Welcome to Scribd, the world's digital library. Read, publish, and share books and documents. See more
Download
Standard view
Full view
of .
Look up keyword
Like this
1Activity
0 of .
Results for:
No results containing your search query
P. 1
2_358

2_358

Ratings: (0)|Views: 5|Likes:

More info:

Copyright:Attribution Non-commercial

Availability:

Read on Scribd mobile: iPhone, iPad and Android.
download as PDF, TXT or read online from Scribd
See more
See less

06/19/2010

pdf

text

original

 
 
TOWARDS THE IMPLEMENTATION OF TEMPORAL-BASEDSOFTWARE VERSION MANAGEMENT AT UNIVERSITI DARUL IMANMALAYSIA
M Nordin A Rahman, Azrul Amri Jamal and W Dagang W Ali
Faculty of InformaticsUniversiti Darul Iman Malaysia, KUSZA Campus21300 K Terengganu, Malaysiamohdnabd@udm.edu.my, azrulamri@udm.edu.my, wan@udm.edu.my
ABSTRACT
Integrated software is very important for the university to manage day-to-day operations. This integrated software isgoing through evolution process when changes are requested by the users and finally the new versions are created.Software version management is the process of identifying and keeping track of different versions of software.Complexity level of this process would become complicated should software was distributed in many places. Thispaper presents a temporal-based software version management model. The model is purposely implemented formanaging software versions in Information Technology Centre, Universiti Darul Iman Malaysia. Temporal elementssuch as valid time and transaction time are the main attributes considered, to be inserted into the software versionmanagement database. By having these two attributes, it would help the people involved in software process toorganize data and perform monitoring activities with more efficient.
Keywords:
version management, temporal database, valid time, transaction time.
1. INTRODUCTION
Software evolution is concerned with modifyingsoftware once it is delivered to a customer. Softwaremanagers must devise a systematic procedure toensure that different software versions may beretrieved when required and are not accidentallychanged. Controlling the development of differentsoftware versions can be a complex task, even for asingle author to handle. This task is likely to becomemore complex as the number of software authorsincreases, and more complex still if those softwareauthors are distributed geographically with onlylimited means of communication, such as electronicmail, to connect them.Temporal based data management has been a hottopic in the database research community since thelast couple of decades. Due to this effort, a largeinfrastructure such as data models, query languagesand index structures has been developed for themanagement of data involving time [11]. Nowadays,a number of software has adopted the concepts of temporal database management such as artificialintelligence software, geographic information systemsand robotics. Temporal management aspects of anyobjects could include:
 
The capability to detect change such as theamount of change in a specific project or objectover a certain period of time.
 
The use of data to conduct analysis of past eventse.g., the change of valid time for the project orversion due to any event.
 
To keep track of all the transactions status on theproject or object life cycle.Universiti Darul Iman Malaysia (UDM) is thefirst full university at East Cost of Malaysia located atthe state of Terengganu. It was setup on 1
st
January2006. UDM has two campus named as KUSZACampus and City Campus. Another new campusknown as Besut Campus will be operated soon. Todate, KUSZA Campus has six faculties and CityCampus has three faculties. The university also hasan Information Technology Centre (ITC-UDM) thatpurposely for developing and maintaining theuniversity information systems and informationtechnology infrastructure.In this paper, we concentrate on the modelling of a temporal-based software version management.Based on the model, a simple web-based webapplication has been developed and suggested to beused by ITC-UDM. The rest of the paper is organizedas follows: next section reviews the concept of temporal data management. Section 3 discusses onthe current techniques in software versionmanagement. Current issues in software versionmanagement at ITC-UDM are discussed in Section 4.The specifications of the proposed temporal-basedsoftware version management model are explained inSection 5. Conclusion is placed in Section 6.
Special Issue on ICIT 2009 Conference - Applied ComputingUbiCC Journal Volume 4 No. 3544
 
 
2. TEMPORAL DATA CONCEPT
To date,
transaction time
and
valid time
are the twowell-known of time that are usually considered in theliterature of temporal database management [2, 4, 6,9, 10, 11, 12]. The valid time of a database fact is thetime when the fact is true in the
miniworld 
[2, 6, 9,10]. In other words, valid time concerns theevaluation of data with respect to the applicationreality that data describe. Valid time can berepresented with single chronon identifiers (e.g.,event time-stamps), with intervals (e.g., as intervaltime-stamps), or as valid time elements, which arefinite sets of intervals [9]. Meanwhile, the transactiontime of a database fact is the time when the fact iscurrent in the database and may be retrieved [2, 6, 9,10]. This means, that the transaction time is theevaluation time of data with respect to the systemwhere data are stored. Supporting transaction time isnecessary when one would like to
roll back 
the stateof the database to a previous point in the time. [9]proposed four implicit times could be taken out fromvalid time and transaction time:
 
valid time
 – 
valid-from and valid-to
 
transaction time
 – 
transaction-start andtransaction-stopTemporal information can be classified into twodivisions; absolute temporal and relative temporal [9].Most of the research in temporal databasesconcentrated on temporal models with absolutetemporal information. To extend the scope of temporal dimension, [12] presented a model which
allows relative temporal information e.g., “event Ahappened before event B and after January 01, 2003”.
[12] suggests several temporal operators that could beused for describing the relative temporal information:{equal, before, after, meets, overlaps, starts, during,finishes, finished-by, contains, started-by, overlapped-by, met-by and after}.In various temporal research papers the theory of time-element can be divided into two categories:
intervals
and
 points
[6, 9, 11]. If 
is denoted anonempty set of time-elements and
is denoted afunction from
to
 R+
, the set of nonnegative realnumbers then:
otherwise,0d(t)i, telement,_time
 point interval
 According to this classification, the set of time-elements,
, may be expressed as T = I
P, where
 I 
 is the set of intervals and
P
is the set of points.
3. RELATED TOOLS IN SOFTWARE VERSIONMANAGEMENT
In distributed software process, a good versionmanagement combines systematic procedures andautomate tools to manage different versions in manylocations. Most of the methods of version naming usea numeric structure [5]. Identifying versions of thesystem appears to be straightforward. The firstversion and release of a system is simply called 1.0,subsequent versions are 1.1, 1.2 and so on.Meanwhile, [3] suggests that every new versionproduced should be placed in a different directory orlocation from the old version. Therefore, the versionaccessing process would be easier and effective.Besides that, should this method be implementedusing a suitable database management system, theconcept of 
lock access
could be used to prevent theoccurrence of overlapping process. Present, there aremany software evolution management tools availablein market. Selected tools are described as follows:
 
Software Release Manager 
(SRM)
 – 
SRM is afree software and supported on most UNIX andLINUX platforms. It supports the softwareversion management for distributedorganizations. In particular, SRM tracksdependency information to automate andoptimize the retrieval of systems components aswell as versions.
 
 Revision Control System
(RCS)
 – 
RCS uses theconcepts of tree structures. Each
branch
in thetree represents a variant of the version. Thesebranches will be numbered by an enteringsequence into a system database. RCS recordsdetails of any transaction made such as theauthor, date and reason for the updating.
 
Change and Configuration Control
(CCC)
 – 
 CCC is one of the complete tools for softwareconfiguration management. It provides a goodplatform for an identification, change control andstatus accounting. CCC allows a simultaneouslyworking for a same version via virtual copies.This can be merged and changes can be appliedacross configurations.
 
Software Management System
(SMS)
 – 
SMSallows all the aspects in software configurationmanagement such as version control, workspacemanagement, system modelling, derived objectmanagement, change detection in the repositoryetc. SMS possesses the desired characteristics,providing resources of version control of systemsand having a good user interface.
Special Issue on ICIT 2009 Conference - Applied ComputingUbiCC Journal Volume 4 No. 3545
 
 
4. THE SOFTWARE VERSION MANAGEMENTISSUES IN ITC-UDM
There are three divisions have been formed at ITC-UDM. These divisions and their function are asfollows:
 
Infrastructure and Application Systems (AIS)
 – 
 to develop and maintain the university software;maintain the university computer networking;
 
Technical and Services (TS)
 – 
to support themaintenance of information technologyhardware, training, multimedia services and helpdesk.
 
Administration and Procurement (AP) - tomanage the daily operation of ITC-UDM such asadministration, procurement etc.Each division is headed by a division leader andsupported by several information technology officers,assistant information technology officers andtechnicians. All the university software modules aredeveloped and maintained by AIS Division. Figure 1depicts the main software modules managed by theITC-UDM. There are over thousands source codefiles are produced by the division. Therefore, it is noteasy for the division to manage all those artefacts.
Figure 1:
University Software ModulesFrom study done by the authors, two mainweaknesses have been found in the current approachfor ITC-UDM in managing all versions of sourcecodes produced:
 
Non systematic procedure used for managingsoftware versions and it is difficult to recognizethe valid time for each version.
 
The current approach does not consider theaspect of relative temporal in representing thevalid time for each version.
 
The current approach maintains only the conceptof 
current view
version of which an existingversion will be overwritten by a new incomingversion during the process of an update.Based on the mentioned problems, we stronglybelieve that the development of temporal-basedsoftware version management tool for ITC-UDMcould gain the following benefits:
 
To support project and software managers inplanning, managing and evaluating versionmanagement.
 
Assigning timestamps (absolute and relative) toeach transaction will provide transaction-timedatabase functionality, meaning to retain allpreviously current database state and makingthem available for time-based queries.
 
To increase the effectiveness and efficiency of the collaborative software version managementprocess.
5. THE MODEL
Version control is one of the main tasks in softwareconfiguration management. For any software versionwould have its own valid time. The collection of software versions should be organized into systematicway for the purpose of retrieval efficiency and torecognize valid time of those versions. Besides theused of unique sign for the associate version, themethod of time-stamping is also needed to beembedded into the version management database.
5.1 The Temporal-Based Version ManagementSpecifications
Temporal elements involved in the model aretransaction time (tt), absolute valid time (avt) andrelative valid time (rvt) which can be denoted as, TE= {tt, avt, rvt}. Transaction time is a date-stampingand it represents a transaction when a new valid timefor a version is recorded into the application database.Absolute valid time is represent by two differentattributes known as valid-from and valid-until and italso using an approach of date-stamping. Meanwhile,relative valid time which involves a time interval, willbe represented by a combination of temporaloperators, OPERATORs = {op
1
, op
2
, op
3
, …, op
n
}and one or more defined event(s), signed as EVENTs= {event
1
, event
2
, event
3
, …, event
n
}. This model,considered only five temporal operators, hence will bedenoted as OPERATORs = {
equal, before, after,meets
,
met_by
}. Table 1 illustrates the generaldefinitions of temporal operators based on timeinterval and time points. Figure 2 shows theAcademic ModuleHuman Resource ModuleStudent Affairs ModuleFinance ModuleDepartment of Development
   U  n   i  v  e  r  s   i   t  y   S  o   f   t  w  a  r  e   M  o   d  u   l  e
 
Special Issue on ICIT 2009 Conference - Applied ComputingUbiCC Journal Volume 4 No. 3546

You're Reading a Free Preview

Download
scribd
/*********** DO NOT ALTER ANYTHING BELOW THIS LINE ! ************/ var s_code=s.t();if(s_code)document.write(s_code)//-->