediately with why this is a useful project, including
possible applications of the work. “What is this dissertation topic good for?” is really the question a good
introduction attempts to answer. After the motivation behind the subject is described, the benefits, and
advantages, and possibly novelty of the dissertation are presented. This is actually where a project could
start. It starts with an idea, and one key to evaluating how good an idea is being able to identify its benefits
and possibly novelty. Each benefit that reading this dissertation provides is identified very explicitly in a
bulleted list. Possible points to mention include:
Words like, contribute, benefit, advantageous, and possibly novel are used in this list. An explicit bulleted list
is used in order to make the subject black and white to the reader otherwise, a reader might not understand
the dissertation. The goal of the project has to be carefully considered in this context. Describe the goal of the
project here.
1.1 Formatting:
The format of the dissertation can be very similar to this document: 11 or 12 point Times (or similar) font,
single column. Excessive white space is to be avoided. Some authors try to pad their dissertation with a lot of
unnecessary white space. Sometimes there are blank pages, or pages that are only half occupied with
content. This is a waste. Don’t start every new section or subsection on a new page in order to make the
documentation appear longer. The goal of a dissertation is not to produce the thickest document possible, but
rather to present good quality. If it looks like too much padding has been used, this may even reduce the final
score of the dissertation.
1.2 On LaTeX:
The best software for writing papers, dissertations, theses, and books is LaTeX. LaTeX produces the highest
quality, most professional-looking output. See Kopka and Daly for an excellent LaTeX reference guide Kopka
and Daly (2004). This is one reason why virtually all conferences and journals provide LaTeX templates for
paper submissions. You can use the LaTeX source from this guideline document if you think it would help.
Simply contact the author. Also, http://www.latex-project.org/ is a great online-guide to LaTeX.
2. BACKGROUND
[included in the Initial Document, Interim Document (if updated), and Final Dissertation]
This section describes the research the author has done in order to prepare for the project. The background
section can be broken down into two sub-sections: (1) related work and (2) previous systems.
2.2 Citations
Any figure, image, or equation that is taken from another source must be cited. Content and terminology from
other sources must also be cited. For more information about citations and their use, see:
http://www.plagiarism.org/. Click on the “How to cite sources” link.
When referring to previous work use names, not only numbers. So, instead of writing, “In Jobard and Lefer
(1997), an algorithm for the creating of evenly-spaced streamlines is presented.”, write, “Jobard and Lefer
present an algorithm for creating evenly-spaced streamlines Jobard and Lefer (1997).” The names are much
more helpful to the reader than the numbers alone. Also, the previous literature was written by real people.
Note that each screen shot is accompanied by a caption, including a citation to the source of that screen shot,
e.g., where the software comes from ( including company name, web page author, web page title, URL, and
last access date, etc). Some examples are given in the references section of this document.
3. PROJECT SPECIfiCATION
[included in the Initial Document, Interim Document, and Final Dissertation]
The Project Specification Section is broken down into two sub-sections: the feature specification and
technology choices.
3.1 Feature Specification
Recall the five stages of software development: (1) requirements specification, (2) software design, (3)
implementation, (4) testing, and (5) documentation, for both programmers and users.
This section is all about stage one–documenting the requirement specification. The requirement specification
describes the functionality of your system. What features is the software to have? What will it be capable of
doing? The feature set is divided into two categories: (1) must-have features and (2) optional features. Must-
have features are the minimum set of requirements the system must have in order to be useful whereas
optional features are those that bring the level of the application up to outstanding quality. Each feature in
both the must-have and optional features sub-sections is presented in a numbered (enumerated) list.
Construction of Table 1 is easy. Simply cut and past each feature listed in the Feature Specification Section
(Section 3.1) and place it under the description column in the table. Then add a completion date. Features
that are too big to be completed in one week can be broken down into smaller pieces that can be
accomplished within one to two weeks.
such as the Introduction, it is not included in the interim document. You may use the minutes of meeting notes
as a history of the project because they document progress that has been made since the initial document
Laramee (2007b).
The table of contents in this document indicates which sections are included in the initial document, interim
document, and final dissertation. This is to be used as a very general guideline and can be customized to fit
most projects. Also read the description of the interim document in your Computer Science Department’s
Course Handbook.
The interim document also requires an updated time table. The time table shows your temporal plan for the
remainder of the project. See Section 4 for a sample time table.
6. PROJECT DESIGN
[included in in the Interim Document, and Final Dissertation]
This section corresponds to the second stage of the software development process Sommerville (2007):
software design. Your design is most likely object-oriented. The design starts with a list of classes and their
responsibilities. This is a very important starting point (the list of classes and descriptions). It forces the
programmer to start thinking about implementation. After identifying a list of candidate classes and writing a
short description of each of them, e.g., 3–4 sentences, the list is reviewed with the project supervisor before
proceeding onwards with the description of relationships between them.
Figure 2: This is an example of a software system showing some of its subsystems. This example shows a schematic of
the design for a visualization system. Only the major subsystems are illustrated.
The design of your system can use the notation of Wirfs-Brock et al. Wirfs-Brock et al. (1990). An example is
shown in Figure 2. A semicircle with an arrow pointing to it represents a contract. A contract is a subsystem or
class interface with other classes or subsystems. It represents the set of services that a subsystem or class
provides. Figure 2 illustrates the different subsystems and the relationships they have with one another.
Class hierarchy diagrams are also very useful (Figure 3). Figure 3 shows different streamline rendering
options displayed in the class hierarchy in which they were designed and implemented. The hierarchy,
following UML notation Fowler (2003), illustrates the is-kind-of relationship between classes. At the top of the
hierarchy we have an abstract base, Streamline Renderer that describes the behavior and contains the
interface that all streamline rendering objects implement.
ITALICS Volume 10 Issue 1 February 2011 - ISSN: 1473-7507 48
Composition diagrams are also very useful (Figure 4). Figure 4 shows the class relationship between the
major components of a texture-based flow visualization subsystem, again using UML notation. However,
rather than showing is-kind-of relationships as in Figure 3 we show composition, a variety of aggregation (is-
part-of relationship). With composition, the part object may belong to only one whole, further,
Figure 3: This is an example of a class hierarchy diagram. This example is from a class hierarchy of streamline and
geometric flow visualization options. UML notation is used.
The parts are usually expected to live and die with the whole Fowler (2003). For example, the Texture Stack
object is part of the Texture-Based Flow Visualizer object and the relationship is one-to-one. Furthermore, an
instance of Texture Stack may be in an instance of Texture-Based Flow Visualizer but not the other way
around. This is indicated by the black diamond shape arrow.
Figure 4: This is an example of a composition diagram. For example, the major components of the texture-based flow
visualization design. Here aggregation, or is-part-of, relationships are shown.
7. IMPLEMENTATION
[started in the Interim Document and completed in the Final Dissertation]
Section 7 is the main content of the dissertation and thus occupies the most space. It is broken down into two
sub-sections: the must-have features and optional enhancements.
8.1 Results
In this section, the nicest images of your system or algorithm are presented. Your system is applied to
interesting data sets and the insight provided is described in addition to a concise description of the data sets.
The characteristics of the test data sets is given. Some possible things to describe are:
1. Where did the data (what domain) come from?
2. How large the are the data sets? What is their resolution in the case of structured data?
3. What are the spatial and temporal dimensionalities of the data?
4. Is the data structured or unstructured?
5. It the data multi-resolution or adaptive resolution?
Hopefully, you’ll be able to show your algorithm being applied to data sets from another domain, e.g., the
visualization of simulation data, in order to demonstrate that your software actually works.
9. CONCLUSIONS
[included in the Final Dissertation]
The Conclusions Section just summarizes the presented work and lists the resulting advantages. Remind the
reader why the presented work is so good and what’s new. Basically, it’s like a concise, one page summary of
the abstract and introduction sections.
12. ACKNOWLEDGMENTS
[included in the Final Dissertation]
This section mentions any people that may have contributed in some way, e.g., with the implementation or
with discussions or feedback. It is tradition to mention your advisor here at minimum. The Acknowledgements
Section also thanks any agency or organization that may have contributed funding to the presented research,
e.g., the EPSRC, KPlus, the NIH, the NSF, etc. An example of a long and complicated acknowledgements
section can be found in previous work Laramee (2004).
Thanks to Daniel O’Sullivan and Nicholas Hammett for providing the impetus behind these guidelines. These
guidelines are the result of approximately four years of third-year project supervision in the Computer Science
Department at Swansea University. We thank the following project students for their valuable discussions and
ITALICS Volume 10 Issue 1 February 2011 - ISSN: 1473-7507 52
input: Essam Farrag, Nicholas Hammett, Marc Harry, Kingsley Hau, Daniel O’Sullivan. Matt Reid, Ferdinand
Vesely, James Walker, and Ho Yuen Wan.
Questions, comments, suggestions for improvement, contributions to these guidelines, or any other feedback
is not only welcome but encouraged. If your department or organization uses these guidelines, we’d like to
hear about it. Please contact Robert S. Laramee.
14. REFERENCES
J. Blanchette and M. Summerfield. C++ GUI Programming with Qt 3. Prentice Hall in associate with Trolltech
Press., Upper Saddle River, New Jersey 07458, 2004. Bruce Perens’ Open Source Series (free for
download).
M. Fowler. UML Distilled: A Brief Guide to the Standard Object Modeling Language. Object Technology
Series. Addison-Wesley, third edition, September 2003.
C. Horstmann. Computing Concepts with C++ Essentials. Wiley, 3 edition, 2003.
B. Jobard and W. Lefer. Creating Evenly–Spaced Streamlines of Arbitrary Density. In Proceedings of the
Eurographics Workshop on Visualization in Scientific Computing ’97, volume 7, pages 45–55, 1997.
Helmut Kopka and Patrick W. Daly. A Guide To LaTeX. Addison–Wesley, 4 edition, 2004.
ABSTRACT
There is an increasing trend towards employer engagement in Higher Education. It is recognised that the
workplace represents a legitimate learning environment in the modern knowledge economy. Higher Education
Innovation Funding is used to foster mutually beneficial links between universities and industry and
commerce. This paper provides a reflective analysis of the utilisation of this funding to establish a software
development team within a department of computer science and information systems to form links with the
regional economy, and to effect synergy between students, staff and external clients. This forms the basis of a
stimulating environment which provides real-world scenarios and brings current skills and knowledge into the
Department. It provides contemporary case studies, offers valuable support to the regional economy, and
enables students to acquire skills to equip them for employment. There is now scope for further curriculum
enhancement and new multidisciplinary initiatives to support applied research.
Keywords
employer-responsive provision, knowledge economy, knowledge-transfer, software development, work-based
learning.
1. INTRODUCTION
There is current concern about motivating and retaining students in Higher Education (HE) and in designing
programmes that will provide them with the skills to function effectively in demanding working environments.
Employer links are increasingly seen as a vital part of university provision, as knowledge generation is
acknowledged to play a significant role, particularly in advanced technical environments. Successfully
integrating provision is not without its challenges, however. This paper presents an account of a recent
knowledge exchange initiative within computer science and information systems, which was co-ordinated by
the University’s Research and Knowledge Transfer Office. This utilised HEIF 4 funding, which is designed to
support and develop a broad range of knowledge exchange activities which result in economic and social
benefit to the UK (HEFCE, 2008). The aim was to establish links with regional businesses and thereby foster
industrial contacts, enhance the curriculum, broaden the Department’s perspective and improve student
employability. The initiative has established firm foundations which promise opportunities to strengthen
regional links and to engage a wider range of students in stimulating, real-world activities to complement
theoretical study.
There is significant debate about the role of work-based learning in HE: Gray (2001), evaluates its
advantages, while Nixon, Smith et al (2006) point up tensions that can arise between HE providers and
employers. There are also questions about the quality of placements as pressure increases to allow students
access to employers (Little and Harvey, 2006). Mason, Williams et al (2003) also challenge the assumption
that employability skills can be developed in HE and transferred into employment. However, it is clear that
there is a marked shift towards recognition of the potential role of the workplace in many HEIs’ provision. The
Quality Assurance Agency for Higher Education’s (QAA) report acknowledges a number of HEIs’ revised
mission statements to accommodate collaboration with business and employers (QAA, 2010). There are
perceived imperatives driving this movement: the HEFCE report (CHERI and KPMG, 2006) identifies the trend
towards high performance management and high performance work practices in organisations featuring
complex job designs, devolved lines of responsibility, and reliance upon access to key business information.
This is consistent with the view that developing countries must rely increasingly upon intellectual capital to
maintain their competitiveness (RAE, 2007). It is also recognised that knowledge creation is not the preserve
4. TECHNICAL ADVANCES
In being exposed to external clients, the team was required to meet a range of user specifications. As
businesses recognised the ways that new technologies could enhance their processes, efficiency, and means
of communicating with the outside world, they sought to use knowledge providers to help them realise these
goals. Equally, as the team acquired knowledge and experience from their developing portfolio, they were
able to advise clients of the kinds of solution that might best meet their business requirements.
Web application development was implemented principally in .NET (Microsoft, 2010) and the team
successfully developed key software components such as a content management system enabling users to
exercise significant control over the material in their web sites. Users were offered facilities such as the
functionality to manage news updates and associated images on their sites. Developers undertook relevant
research to expand their own capability in order to integrate functionality such as shopping cart and on-line
payment facilities into user applications. They also acquired the expertise to integrate features including
search engine optimisation and interfaces to social networking sites into client systems. Noticeably, in a fast-
changing environment where these technologies are becoming more commonplace, there are increasing
expectations as to what clients might regard as basic requirements.
The team has also embarked on projects that require systems to be tailored to more specific business tasks.
Examples include a bespoke mail ordering system, a software version of a business board game and
promotion of products using animation software.
6. CURRENT ACTIVITIES
As the team’s portfolio of completed work has expanded, opportunities to respond to requests for support
within the University, and from external enquiries, have increased. Consequently, the Centre was able to
recruit two more students at the beginning of the last academic year. The team has completed or is working
on over 40 small to medium sized projects to date.
In addition to software development tasks, the Centre has provided training in web application development to
a group of computer programmers from a regional branch of a national organisation. Selected members of the
team composed learning materials from examples of their work. Their experience using a range of
technologies and techniques, such as AJAX to support dynamic web interaction (Garrett, 2005; Shklar and
Rosen, 2009; Welling and Thomson, 2009), enabled them to enhance the provision with informed, timely
insight.
10. REFERENCES
Boes, N. (2010). Vom Bolzplatzhelden zum Bachelor. Frankfurter Allgemeine Zeitung. Frankfurt.
Brodie, P. and K. Irving (2007). Assessment in work-based learning: investigating a pedagogical approach to
enhance student learning. Journal of Assessment and Evaluation 32(1).
BSI (2008). Quality management systems - Requirements. CEN, Brussels.
BusinessLink (2010). Practical advice for business. Retrieved 19th May, 2010, from
http://www.businesslink.gov.uk/bdotg/action/home.
CEPLW (2010). Centre for Excellence in Professional Learning from the Workplace. Retrieved 20th
November, 2010, from www.wmin.ac.uk/ceplw.
Ceppl (2010). Centre for Excellence in Professional Placement Learning. Retrieved 25th November, 2010,
from www.placementlearning.org.
CHERI and KPMG (2006). Towards a strategy for workplace learning, HEFCE.
DevGuru (2010). DevGuru Home. Retrieved 19th May, 2010, from http://www.devguru.com/.
Garrett, J. J. (2005). Ajax: A New Approach to Web Applications. Retrieved 12 January, 2011, from
http://www.adaptivepath.com/ideas/essays/archives/000385.php.
Gray, D. (2001). A Briefing on Work-based Learning, LTSN Generic Centre.
HEFCE (2008). Higher Education Innovation Fund. Retrieved 14th January, 2011, from
http://www.hefce.ac.uk/econsoc/buscom/heif/.
Hislop, D. (2009). Knowledge Management in Organizations A Critical Introduction 2nd Edition. Oxford
University Press, Oxford.
Kerins, J. (2007). Collaborative Initiatives and Work-based Learning Opportunities: A Case Study. ITALICS
6(1).
Leon, W., W. Tynes, et al. (2007). Microsoft SharePoint Server 2007. John Wiley & Sons Ltd, Indianapolis IN.
Little, B. and L. Harvey (2006). Learning Through Work Placements and Beyond. Retrieved 12th October,
2006, from http://www.hecsu.ac.uk/cms.
Mason, G., G. Williams, et al. (2003). How Much Does Higher Education Enhance the Employability of
Graduates?, HEFCE.
Microsoft (2010). Microsoft Developer Network. Retrieved 17th May 2010, from http://msdn.microsoft.com/en-
gb/default.aspx.
Microsoft (2010). The .NET Framework. Retrieved 19th May, 2010, from http://www.microsoft.com/net/.
Nixon, I., K. Smith, et al. (2006). Work-based learning: illuminating the higher education landscape, The
Higher Education Academy.
Pachler, N. and K. Evans (2008). Turning Workplace Learning on its Head. Education Today 57(4).
Prusak, L. and E. Matson, Eds. (2006). Knowledge Management and Organizational Learning A Reader.
Oxford University Press, Oxford.
QAA (2010). Employer-responsive provision survey: A reflective report, The Quality Assurance Agency for
Higher Education.
ABSTRACT
This paper presents an overview of a real-world interaction design assignment where the focus was on
encouraging students to explore the conceptual design space to find creative solutions. By providing them
with background information extracted from a real research study with access to a domain expert, we were
able to assist with enabling a deeper analysis of the context, which is often difficult to do within the resource
limitations of an academic course.
Also by engaging the students in a real case-study we were able to explore the potential for further developing
the work as a cross-disciplinary collaborative research project. Integrating research and teaching in this way
creates an exciting environment where both students and teachers become co-learners.
Keywords
Interaction design, mobile, mental health, anxiety, practice, hci.
1. INTRODUCTION
Mobile interaction design involves the development of interfaces, functions and user feedback for applications
running on mobile devices. To maximise the quality of interaction and ensure that an application meets user
needs, a user-centred design process should be followed which helps developers understand the usage
domain and user motivations. Through a collaborative process between the client, end-users and designers,
and as an outcome of user research, requirements can be determined for building prototype applications. In
the learning situation the provision of realistic contexts for design helps provide crucial exposure to the levels
of detail and nuanced user characteristics that designers will meet in the real world. Providing students with
the opportunity to develop practical design skills can help to bridge the gap between theory and practice. The
challenges of teaching interaction design have been discussed by Sas et al. (Sas, 2006) and include giving
consideration to the fact that as design is mostly a heuristic process, it requires a more reactive and flexible
approach by educators. It is important to encourage students to find alternative ways of thinking about the
material and develop the habits of critical thought.
This paper describes an assignment developed for a module in Interaction Design at Level 3 for the B.Sc.
Web Design at the University of the West of England (UWE) which takes these challenges into consideration
when defining the design brief. The proposed design brief was written to engage students with a real-world
scenario and give them the grounding to go from conceptual design through to a high-fidelity prototype using
an iterative user-centred design process. The assignment brief included background research evidence for
the design idea and high level requirements that students could then use to design and develop a high-fidelity
interactive prototype. The application to be designed was an interactive mobile solution that would enable
students to obtain support for social anxiety in relation to university work (presentations, seminars etc).
This application for the assessment is the outcome of an ongoing collaboration between the Department of
Psychology and the Department of Computer Science and Creative Technologies at UWE.
The module also has a strong research element requiring students to seek primary and secondary information
beyond the teaching delivered in class.
In the teaching of interaction design our focus is to engage students in human-centred design and as such
learn how to consider the close connection between the functionality of what they are designing and the end-
users’ needs and context, with a view to developing applications that are usable and useful, as well as
desirable.
Our goals for assessments at this level therefore include engaging students in work that not only provides a
realistic context to help them develop their skills for translating user needs into tangible design features but
also provides them the opportunity for creativity and to produce work that they can use to showcase their
skills. Students are encouraged to use their assessment work as part of their portfolios (Example Student
Portfolio, 2010) as an illustration to future employers of their practical skills for interaction design.
This assignment followed on from a previous requirements elicitation and design assessment where students
had researched a problem, and had gained skills in using a variety of ethnographic, qualitative and
quantitative methods to collate data from which they derived personas (archetypes of users) and scenarios of
use, using the evidence and insights gained to define a set of requirements. While this assignment gave
them some key practical skills and experience of data gathering, due to the limitations in time and resources
they were able to only focus on a limited perspective.
For their next assignment we wanted them to gain design skills, and work with data that provided multi-
dimensional perspectives of the context to enable taking a more holistic view when considering possible
design solutions.
Linking the design brief to the social anxiety research study provided access to background data collated by
psychologists and provided a high level of ethnographic validity and reliability. From this data we were in a
position to provide robust examples of well-rounded personas, contexts of use and other evidence as a
starting point for this next assignment.
Experience shows that students with social anxiety can learn to cope with it over time through practicing
cognitive behavioural tasks, relaxation techniques and through accessing counselling support from the
institution. Unfortunately, however, the condition means that students are wary of seeking support or even
admitting they are experiencing problems.
As a result of these findings, the Department of Psychology and the Counselling service, in collaboration with
colleagues from Plymouth University had been contemplating the provision of an online diagnostic, self-help
and informational tool to provide additional support to students. As one student put it in the study: "I would
personally like to see more provision for anonymous advice and help, identifying myself to the university as
having problems is in my mind like admitting failure" (Topham, 2009).
Students were advised that their prototypes did not need to fulfil all of these criteria, but that they should
address at least two of the essential requirements. This was an individual assignment, however group
discussions and prototyping activities in class gave the opportunity for collaborative peer learning.
Their designs needed to illustrate consideration of interaction design issues, which included ease of the
communication and interaction, suitability of the interface to support different modes of communication,
efficient mechanisms for switching between modes and smoothness of the transition; as well as social issues,
such the ability to share experiences and content, regard for social protocols especially anonymity and
privacy, facilities for providing awareness mechanisms and limiting scope for negative social phenomena to
occur.
There were two components to the assignment. The first component comprised a report on the design project,
which included their final prototypes illustrating how the different features of their application would be used,
and a reflective commentary discussing their design process and the impact of their approach on the final
prototype. The students were encouraged to maintain a design journal as part of their design process and use
excerpts to underpin their discussion. The second component was a presentation and viva where they were
required to demonstrate how they had interpreted and addressed the requirements of the SASHA project and
employed principles and guidelines for mobile interaction design.
In addition to the website, we also provided opportunities for the students to quiz the “client” for the project,
the psychologist researching social anxiety, who was able to answer queries based on his experience as the
domain expert.
The students also conducted competitor analysis of other similar apps with a view to identifying poor, as well
as useful, usability and interaction design features. The analysis was based on both co-operative and
heuristic evaluations, drawing on specific design guidelines for mobile applications.
For high-fidelity prototyping students were shown how to use MockApp (http://mockapp.com/), Photoshop
mobile design templates (http://www.teehanlax.com/blog/2009/06/18/iphone-gui-psd-30/) and
Genuitec’s Mobione Studio (http://www.genuitec.com/mobile/) which also works as an emulator.
ITALICS Volume 10 Issue 1 February 2011 - ISSN: 1473-7507 66
5. OUTCOMES
The overall work produced had a high element of creativity with the students identifying some innovative
solutions in interpreting and responding to the requirements.
The students followed an iterative user-centred design process which included an analysis phase where they
had to explore the design context in more depth, a conceptual design phase where they focused on particular
scenarios to identify ways in which the requirements could be interpreted (Figure 1) and considered the
holistic nature of the interaction using techniques such as storyboarding (Figure 2), followed by iterative lo-
fidelity paper prototyping (Figure 3), leading to hi-fidelity prototyping (Figure 4).
Realising scenarios via the use of storyboarding is a useful technique for encouraging engagement with the
interaction process in context (Holtzblatt, et al., 2005). Storyboarding helped students to explore different
contexts of use, which in turn helped them to develop clearer ideas regarding the pragmatics of use in
different situations and identify corresponding supporting features.
Students typically addressed requirements for monitoring anxiety status, a self-assessment questionnaire,
semi-anonymous social profiling and sharing of advice and support with other users (Figure 4). The best
quality solutions combined an appreciation of the problem domain with awareness of emerging norms for
touch screen interaction, economy of display usage and social knowledge sharing. Many of the delivered
prototypes were richly annotated, showing links between the design decisions and the contextual evidence.
They also illustrated interactivity through the use of branching or animation within the presentation slides.
We invited our client for the project, Phil Topham, from the Department of Psychology, to the presentations
and vivas and he was able to provide a useful feedback from his perspective in relation to the proposed ideas
ITALICS Volume 10 Issue 1 February 2011 - ISSN: 1473-7507 69
and solutions. The range of ideas and interpretation of the brief by the students impressed him. The students’
work also provided him with an insight into the potential of the technology and enabled him, and us, to think
more broadly about the scope of using this in different situations; this would have been difficult to do without
these tangible exemplars.
Students were able to gain practical experience of using these design methods. Feedback on a student
portfolio incorporating this assignment work, from an international design consultancy, based in Bristol,
provides further acknowledgement of the relevance of the project. Here is a quote from their Managing
Director “…it's great to see a really good and well rounded portfolio coming out of the Web Design course with
a really practical set of skills that can be applied to the workplace.”
From a research perspective, another key outcome of this assessment was that on the basis of this
collaboration and conceptual design activity started by students, we have been able to successfully apply for
funding for a formal research study to design, develop and formally evaluate a social anxiety app. We hope to
use components of some of the design ideas generated by the students, and following contractual IPR
agreements, will be able to offer the students a proportional remuneration for their contribution as appropriate,
together with formal acknowledgement of their work. This student collaboration has proved to be an efficient
model for piloting concepts that can then be taken further in a research project and one that works in an
environment of competing demands on the academic’s time.
6. CONCLUSION
This paper presents an overview of a real-world interaction design assignment where the focus was on
encouraging students to explore the conceptual design space to find creative solutions. By providing them
with background information extracted from a real research study with access to a domain expert, we were
able to assist with enabling a deeper analysis of the context, which is often difficult to do within the resource
limitations of an academic course.
Also by engaging the students in a real case-study we were able to explore the potential for further developing
the work as a cross-disciplinary collaborative research project. Integrating research and teaching in this way
creates an exciting environment where students and teachers become co-learners.
We hope to involve the next batch of students in a follow-up study where we can explore implementation
issues and evaluation studies with real-end users.
7. REFERENCES
Holtzblatt, K., Wendell, J.B., Elsevier.Wood, S. (2005) Storyboarding, In Rapid Contextual Design, A How-to
Guide to Key Techniques for User-Centred Design, pp229-243, edited by Holtzblatt, K., Wendell, J.B.,
Elsevier.
Rosson, MB., Carroll, JM.(2002), Interaction Design, In Usability Engineering, Scenario based Development
of Human-Computer Interaction. pp159-195, edited by Rosson, MB., Carroll, JM. Elsevier.
Sas, C. (2006) Learning Approaches for Teaching Interaction Design, In HCI Educators Workshop, 23-24
March 2006, Limerick, Ireland.
Topham, P. (2009) Feeling stupid: A survey of university students’ experience of social anxiety in learning
situations. Project Report. University of The West of England Bristol, Bristol UK. (Unpublished). Available at:
https://eprints.uwe.ac.uk/164/2/SA_FINAL_REPORT_FEB_2009.pdf
8. Acknowledgements
We would like to thank Phil Topham for providing his expertise on social anxiety and inspiration for the
assignment. We would also like to thank the Interaction Design students, Class of 2009/10 for their work
towards the coursework assignment, in particular Chris Anstey, Jenna Lopez, Karis Cooper, Kieron Gurner,
Sean Payne and Alex Older whose work is featured in this paper.
ABSTRACT
Changes within our approach to teaching can make some students feel uncomfortable. To overcome this,
inquiry based learning, which is strongly supported by research in the areas of intellectual development and
approaches to learning (Prince, 2007), can be used. Inquiry based approaches should be introduced in
combination with existing teaching styles in order to address the needs of all students. Pair programming
enhances the communication among peers and encourages students to ask questions of each other and be
more ambitious in their computer programming practicals. The students subsequently gain confidence from
one another to try different approaches to solving programming problems; this enhances deeper learning.
Additionally, working in pairs provides some students with the courage to ask questions of the teacher while
with their pair, which they may not do alone. This paper presents a case study on using pair programming to
encourage inquiry based learning within programming modules, to improve attendance and practical
assessment results.
Keywords
Pair-programming; assessment
1. INTRODUCTION
The concept of pair-programming has recently been introduced in U.S. universities and companies. By
definition, pair programming is a technique in which two programmers work together on one task at one PC.
One programmer does the coding (the driver) while the other programmer reviews each line of code as it is
typed in (the observer); the two programmers frequently switch roles. The observer should also be coming up
with ideas for code improvements allowing the driver to focus all of their attention on the tactical aspects of
completing the current task (Williams). Within industry, pair programming improves many aspects of
development including design quality, reduced defects, enhanced technical skills and improved team
communications (Cockburn 2001). These are skills that we also require in our students. In 2000 (in Missouri
State University) students were asked their views of pair programming, 70% of them favoured it (Sanders,
2003). Much research has been carried out and published on the benefits of pair programming within both
industry and academia. For example, within an industrial setting, Arisholm et al., (2007) carried out a
controlled experiment in pair programming using a combination of junior, intermediate and senior Java
consultants within Europe. The results suggested that the most benefit was gained by the junior developers in
that they could complete tasks in less time and more successfully than normal with the support of a partner.
The research in (Dyba, 2007) indicated that, given the correct pairing, pair programming could be very
successful. Hence in this case study the students are permitted to select their own partners with the aim of
finding the right pairs. Many projects have been undertaken in pair programming in academia, but primarily in
the U.S. rather than the U.K. The work of Braught et al., (2008) determined that students are more likely to
successfully complete their course using pair programming and are also strengthened as individual
programmers. In (McDowell, 2006; Simon, 2008) this was also found to be true as students have more
confidence in their work and particularly in the case of females in computer science.
Much research has also been carried out and published on the benefits of both inquiry based learning and
working in small groups. Inquiry based learning is an approach to increasing retention through the use of a
teaching approach that allows student-constructed learning rather than teacher transmitted information (The
Sparks Foundation, 2010). An old adage states: "Tell me and I forget, show me and I remember, involve me
and I understand." This is the basic principle behind inquiry based learning: it is any task or activity that
In this paper we present initial findings on the use of pair-programming as a means of embedding enquiry
based learning activities into an algorithmic programming module. We find that the students enjoy working in
pairs and it provides an overall improvement in the module assessment marks.
2. METHODOLOGY
Pair programming as a means of inquiry based learning was introduced into the Algorithmic Programming II
module in semester 2 of the academic year 2009-10 (AY0910). There is a wide range of first year students
who take this module including BSc Computer Science, BEng electronics and Computer Systems and BEng
Games Development. With respect to carrying out practicals and practical assessments, the students were
put into pairs.
2.1 Pairing
In total there were 58 students on this module. The students were allowed to choose their own partners, as
randomly selecting partners can lead to many problems during the course of the module. In week 1 the
students were ask to email the author with details of the partnership (one email per pair). There were initially
some problems with this. Some students selected their partner without asking them if they were willing to be
their partner and this resulted in some students being in two pairs. These issues were raised in the lecture
and students were asked to solve it among themselves. Some students were also reluctant to work in pairs,
but it was not optional, so they had to commit to working with someone. Of course, by week 3, a few students
had dropped out and so their partners had to be paired with someone else. However, none of these issues
were major, and were easily resolved prior to the first assessment taking place.
2.2 Assessment
In their pairs, the students completed the practicals and were assessed. However, assessments were still
conducted on an individual basis in the following manner:
a) both students had to be present at assessments in order to obtain their individual mark;
b) both students were asked individual questions about the practical, hence both students had to
demonstrate their own understanding.
By conducting the assessment in this way, no individual student received credit for their partner’s work.
The assessments were conducting during the timetables practical time. Approximately 5 to 10 minutes was
spent with any one pair and so they were able to return to their normal practical work. At the beginning of the
practical class, all students were requested to select which PC the practical would be running on and have it
running ready for the assessment to be conducted. The practical class would be commenced as normal and
then the assessment would begin. The assessor simply asked each pair to run their program, demonstrate
their output and then show their code.
At the end of the module, the success of pair programming was assessed in two ways: a statistical analysis
was conducted to assess the overall improvement of coursework marks and attendance at the practical
sessions in contrast to the previous year; the students were given a questionnaire in order to collect and
analyze their views on pair programming. As this project was carried out in second semester, the first year
students had already completed one programming module in first semester without using pair programming,
so they were able to compare and contrast the advantages and disadvantages of both learning styles.
It should be noted that practical assessment was only one component of the coursework for this module.
There were two class tests which each student took independently and these were combined with the paired
assessments marks to create a individual assessment mark for each student.
In addition, after the exam component of the module was completed, additional statistical analysis was carried
out to ascertain if the overall exam performance has improved in comparison to the previous year, thus
providing evidence that pair-programming leads to deeper learning and understanding. However, this will also
have to be cross correlated with other module marks to determine if an overall improvement is unique only to
the Algorithmic Programming module, or if indeed the student cohort is academically superior to the previous
year.
3. STUDENT VIEWS
Question % response of ‘yes’
Do you enjoy working in pairs? 76.2
Would you have preferred to work on your own? 28.6
Would you have preferred to work in larger groups? 28.6
Do you feel working in pairs has helped your understanding 57.1
of programming?
Do you feel you have learnt from your programming partner? 57.1
Do you feel your programming partner has learnt from you? 42.9
Did you and your partner work together on programming 47.6
practicals outside the assigned time?
At the end of the module, the students were issued with a questionnaire in order to obtain their views and
opinions on pair programming. They were asked to compare and contrast the pair programming technique
used in this project with the solo programming approach used in their first semester programming module.
This provided us with the student’s overall view of pair programming. In total, 21 students completed the
questionnaire out of a possible 58. The questions and the percentage of students that answered ‘yes’ to each
question is summarised in Table 1.
As illustrated in Table 1, 76.2% of the student cohort enjoyed working in pairs. The general feedback
suggests that knowledge transfer within the partnerships did occur. The evidence also indicates that the
students prefer working in pairs rather than larger groups which could potentially have impact on many other
modules that incorporate group work as means of assessment. It is also interesting to note that 47.6% of the
student cohort worked on their practicals together outside of the timetabled practicals; perhaps the use of pair
programming helps to strengthen bonds and friendships amongst first year students.
An overall aim of embedding pair programming as a means of enquiry based learning in the programming
module was to improve practical attendance, enhance deep learning and thus improve practical assessment
marks with the ultimate goal of improving first year retention. With respect to practical attendance, we found
that there was an overall increase of 2.5% in practical attendance in comparison to the AY0809. Additionally,
when we cross-correlated the second semester data with first semester data for the same cohort we found
that the attendance was significantly higher than expected: practical attendance was 27.8% better than
expected. We investigated whether working in pairs during the practical was having any impact on the
attendance at lectures and found lecture attendance was 10.35% better than expected. This implies that
working in pairs encouraged students to attend in second semester, however there is still room for significant
improvement in the overall attendance.
We subsequently analysed the overall practical assessments for the module to determine if pair programming
improved the practical assessment results. Analysis showed the practical assessment marks had increased
by 31.2% in comparison to last year’s practical assessment marks and were 26.2% higher than the
expectation. Cross-correlation with other modules has also been conducted which highlights that attendance
and performance in the Algorithmic Programming module is slightly better than the Mathematics module.
However, it is not yet clear whether this is strictly due to the pair-programming or the students’ module
preference and hence next year we will try to gauge this via the student questionnaire by including questions
which will advice us about which modules students like and dislike.
In addition to the analysis conducted, we analysed the attendance and performance of the students we
deemed to have engaged in the module, i.e., those that submitted all of their coursework and completed the
exam. If we look only at these students and compare their performance in Algorithmic Programming II with
that of other modules which the same student cohort has completed, we find a significant improvement in both
attendance and assessment as demonstrated in Table 1.
% improvement
The use of pair programming enabled practicals to be assessed in an easy manner particularly with large
module sizes and as the module size increases, it is possible to increase from pairs to small groups. It should
be noted however, that although the pair programming did encourage students to attend practical and lecture
and the module coursework marks were notably better than previous years, this was not reflected in the
examination marks. The examination marks were similar to previous years with no significant improvement
noted.
5. DISCUSSION
The idea behind this project was to introduce a practice in the Algorithmic Programming II module that would
encourage the students to engage and participate in the module and ultimately improve first year retention.
Now we need to consider: did we achieve this? Did we go some way towards achieving this?
Within the U.K. and Ireland, the issue of student retention is prominent. However, throughout the academic
year we can often determine which students are likely to fail first year based on their attendance and amount
of engagement in the degree program. This is particularly prevalent in computer science related disciplines as
students often learn Information and Communication Technologies ICT in schools and think this is computer
science. Therefore when entering the degree program many discover it is not what they first thought.
The main reason for introducing pair programming into the programming module was to get students to
engage with each other and the module, improving attendance, performance and overall retention. The
attendance at practicals improved in comparison to the attendance in AY08/09. As illustrated in Table 1, the
practical assessment marks increased on average by almost 54% compared with the expectation. Table 1
also illustrated improved performance among the females in the class which is also an expectation of pair
programming.
The next major aim of this project was to enhance learning by encouraging students to attend their practicals
and practical assessments and therefore develop a better understanding of programming than in previous
years. The improvement in learning would have been determined by improved class test and examination
marks. Unfortunately this was not the case. The average class test mark dropped by 2% and the overall
examination marks were similar to previous years. However, this was primarily due to approximately 8
students who never engaged in the module and did not sit the exam. This is still a relevant point as the aim of
the project was to improve engagement and attendance. Pair programming cannot benefit those who do not
take part. However the students who participated in pair programming clearly benefitted.
6. CONCLUSION
We have presented an approach to embedding inquiry based learning into algorithmic programming modules
via the use of pair programming. With respect to pair programming students work together in pairs and are
also assessed in pairs in their practical assessments. We have presented initial findings that illustrate that the
use of pair programming encouraged students to attend lectures and practicals and also, overall, the module
assessment marks were significantly higher than expected based on the cohorts experience performance and
the assessment marks from previous years. However, it is also evident that the use of pair programming did
not improve the examination marks in any significant way.
APPENDIX A
In order to obtain the expected results, the total possible attendance from the first semester's module was
obtained by multiplying the number of students by the number of weeks. Next the attendance every week of
the semester was summed and divided by the total possible attendance to obtain the attendance percentage.
This was carried out for the current and previous academic years. The comparison of these first semester
attendance records was then used against the previous year's Algorithmic Programming II module attendance
to achieve the expected attendance for the current Algorithmic Programming II module. The same method
7. ACKNOWLEDGMENTS
The authors would like to thank the Higher Education Academy Information and Computer Science Subject
Centre for funding this project.
8. REFERENCES
1. Arisholm, E., Gallis, H., Tore, D., Sjoberg, D.I.K., “Evaluating Pair Programming with Respect to System
Complexity and Programmer Expertise” IEEE Trans on Software Engineering, Vol. 33, No. 2, Feb 2007.
2. Braught, G., Eby, L.M., Wahls, T., “The Effects of Pair-Programming on Individual Programming Skill”
SIGCSE, March 2008.
3. Cockburn A., Williams, L., “The Costs and Benefits of Pair Programming” The XP Series, Extreme
Programming Examined, pp. 223-243, 2001
4. Dyba, T., Arisholm, E., Sjoberg, D.I.K., Hannay, J.E., Shull, F., “Are Two Heads Better than One? On the
Effectiveness of Pair Programming” IEEE Software, Vol. 24, No. 6, pp12-15, Dec 2007.
5. McDowell, C., Werner, L., Bullock, H. E., Fernald, J., “Pair Programming Improves Student Retention,
Confidence and Program Quality” Communications of the ACM, Vol. 49, No. 8, August 2006.
6. Prince, M., Felder, R.W., “The many faces of inductive teaching and learning” Journal of College Science
Teaching, Vol. 36, No. 5, pp14-20, March/April 2007
7. Prince, M., Felder, R.W., “Inductive teaching and learning methods: Definitions, comparisons, and
research bases”. Journal of Engineering Education 95 (2): 123–38, 2006
8. Rice, P., “Facilitating learning in small groups” http://www.heacademy.ac.uk/
9. Sanders, D., “Student Perceptions of the Suitability of Extreme and Pair Programming” Computer Science
Education, 2003
10. Simon, B., Hanks, B., “First-Year Students’ Impressions of Pair Programming in CS1” ACM Journal on
Educational Resources in Computing, Vol. 4, No. 7, Jan 2008.
th
11. The Sparks Foundation (2010). Why promote inquiry based science programming? Retrieved August 6
2010. http://scienceexplorers.com/wp-content/uploads/promoting-inquiry-based-learning.pdf
12. Williams, L., "Integrating Pair Programming into a Software Development Process" (PDF). 14th
Conference on Software Engineering Education and Training, 2001
ABSTRACT
The problems associated with learning and teaching first year University Computer Science (CS1)
programming classes are summarized showing that various support tools and techniques have been
developed and evaluated. From this review of applicable support the paper derives ten requirements that a
support tool should have in order to improve CS1 student success rate with respect to learning and
understanding.
Keywords
Introductory programming, learning, design, human factors, theory.
1. INTRODUCTION
Students learning to program are often initially frustrated by not only having to learn code syntax, but by
having deeper issues of understanding than is evident from any compilation failure. Students require to
engage with programming on both a cognitive and a conceptual level and, as help is not always immediately
at hand in a busy laboratory or in private development, an automated tool supporting those concepts, as well
as fundamental syntax and semantics, would be a useful addition to their educational armoury. The work
described here is the first part of a doctoral research project to determine what was necessary in a student
support tool for learning to program, and, secondly, to build and evaluate such a tool. The aim of the initial
research, therefore, was to review and then synthesize current research and practice in the field and to
determine fundamental support requirements that any learning system should have.
The paper outlines the current state of the art in learning to program research and practice followed by a
discussion of the applicable learner models. These give rise to the emergent requirements for a support tool
and the conceptual architecture for such a tool is then outlined. The paper concludes with a short summary of
the second part of the research project in which the requirements were substantiated within a prototype tool.
2. RELATED RESEARCH
The specific task of learning to program attracts particular challenges, and it is well recognized as a time-
consuming and frustrating process for the majority of novices (Johnson 1990; Edwards 2003; Kelleher and
Pausch 2005; Bennedsen and Caspersen 2007; Eckerdal 2009). Progress has been made in terms of
identifying the underlying difficulties faced by novices by investigations of possible underlying influences
(Simon, Fincher et al. 2006) and, separately, direct comparison with expert programmers (Mead, Gray et al.
2006; Kay 2007). Some older studies, for example, Soloway and Ehrlich (1984) explored the difference
between novice programmers and experts. These older studies lay the foundation of our current
understanding of the programming process. The Soloway and Ehrich paper states that the process of
programming requires at least two different skill sets: program comprehension (i.e. the ability to understand
each program part) and, secondly, plan recognition (i.e. the ability to understand the interactions among each
program part). Program comprehension is dependent on an understanding of the form and meaning of each
symbol and each line, and this skill is necessarily evident in experts. This study also demonstrates that
understanding arises as a result of recognition of the interaction among program lines in the form of a
More recently, Kay (2007) suggests using errors to determine and then ultimately predict erroneous actions of
beginners, intermediates and then advanced programmers. He also states, along with Ebrahimi and
Schweikert (2006) that mistakes came from incorrect plans or mental models and a previous experiment
(Lazonder and Van Der Meij 1995) demonstrates that novices require context specific information when an
error has occurred. Ebrahimi (1994) undertakes a review of existing work on the difficulties faced by novice
programmers and clearly states that the two challenges noted by the studies above impact significantly on the
learning process for novices. It is clear that knowledge of the syntax and underlying semantics is required to
develop programs. However, Ebrahimi states unambiguously that program plans are also a fundamental
requirement to successful program development. Ebrahimi also shows, through observational study, that the
concepts of “language constructs” and “plan composition” are central to programmer development and
intimately linked. Ebrahimi concludes that programmers cannot form plans without language and cannot use
language without plans and that, consequently, these must be taught concurrently.
The literature described above demonstrates that there are two challenges in learning to program. The first,
here termed program formulation is the syntax and semantics of the individual program parts that are
fundamental to the formulation of any program. The second, here termed problem formulation is therefore the
challenge of identifying and structuring the necessary program components to formulate a programmed
solution to a specific problem.
Both Fuller et al. (2007) and Rountree et al. (2005) state that code comprehension and programming are two
separate skills and need to be developed. Lister et al. (2004) investigated students’ ability to trace programs.
All these skills are considered essential for developing a student along the path towards experience. Malan
and Halland (2004) identified pitfalls to avoid when teaching programming, specifically identifying the problem
of teaching being too abstract and confusing the novice student. Edwards states that students need “explicit,
continually reinforced practice in hypothesizing about the behaviour of their programs….[and] need frequent,
useful and immediate feedback” (2003). Students may believe that once their program compiles then it works
and must be correct. This suggests that their mental model is perhaps not accurate or that they have failed to
understand the logic of their code. Murphy et al. (2008) states that novices behave unproductively when
debugging their programs implying confused mental models. Norris et al. (2008) notes that better performing
students spend less time on a task, using fewer compilations and writing less code implying better mental
models of their code.
Lahtinen et al. (2005) suggest that studying alone and example programs were the best mechanism for
learning and Fetaji et al. (2007) states that learning to program in a classroom approach is not the most
appropriate strategy. Hassinen and Mayru (2006) support a learning-by-doing theory and Shaffer (2003)
further suggests limiting the number of concepts introduced in a class. Further, Pillay (2003) states that one
on one tutoring is the best mechanism for helping novice programmers.
3. LEARNER MODELS
In order to improve the learning process that underpins the development of expertise in programming, a
number of models of student learning have been devised (Mayer 1981; Mayer 1997; McGill and Volet 1997).
These models are typically formulated in terms of stages in progressing from a novice to an expert
programmer and can be used to identify and describe the composite aspects of the activity of learning to
program. In turn, these activities, and the accordant skill base, can be used to determine the direction that we
expect students will follow and provide guidance and structure on how we need to help students begin this
process of learning.
Program Formulation
Program Formulation encapsulates both the Syntactic and Semantic knowledge levels of Mayer (1997).
According to Lund (2002), Syntactic knowledge is specifically concerned with the syntax, or form, of a
programming language and the ability to apply that knowledge. Correct syntactic expression of constructs
within a particular programming language allows development of Semantic knowledge, i.e. the operation of
those individual constructs, which will enable them to develop a mental model of program execution or “what
goes on inside the computer as a result of a line of code” (Lund 2002). This mental model, combined with
knowledge of the underlying syntax, represents the skill set needed for program formulation.
Problem Formulation
Lund also defines Schematic knowledge as the ability to recognize patterns in code developed for previous
problems, also known as programming plans, and apply these plans to form a solution to the current problem.
As noted earlier, novice programmers have difficulty recognizing these patterns and, even when the patterns
are made explicit in one context, they are unable to transport these patterns to a given problem, i.e. another
context. The final layer in the model, Strategic, may be defined as the techniques employed to create and
monitor plans, so following a software development process (Lund 2002). Therefore, knowledge at the
syntactic and semantic levels maps directly on to program formulation and knowledge at the schematic and
strategic levels maps directly on to problem formulation.
Simon et al. (2006), state that students with pre-existing strengths in strategic or algorithmic articulation are
invariably successful programmers. One could interpret this as these students progress quickly through the
levels because they are already thinking at the problem formulation level and can therefore cope with lower
level issues such as syntax errors. Kordacki (2007) implies the ability of learners to use representation
systems to express their problem-solving strategies will make for better learners, suggesting the use of both
program and problem formulation. Soloway and Ehrlich (1984) previously claimed that a major source of code
problems came from plan composition failures, i.e. program formulation. A similar claim was also made by
Winslow (1996) when he stated that “..novice programmers know the syntax and the semantics of individual
statements, but they do not know how to combine these features into valid programs.”
4. EMERGENT REQUIREMENTS
From reviews of existing tools (Kelleher and Pausch 2005; Coull 2008) it is evident that a support tool may be
used to facilitate the learning process of novices who are learning to program. Further, from this review, a
series of requirements emerge. Generally, there is diversity in methods and approaches used to teach the
complex, interleaved concepts associated with programming. Therefore, a support tool that is sufficiently
flexible to accommodate many different instructional and delivery designs is of the most value. Ideally, the
support tool should be decoupled from any specific development requirement and further operate as an add-
on to any chosen programming environment. No support tool that aids program formulation with respect to
syntax should promote dependence on that tool to such an extent that the student is not able to progress
beyond the remit of that tool. To remove this dependence, and to benefit from support that takes account of
the taught context, it is necessary to present the student with both the standard compiler messages and any
supplementary error messages enhanced with context-relevant support. Again at the program formulation
level, and in recognition of the frequent semantic errors which students make, a support tool should detect
these common mistakes and provide relevant warnings to the novice. To further reduce the risk of
dependency the level of support should be reduced over time. In addition to these steps, the use of a support
tool should be voluntary as this affords students the opportunity to direct their own learning and allows those
students with existing experience to opt out of extended support provision and develop programs more
quickly.
To support a student more fully in the development of a solution to a given problem, i.e. problem formulation,
a support tool needs to have knowledge of both the key constructs and the relationship among those
constructs necessary for a working solution. A support tool should be able to offer assistance to the student in
identifying which constructs are needed to solve a particular problem and make sure that their solution meets
The architecture comprises five main dimensions: Syntactic, Semantic, Schematic and Strategic structure
support together with interaction, i.e. invocation and feedback, with the support tool itself. To provide
Syntactic Support for program formulation it is necessary to supplement the (original) error messages arising
from the compilation process with extended text, where this text is sensitive to the taught context. To provide
Semantic Support for program formulation it is necessary to perform checks on the program code for common
(logic) errors and provide warnings to students, which indicate these logic errors. Knowledge of the problem
domain (a specific practical exercise) may be used to inform the development of strategic and schematic
checks for solution formulation, i.e. checks for key constructs and the relations among those key constructs
respectively. Through this mix of syntax, semantic, schematic and strategic checks, the support tool may
provide feedback to the student. This support must be integrated into the current development environment to
provide a seamless interface that is available whenever students elect to use it.
Similarly, Semantic Support is also very important for program formulation. Although problems at this level do
not occur with the same frequency as syntax errors, their presence is less obvious to the novice, and so they
are harder to detect, since they are manifest only at run-time. As there is no compiler support for these errors
students may spend a long time trying to identify and then rectify the problem. Indeed, based on observations
(Jadud 2006) students often make structural changes to their program in an attempt to correct its behaviour
rather than identifying a (more trivial) semantic error. Some other support tools, for example BlueJ (Johnson
2006), recognize the value of providing support for semantic problems and have incorporated a series of
semantic checks into the software support. To address Requirement 3 a support tool should perform checks
on syntactically valid code for commonly observed semantic errors. Again, the dialogue used in feedback
should be something that the student is familiar with and, where possible, relate to the content of the taught
material (Requirement 2).
Feedback
To enable students to become more confident in their abilities, and develop a skill set independent of any
additional support, it is necessary to reduce the amount of feedback provided as teaching progresses
(Requirement 9). The progressive reduction in support over the taught period will ultimately eliminate
dependence on a support tool. To ensure that the feedback is relevant and contains terminology that the
student is familiar with, the feedback should make a direct relation to core and/or supplementary teaching
material where possible (Requirement 2). To direct their own learning students must have the opportunity to
work with just the basic toolset, for example the compiler messages, and so invocation of any software
support must be at the individual student’s discretion (Requirement 10).
6. EXPERIMENTAL EVIDENCE
To evaluate the derived requirements and the generic architecture, a prototype implementation, SNOOPIE,
was built and linked to specific programming courses. The remit and feedback of the tool support was initially
informed by an extended series of laboratory observations. These demonstrated the multi-faceted nature of
problems that students encounter whilst they are learning to program and how these problems can be
mapped onto the different levels of program and problem formulation. Detecting erroneous problem
formulation is not easy to identify and analysis was done through interview, questionnaires and experimental
evidence. Erroneous program formulation can be detected through compilation errors and the number of
attempts and time taken to achieve a solution. The support tool was fully evaluated and demonstrated to have
a significant impact on the learning process for novice programmers. From a series of experiments with
multiple student cohorts over two years in two universities each using a different series of laboratory practicals
and lectures, five hypotheses were evaluated: that program formulation support improved short-term student
performance, that problem formulation support improved short-term student performance, that problem
formulation support reduced the time taken to achieve an ideal solution, that a combination of problem and
program formulation improved long-term student performance and that the system was perceived as having a
positive impact on the student learning process. Hypotheses 1 to 3 were tested using laboratory experiments,
hypothesis 4 was analysed using student performance at the module level combined with interviews with staff
and students and finally, hypothesis 5 was tested via experiments, questionnaires and interviews with
students. The results demonstrated that, using the SNOOPIE tool:
1. Students receiving program formulation support were able to solve errors faster than those without
support, i.e. program formulation support improves short-term performance
2. Students receiving problem formulation support were more able to produce a solution that met the
criteria than those without support, i.e. problem formulation support improves short-term performance
3. Students receiving problem formulation support took longer to produce a solution than those without
support. It was concluded that students were not being “spoon-fed” solutions and took time to think in
7. SUMMARY
Learning to program is recognized as a complex task that novices find challenging. Existing literature
demonstrates that learning to program is difficult because of the need to learn the rules and operation of the
language (program formulation), and the concurrent need to interpret problems and recognize the required
components for that problem (problem formulation). There exist many endeavours to support the novice in
this activity, including software tools that aim to provide a more supportive environment than that provided by
standard software facilities. This paper presents a new methodology for developing learning support tools that
addresses the dual task of program and problem formulation. A review of existing teaching tools that support
the novice programmer lead to a set of requirements for a support tool that encompasses the processes of
both program and problem formulation. The set of requirements is encapsulated in a generic architecture for
software tool development. A prototype support system was then built from the architecture and a series of
hypotheses was developed to determine system validity. By using a combination of evaluation methods
including anecdotal evidence, interviews, questionnaires and experiments, the results demonstrated how a
series of automated analyses at different stages in the student's development of a solution improved both
student confidence and performance.
8. REFERENCES
Bennedsen, J. and M. E. Caspersen (2007). "Failure Rates in Introductory Programming." Inroads (SIGSCE
Bulletin) 39(2): 33-36.
Coull, N. J. (2008). Support for Novices in an Object Oriented Programming Integrated Environment. School
of Computer Science. St Andrews, Scotland, KY16 9SX, University of St Andrews. PhD.
Ebrahimi, A. (1994). "Novice programmer errors: language constructs and plan composition." International
Journal of Human-Computer Studies 41: 457-480.
Ebrahimi, A. and C. Schweikert (2006). "Empirical Study of Novice Programming with Plans and Objects."
Inroads (SIGSCE Bulletin) 38(4): 52-54.
Eckerdal, A. (2009). Novice programming students' learning of concepts and practice. Information
Technology. Upsalla, Sweden, University of Upsalla. Ph.D.
Edwards, S. H. (2003). Rethinking Computer Science Education from a Test-First Perspective. OOPSLA’03
148-155.
Fetaji, M., S. Loskovska, et al. (2007). Combining Virtual Learning Environment and Integrated Development
Environment to Enhance eLearning. International Conference on Information Technology Interfaces.
Croatia.
Fuller, U., C. G. Johnson, et al. (2007). "Developing a Computer Science-specific Learning Taxonomy."
SIGCSE Bulletin 39(4).
Hassinen, M. and H. Mayra (2006). Learning Programming by Programming: A Case Study. Baltic Sea '06
Koli Calling.
Jadud, M. C. (2006). Methods and Tools for Exploring Novice Compilation Behaviour. Proceedings of the
2006 International Workshop on Computing Education Research. Canterbury, UK, ACM: 73-84.
Johnson, C. (2006). Abstract interpretation of student programs as a strategy for courseware development. In
Methods, Materials and Tools for Programming Education, Tampere Polytechnic: 14-19.
Johnson, W. L. (1990). Understanding and debugging novice programs, Artificial intelligence and learning
environments. Scituate, MA, Bradford Company.
Kay, R. H. (2007). "The role of errors in learning computer software." Computer & Education 49: 441-459.
Kelleher, C. and R. Pausch (2005). "Lowering the Barriers to Programming: A Taxonomy of Programming
Environments and Languages for Novice Programmers." ACM Computing Surveys 37(2): 83-137.
ABSTRACT
This paper compares two different approaches of teaching introductory programming by quantitatively
analysing the student assessments in a real classroom. The first approach is to emphasise the principles of
object-oriented programming and design using Java from the very beginning. The second approach is to first
teach the basic programming concepts (loops, branch, and use of libraries) using Python and then move on to
oriented programming using Java. Each approach was adopted for one academic year (2008-09 and 2009-
10) with first year undergraduate students. Quantitative analysis of the student assessments from the first
semester of each year was then carried out. The results of this analysis are presented in this paper. These
results suggest that the later approach leads to enhanced learning of introductory programming concepts by
students.
Keywords
Teaching programming, object oriented approach, procedural approach, java, python.
1. INTRODUCTION
Teaching programming is a complex task. The task is even more challenging for introductory modules. Failure
rates are not marginal (Bennedsen & Caspersen, 2007). One of the common problems shared by Computer
Science Departments is the lack of basic programming skills reported by module leaders of courses following
first programming courses and how to equip students with better programming skills after the introductory
courses. First programming courses typically emphasise the principles of object-oriented programming and
design from the very beginning. An alternative approach is based on starting with a more traditional
procedural approach first. The evidence from past research seems to suggest that Object Oriented
Programming is generally more complex than procedural (Robins et al., 2003). McCane (2009) has
demonstrated the effectiveness of Python to teach introductory programming by using qualitative analysis.
In this paper it is therefore suggested to introduce Python (Python Software Foundation, 2010) for the basic
(procedural) aspects of programming and then introduce Java (Java, 2010) to focus on object-oriented
aspects. The use of Python is expected to both reduce the overhead attached to Java syntax and allow
immediate feedback while the students practice with basic instructions (due to the interpreted nature of
Python). In other words, at the beginning of the module, novice student are able to devote particular attention
to procedural concepts, flow of control, flow of data, etc.
The aim of this paper is to evaluate the success of using Python to introduce basic concepts (loops, branch,
and use of libraries). Such evaluation is based on the analysis of student assessments. To measure the
success of the proposed method, quantitative analysis of the assessments outcome are presented and
discussed. Ehlert (2009) highlights the importance to develop an experimental setting for a fair comparison
and the importance to control the variables carefully. The experimental setting used in this paper controls and
defines the variables carefully and measures them using clear indicators.
3. METHODOLOGY
Within this paper Python assessment submissions are compared with Java ones. The Python and Java
assessments are very similar and therefore comparisons should give a reliable measure. The aim of the
comparison is to quantitatively measure a student’s ability to master basic programming concepts when
Python is used instead of object-oriented Java. Both assessments consisted of students having to implement
a program (See below Section 4 for further details).
The following three comparisons have been carried out: grades, programs with bugs and frequency of
keywords. The grades of the submitted program are a measure of success of the student ability to implement
programs. Bugs are interpreted as a measure of their overall understanding of programming. Frequency
values of keywords are interpreted as a measure of the familiarity with basic concepts of programming. The
four keywords used in the frequency measure are “if”, “for”, “while” and “import” (in combination with the
use of “random” class). They are indicators of student’s use of conditional, loops and import (the random
library) statements. Occurrences of these keywords in commented sections of the code have been discarded.
No distinction has been made between correct and incorrect use of the statements in the frequency
measurement.
Two different academic years have been used to collect data. During the 2008-09 academic year, the
approach to emphasise the principles of object-oriented programming and design using Java from the very
beginning is followed. During the 2009-10 academic year, the approach to first teach the basic programming
concepts (loops, branch, and use of libraries) using Python and then move on to oriented programming using
Java is followed. In both cases, students were assessed after approximately ten weeks of teaching. Java
assessments are from 2008-09 academic year, whereas Python assessments are from the 2009-10 academic
year. Also all students in a cohort have (some) similar background since they have to fulfil similar
requirements to be accepted. Some may have a better proficiency than other but this is true in equal measure
for both cohorts. The research is based on the assumption that students cohort for the academic year 2008-
09 and 2009-10 are of similar academic level. This assumption is supported by the fact that the admission
criteria for both the years were the same. No student was in both cohorts. Table 1 summarises the student’s
cohort.
Students are assessed for their ability to implement programs. Criteria for the assessments include:
generalizability, complexity, coverage of loops, conditionals, etc. Moreover, the ability to produce a program
that could compile and run without errors has been considered a requirement for a pass. However, for both
Java and Python data, small typos (a missing semicolon, etc.) have not been considered bugs (even if the
submitted program did not run correctly). That is, they have not been graded as a fail grade, although they
have been graded as just above a pass. To assess the style of the submitted program the source code was
inspected.
Students have an hour to complete the task and they can use the same environment as the one they used
during their lab tutorials (BlueJ, Python, etc.). That is, during the test they can implement, run and debug their
programs. They have access to both the standard Java/Python documentation and to the module resources
(lecture notes, lab examples, etc.). Tasks are kept as similar to one another as possible in both assessments.
Table 3 shows the grade distribution. The use of loops, conditional, etc. to achieve the given task was part of
the marking criteria. For example, a grade A requires the correct use of conditional and the correct use of
loops and the correct use of libraries in the program implemented. The last column (D(40)) shows the number
of occurrences of programs that did not run without debugging but where bugs have been nevertheless
considered of minor entity(and therefore graded as just a pass).
The results presented in Table 2 show that there is evidence of a significantly increased use of all the four
markers when Python is used as programming language. Moreover, the occurrence of bugs is lower with
Python than with Java. The results presented in Table 3 show that grades distribution shows a higher
percentage of programs able to meet the minimum requirements for a pass when Python has been used.
These results imply that our approach to introduce Python for the basic (procedural) aspects of programming
and then switch to Java to focus on object-oriented aspects enhanced students understanding of the
introductory programming.
6. CONCLUSIONS
This paper presents experimental results from real classroom comparing two different approaches to teach
introductory programming. The first approach is to emphasise the principles of object-oriented programming
and design using Java from the very beginning. The second approach is to first teach the basic programming
concepts (loops, branch, and use of libraries) using Python and then move on to oriented programming using
Java.
The first teaching approach using Java has been followed for the academic year 2008-09 and the second
teaching approach using Python has been followed for the academic year 2009-10. The student assessments
towards the end of first semester in each academic year have been used to carry out the experiments
described in this paper. Three indicators namely grades, programs with bugs and frequency of keywords have
been used for comparison. The grades of the submitted program are a measure of success of the student
ability to implement programs. Bugs are interpreted as a measure of their overall understanding of
programming. Frequency values of keywords are interpreted as a measure of the familiarity with basic
concepts of programming. The four keywords used in the frequency measure are “if”, “for”, “while” and
“import” (in combination with the use of “random” class).
The experimental results discussed in this paper show a measured positive (increase) in all the three
indicators in favour of the second approach which is to first teach the basic programming concepts (loops,
branch, and use of libraries) using Python and then move on to oriented programming using Java. The use of
Python could facilitate the mastering of basic concepts such as loops, branch, and use of libraries for novice
students. A possible explanation could be that with Python students can focus on the crucial basic issues
without being distracted by the overheads. Moreover, the complexity of the programs used during their
practice could be tailored in a more accurate way to their level of proficiency.
One limitation of this research is that it considers only one case study. Such case studies require considerable
amount of time as they have to fit in with the academic calendar and hence we could only consider one case
study. We hope that similar case studies will be replicated by other institutions as well. The results from our
case study are positive and statistically significant in favour of our proposed approach which is to first teach
the basic programming concepts (loops, branch, and use of libraries) using Python and then move on to
oriented programming using Java. We hope that our research will encourage debate over different
approaches to teaching introductory programming and in turn will lead to adoption of better teaching
methodologies.
Also this research focuses only on the quantitative analysis of the student assessments and does not elicit the
views of the students about their learning experience. So an interesting direction for future work would be to
collect and analyse the qualitative data from the students about their perspective.
This research takes into account the performance of students up until the first semester. We are analysing the
performance of students over the whole academic year and will be publishing it at a later stage.
BlueJ, (2010), BlueJ - Teaching Java - Learning Java, http://www.bluej.org/ (01 Dec 2010)
Bruce, K.B. (2005), Controversy on how to teach CS 1: a discussion on the SIGCSE-members mailing list,
ACM SIGCSE Bulletin, 37(2), pp. 111-117
Decker, A. (2003), A tale of two paradigms, Journal of Computing Sciences in Colleges, 19(2), pp. 238-246
Engel, G., Roberts, E. (2001), Computing Curricula 2001 Computer Science: final report, IEEE Computer
Society and Association for Computing Machinery, 28, pp. 2004
Ehlert, A., Schulte, C. (2009), Empirical comparison of objects-first and objects-later, ACM Fifth international
workshop on Computing education research workshop, pp. 15-26
Lemos, R.S. (1979), Teaching programming languages: A survey of approaches, ACM SIGCSE Bulletin,
11(1), pp. 174-181
Lister, R., Berglund, A., Clear, T., Bergin, J., Garvin- Doxas, K., Hanks, B., Hitchner, L., Luxton-Reilly,
A.,Sanders, K., Schulte, C., and Whalley, J. L. (2006), Research perspectives on the objects-early debate.
ACM SIGCSE Bulletin, 38(4), pp. 146-165
McCane, B. (2009), Introductory programming with python, The Python Papers Monograph, 1, pp. 1-18.
Pears, A., Seidman, S., Malmi, L., Mannila, L., Adams, E.,Bennedsen, J., Devlin, M., and Paterson, J. (2007),
A survey of literature on the teaching of introductory programming, ACM SIGCSE Bulletin, 39(4), pp. 204-223
Python Software Foundation, (2010), Python Programming Language, http://www.python.org/ (01 Dec 2010)
Robins, A., Rountree, J., Rountree, N. (2003), Learning and Teaching Programming: A Review and
Discussion. Computer Science Education, 13(2), pp. 137-172
Reges, S. (2006), Back to basics in CS 1 and CS 2, ACM SIGCSE Bulletin, 38(1), pp. 293-297
The R Foundation, (2010), The R Project for Statistical Computing, http://www.r-project.org/ (01 Dec 2010)
Vilner, T., Zur, E., Gal-Ezer, J. (2007), Fundamental concepts of CS1: procedural vs. object oriented
paradigm-a case study, ACM SIGCSE Bulletin, 39(3), pp. 171-175
ABSTRACT
The experiences of students studying Computer Forensics at the University of Derby, together with the views
of the teaching and technical support staff, strongly suggests a growing mismatch between expectations and
reality, an underdeveloped ability to problem solve and communicate findings and unrealistic expectations of
the job market.
In improving the conceptualisation of how our courses can better equip students for employment and how
their understanding of some of the intellectual and advocacy aspects of investigating cases and presenting
evidence can be improved we make suggestions on what works and what needs to be improved. Emphasis is
put upon investigative skills rather than investigative tools (the reverse of student expectations), marshalling
and presentation of facts and the growing importance of triage approaches.
We conclude with asset of recommendations on how digital forensics courses can be improved in terms of
delivery, managing student expectations and aspirations, and increased engagement with industry and
commerce.
Keywords
digital forensics, information security, expectations, employment
1. INTRODUCTION
The University of Derby has been running Digital Forensics and Information Security courses at both
undergraduate and postgraduate levels for more than 4 years. Digital Forensics has become increasingly
popular as a destination course for many computing students. Perhaps as a result thereof, the number of UK
institutions offering this as a bachelor’s degree subject has increased dramatically, at the time of writing there
were 75 undergraduate courses listed on UCAS.
An important underlying assumption is made within this study that digital forensics is a multi-disciplinary field
which encompasses areas of computing, law and science. This view is consistent with that presented by
Kessler and Schirling (2006) but somewhat contradictory to that presented by Gottschalk, Liu, Dathan,
Fitzgerald and Stein (2005) who suggest that it is in fact a subfield of forensic science. A fuller exploration of
these views is outside of the remit of this article but is nevertheless an important area for further research as
the set of key professional and technical skills involved in either area of study can differ quite remarkably.
The course at Derby includes three modules related to criminology and IT law (delivered by colleagues from
the School of Law), two modules based around forensic science (delivered by academics from the School of
Sciences) and has a significant emphasis on Information Security. It is not uncommon for digital forensic
courses to offer modules related to Information Security, this means that students undertaking such courses
are well placed to pursue opportunities in the Information Security Industry. A combination of the subject
areas described here is likely to lead to the emergence of a number of pedagogic issues some of which have
previously been identified by Yasinsac et al (2003).
The problem of students accepting and recognising the justification for certain elements of the curriculum to
exist within their chosen degree is one that has been faced by many institutions and courses. For instance,
students on the undergraduate Computer Networks course at Derby have in the past questioned the
existence of even the most introductory programming modules. This does not seem to have been a problem
for the digital forensics course in Derby. Students have not generally questioned the existence of law and
even forensic science within the digital forensics curriculum. The induction process (the point at which
curricular justification is presented) has in this respect served the course well.
• the perceptions and skill levels of students who come onto the course
• the manner in which we deconstruct and manage those perceptions and skills
• the complexities of teaching a diverse range of skills.
2. OBSERVED PROBLEMS
Some of the problems relating to student expectations may be linked to what the term computer forensics
conveys and means to students. Other relevant terms to describe the sub-discipline are digital forensics (the
term preferred by the authors), network forensics and forensic computing. Some of these terms have been
explored by Yasinsac et al (2003) and are not the subject of discussion here.
• Increased University graduate competition. An increasing number of Universities are offering courses in this
area and there is likely to be a corresponding increase in competition for employment in this sector. There is
probably more opportunity in the Information Security sector. However, students generally do not recognise this.
• Career progression and pay. Graduates entering the digital forensics market are likely to be paid less than those
entering the Information Security sector. A student who completed his master’s degree in Forensic Computing and
Security in 2010 recently reported that the only suitable job he was able to find was as a “forensic technician” at a
starting salary of £14,000 per year. He is currently widening his job search beyond the UK. It is not clear that this is
a typical experience partly because we are yet to have students graduate from our undergraduate course. Our
postgraduate students have reported that the job market is difficult to negotiate. It is however clear that the job
market currently shows no signs of being able to absorb forensics specialists in the numbers graduating from UK
degree courses. This situation is likely to have a massive effect on the employability of our students and should
impact upon the way in which we design and market our courses. Some further discussion of the jobs market can be
found in section four.
• Unrealistic Understanding of Job Roles. Whilst courses such as the one offered at Derby interface well with
employers and endeavour to provide students with a realistic understanding of the digital forensics industry,
students do nevertheless have an unrealistic understanding of what is actually involved in working in the sector. For
instance, students have often not thought through realistically in terms of the potential emotional impact of the
material that they will be handling on a daily basis.
Students may well have developed an interest in digital forensics based on portrayals in the media (e.g.
C.S.I.). Not only is the reality of the job less glamorous but students are often surprised at how painstaking
and time consuming real investigative processes are. Searches which appear to yield results instantaneously
in media portrayals can take hours or days in real cases. This isn’t helped by experiences on digital forensics
courses wherein the digital forensic images that students have to investigate are often deliberately small so as
to ensure that tutorial outcomes are achievable, if unaccompanied by appropriate disclaimers, this can
contribute to enforcing unrealistic views.
Students coming onto the course generally expect and aspire to careers within a digital forensics technician
role. This manifested itself more clearly at Derby when the majority of the first year two cohort seemed
• Students have had particular problems with developing the ability to debate and present strong conclusive legal
arguments. Some seem fearful of the need to stand up as a witness and defend their findings.
• Some students find the necessity to develop unambiguous and clear reports which defend their arguments, quite
challenging.
We have attempted to deal with some of the issues described above by: greater industrial involvement,
assessments designed to focus on industrial issues and a deliberate and focussed deconstruction of
preconceptions. It is probably too early to present conclusive arguments and responses to some of the
problems described herein, in fact there may not be conclusive approaches to resolving these. The responses
can therefore be best described as ‘work in progress’.
For any academic discipline that has been glamorised somewhat by media portrayal, it is important that a
process of deconstruction takes place early in the delivery of the course.
Following the experience of the first cohort, the course team undertook to make a deliberate attempt to
deconstruct these views. This was done primarily through industrial engagement and industry focussed
assessment, however the team also considered it important to address these preconceptions through the
induction process at the outset of the cohort experience. There is some focus during the induction process
aimed at not only developing a realistic understanding of the industry but also alerting students to
opportunities available in the areas of Information Security and also in computing generally.
At least one guest lecture is organised to be delivered by a director or practitioner from within the forensics
industry, this is an approach that other academic practitioners within this discipline have also adopted (Troell,
Pan and Stackpole 2004). The focus of the talk is to emphasise the nature of work that a forensic technician
may be involved in. The talks thus far have emphasised that:
• There is a growing demand for digital forensic practitioners, however there is also a disproportionate increase in
courses focussing on digital forensics.
• Digital forensic practitioners may be required to conduct investigations involving contraband material which
includes indecent images of children, bestiality and violence. Whilst larger organisations may be better prepared to
managing resources so that certain staff are not exposed to this, smaller organisations have less flexibility in this
respect. There is also a duty of care on the part of the employers to provide appropriate counselling and access to
psychological services for staff exposed to extreme materials. Arguably University courses in digital forensics
should also incorporate some techniques on maintaining mental well-being when dealing with such evidence.
• The digital forensics industry is new, it is undergoing growth and developing in its maturity. Career structures and
progression pathways are not as clearly defined as they may be in other industries.
A common problem in computing related courses is that of teaching communication and professional skills
(CPS). This issue has been addressed at Derby through the tighter integration of CPS within the curriculum.
For instance digital forensics is used as the vehicle for all CPS.
The CPS assessment in this case involves the defence of a digital forensics related business proposal under
heavy and strict scrutiny by course team members and industrial invitees. The presentation portion of this
assessment places the students under heavy pressure.
Amongst other tasks, the students must:
As noted earlier there is are questions about the level and type of job that will be available to our graduates.
One of the problems we face is tracking and quantifying the kinds of jobs that might be suitable. Very few
resources exist which quantify the opportunities. An examination of jobs advertised on ITJobsWatch (2010)
shows 1100 Information Security jobs were advertised in the three months to 11thh November 2010 with an
average salary of £52,500. Jobs in this sector ranked 230 compared to 198 in the same period in 2009 and
256 in 2008. By comparison average wages for Computer Forensics specialists were lower and there were
only 31 posts advertised in the three month period to November 11th. 30th. Job ranking within IT was 784
compared to 625 in the same period last year. The number of jobs for Forensic Analysts in IT in the same
period was only 3. Further interrogation of the statistics available on this one site failed to yield any other
obvious digital forensic employment. It is unlikely that our graduates are starting on the salaries quoted on
this site yet it is unclear where else they are being employed. Irrespective of the question of how much lower
entry level salaries must be, it is abundantly clear that advertised jobs in digital forensics are few whereas the
demand for information security professionals is much higher.
A search for “digital forensics”, “digital analyst” and related terms turns up a number of individual
advertisements but they are limited in number compared to other IT sector opportunities. An example from
www.jobsite.co.uk yields an advert for a “computer forensic analyst” with a salary range of £25-£30K, the
skills requirements include case management and forensic analysis of computer-based equipment and mobile
phones, securing and preserving digital evidence, EnCase, familiarity with ACPO guidelines and
“unimpeachable integrity.” However, this is a standalone job with no evidence of multiple job opportunities.
Similarly, adverts can be found for “mobile phone examiner” and “cell site analyst”, for example, but they are
few and far between.
Much more work needs to be done to understand the demands of the marketplace and we are currently
undertaking a survey of companies working in the forensics area to determine how they recruited existing staff
and what their likely demands are for the future. There is an uncomfortable sense that we may be producing
too many graduates with the wrong mix of skills to match employment demands.
5. CONCLUSIONS
Universities develop and promote courses based on student demand and the proliferation of digital forensics
courses and application figures bear testament to the fact that demand is still high. In contrast demand for
employers for digital forensic practitioners is demonstrably low at present although the current political climate
may indicate demand is likely to rise. The authors believe that we have an ethical responsibility to ensure that
student expectations are as closely aligned with the reality of the job market as we can achieve. Current
employment trends indicate therefore that our courses should be equipping students with broad skills as
information security professionals rather than just as digital forensics experts.
Assuming our initial assumptions about the job market are correct and that our observations of student
capabilities are representative we make the following recommendations:
• Digital Forensics courses should ensure their curricula include information security skills and course titles
should reflect this
• The recruitment and induction processes are vital in managing student expectations
• Career planning should start in year one
6. REFERENCES
Brinson A., Robinson A.,Rodgers M., (2006), “A cyber forensics ontology: Creating a new approach to
studying cyber forensics”, Digital Investigation, Vol 3s, S37-S43
Carlin A., Curl S., and Manson D., (2005). “To catch a thief: Computer forensics in the classroom”. In
Proceedings of the 22nd Annual Information Systems Educators Conference (Columbus, OH, Oct.),
Association of Information Technology Professionals, Chicago, IL.
Gottschalk L., Jigang, L., Dathan B., Fitzgerald S., and Stein M., (2005), “Computer Forensics Programs in
Higher Education: A Preliminary Study”, SIGCSE'05, February 23-27, 2005
HM Government, Securing Britain in an Age of Uncertainty: The Strategic Defence and Security Review,
2010, The Stationery Office.
ITJobsWatch. 2010. Information Security Jobs. [Online] (Updated 30th June 2010) Available at:
http://www.itjobswatch.co.uk/jobs/uk/information%20security.do [Accessed 11th November 2010]
Kessler G.C., and Schirling M.E. (2006). “The Design of an Undergraduate Degree Program in Computer &
Digital Forensics”. Journal of Digital Forensics, Security and Law, 1(3), 37-50.
McGuire T., Murff K., (2006), “Issues in the Development of a Digital Forensics Curriculum”, Consortium for
Computing Sciences in Colleges
Parsonage, H. 2009 Computer Forensics Case Assessment and Triage. [Online] (Updated November 2009)
Available at
http://computerforensics.parsonage.co.uk/triage/ComputerForensicsCaseAssessmentAndTriageDiscussionPa
per.pdf [Accessed 11th November 2010]
Rogers M., et al(2006) “Computer Forensics Field Triage Process Model.” Proceedings of Conference on
Digital Forensics, Security and Law pp27 -40
Smith F., Bace R., 2003. A guide to Forensic Testimony, Boston Addison Wesley
Troell L., Pan Y., Stackpole B., (2004), “Forensic Course Development – One Year Later”, SIGITE’04,
October 28–30, 2004
Yasinsac A, Erbacher R.F. et.al, “Computer Forensics Education”, IEEE Security and Privacy, vol 1, no. 4, pp.
15-23, July 2003