You are on page 1of 5

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/318208261

Agile with Animals: Towards a Development Method

Conference Paper · September 2017


DOI: 10.1109/REW.2017.11

CITATIONS READS
10 894

2 authors:

Dirk van der Linden Anna Zamansky


Northumbria University University of Haifa
71 PUBLICATIONS   532 CITATIONS    149 PUBLICATIONS   1,246 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Privacy impact of technology for animals View project

The human side of software development View project

All content following this page was uploaded by Dirk van der Linden on 23 October 2017.

The user has requested enhancement of the downloaded file.


Agile with Animals:
Towards a Development Method
Dirk van der Linden, Anna Zamansky
Department of Information Systems, University of Haifa, Haifa, Israel
Email: {djtlinden,annazam}@is.haifa.ac.il

Abstract—The advance of modern technologies opens new human users, which call for non-verbal elicitation techniques
exciting opportunities for computer-based systems for animals. and the need to involve animal experts, making reference to a
There is an explosion of products and gadgets for animals, such large body of knowledge of animal behavior.
as wellness monitoring applications (e.g., FitBark and PetPace),
automatic food dispensers, digital enrichment, and many more. In this paper we take a broader look at the development
The emerging discipline of Animal-Computer Interaction marks cycle for non-human users, and discuss its unique, agile nature
a new era in the development for non-human users, stressing a and the need for adaptation of standard agile methodologies to
user-centric approach. the context of non-human users. We review reports on devel-
As most available elicitation techniques in RE are verbal opment processes for non-human users, and propose AWAX,
and little reusable requirements data yet exists, this calls for
developing new approaches which can be adapted to non-human a development model incorporating Agile, Welfare, Animals
users. This paper takes a step in this direction by proposing and Experts tailored for the unique needs of such development.
a development model which incorporates iterative prototyping, The model takes into account aspects of animal welfare as non-
animal welfare as a central value, and a direct involvement of functional requirements, as well as the necessity to explicitly
animal experts at all stages of the development process. rely on animal experts throughout the development process.
I. I NTRODUCTION II. D EVELOPING W ITH AND F OR A NIMALS
Animals have been active users of technology for decades, What is so unique about developing products for non-human
but with the advent of recent technologies, new exciting users, and why do we need to adapt existing methodologies?
possibilities for animals to interact with technology emerge. The literature focusing on the development processes of ACI
Samsung has recently unveiled a futuristic smart kennel for technologies reveal several aspects which we discuss below.
dogs, featuring an automatic snack dispenser, a treadmill and Agility: Rapid iterative prototyping and getting quick
a tablet [1]. At Microsoft Research Centre at the University feedback from the user seem to be the prevalent techniques
of Melbourne researchers are introducing digital enrichment to in development of ACI products. Such methods were used,
improve the welfare of captive orangutans kept at Melbourne e.g., in [22], [23], [24] for designing sensor-based canine
Zoo [2]. Gadgets and software-enhanced products such as interfaces, and in [25], [26], [15] for designing enrichment
wellness monitoring [3], smart toys [4] and automatic feeders for elephants and apes kept in captivity. Working in iterations
[5] are an emerging market with great potential. is particularly important for being able to get feedback from
The emergence of the field of Animal-Computer Interaction the animal. As pointed out by [26]: “As we did not know
(ACI) ([6], [7]) marks a new era in designing technologies what types of controls an elephant was capable of using, nor
for animals and changes the rules of the game by taking what kinds of output held interest for her, it was vital to
a user-centric approach, placing the animal in the center obtain feedback from Valli during the design process. This
of an interactive development process. Examples of animal- led us to prototype iteratively until a useful solution was
centered technologies include various systems for supporting reached.” [22] also explains iterative development by getting
the activities of working dogs and assistance dogs ([8], [9], feedback from a dog behavior: “We also had to find a way of
[10]), providing environmental and cognitive enrichment for stabilizing the stand to prevent it from being pushed back by
pets ([11], [12], [13], [14]) and zoo animals in captivity the dog and thus dispersing the pressure we wanted to record”.
([15], [16]); or technologies for conservation and other animal
research ([17], [18], [19]. Welfare Value: The ACI Manifesto [6] states improving
The above poses new and exciting challenges in user- animals’ quality of life as one of the main aims of designing
computer interaction, requirement engineering and software ACI technologies. As proposed by North, [20] “We have a
development for non-human users. These challenges have moral commitment to build only what they want and need.”
recently been discussed [20] in the context of interaction, and Many ACI works take improving animal welfare as their
stresses the need for adapting research methods in HCI to the primary objective. For example, mobile technology to improve
context of non-human users. Additionally, we have discussed the welfare of poultry,[27] digital game for stress reduction
[21] the unique challenges of requirement elicitation for non- in stay-at-home dogs,[11] and other works developing digital
enrichment for captive animals [25], [15], [2]. But even ACI
technologies which do not aim directly at improving welfare
(such as technologies for dogs with occupations), state
their commitment to maintaining the welfare of their users
([28], [24]). The ACI community identifies strengthening the
connections to animal welfare science as a primary objective
[20]. Therefore, we argue that ACI development should be
viewed as value-driven development (cf. [29], [30]) with
animal welfare as a main value.

Domain experts: guiding vs. informing Due to the specifics


of our users, ACI products cannot be developed without
relying on body of knowledge from animal behavior, ethology Animal  expert  team   Development  team  

and biology. As explained in [20]: “From the perspective of


behavioral science, it is important to take full account of the
Animals  
ecological niche and innate behavioral tendencies, perceptual
abilities, and social needs of the species in question, and the Fig. 1. Diagram showing the typical cycle of an AWAX agile development
impact that past human interactions (and other experiences) and the involvement of relevant stakeholder (teams) in each phase.
may have on an individual animal. In addition, the size and
shape (anatomy) of the species and the age of the individual
animal must be taken into account. ” According to [20], “pre- for the animal, representing their interests to the best of their
senting an animal with a task that it cannot complete because ability. Tasks placed on the dotted line are collaboratively
it does not have the physical and/or perceptual capability will executed, as can be seen from the stakeholder involvement.
lead to inaccurate conclusions and will be detrimental from an Fig. 2 focuses on detailing the tasks in the actual itera-
animal welfare perspective. ACI needs to consider apparatus tions of design, develop, and test & review. When starting
and experimental design from the animals point of view.”. a project, an initial specification is typically derived from
Some examples of active involvement of animal experts in requirements imposed by business or research stakeholders.
the development process of ACI artifacts are [24], [25], [28], This tends to involve little a priori requirements specific to the
[11]. Therefore, animal experts take a much more pivotal animal yet, because very few re-usable existing requirements
role in development processes than usual domain experts – fragments and patterns yet exist. This is further complicated
functioning as a sort of surrogate stakeholder for the animal. because requirements tend to differ between species (e.g.,
While the latter inform the design process, the former rather dogs, cats, elephants), and may be additionally fragmented
guide it, and any development model should make explicit the between breeds with distinct physiological characteristics and
points at which their expertise affects design decisions. intelligence levels (e.g., poodles vs. bulldogs).
The unique combination of considerations identified above Figure 2 also captures the explicit points at which welfare
provide the basis for our proposed development model, which considerations are reviewed by the animal expert team. Wel-
is described next. fare issues noted at this point, together with other issues of
usability (e.g., determining more effective ways for the animal
III. AWAX: A D EVELOPMENT M ODEL FOR ACI to interact, to make them want to use the product) yield just-
in-time requirements for the current iteration. When testing a
In this section we propose a development model called
build of the product, interpretation of the animal’s interaction
AWAX1 for ACI technologies, which takes into account the
with the build will likely lead to unforeseen issues in terms of
iterative nature of development, animal welfare as guiding
usability , or unforeseen issues impacting their welfare (e.g.,
value, and an active involvement of animal experts team. Let us
unwittingly causing undue stress for the animal, or unforeseen
further examine the way in which these components, identified
safety issues - see the discussion below). This yields ‘just-too-
in the ACI literature and discussed above, are manifested in
late’ requirements, which can contribute to the next iterations.
the proposed model.
The agile nature of development is captured on Fig. 1, which The main difference between AWAX and other agile de-
also shows the involvement of the different stakeholders in the velopment models, is in the extremely rigid limitations the
process. The envisioned collaboration between animal experts intended user places on all phases of the development process.
and the development team is captured on Fig. 2: the animal First of all, only indirect ways to elicit requirements can
team guides the development process (as opposed to just be used, mainly but via animal experts hypothesizing on
informing design as usual for domain experts) at the phases possible ways an animal is likely to interact with the developed
of design, test and review, acting as a surrogate stakeholder artifact. This calls for rapid prototyping to check the proposed
hypotheses, which in its turn may compromise the value of
1 AWAX stands for Agility, Welfare as value and Animal eXpert involve- maintaining animal welfare – such situations may yield just-
ment. too-late requirements, which raise new welfare and usability
Development  team   Animal  expert  team   On the one hand, these just-too-late requirements lead to
valuable and indeed needed revised specifications, while on the
Start  
other hand they cause us to violate one of the core principles
of ACI: we are stuck in a catch-22 of having to violate the
Create   welfare of animals in order to guarantee it.
Review  
specifica-on  
specifica-on  
for  build   One illustrative example in this context is the report in [31]
must  address  all  fore-­‐ on the trial of radio telemetry collars to track the movements
seeable  welfare  issues  

DESIGN  
any  
of agouti birds. The weight of the collars turned out to be
no  
Revise   foreseeable   too heavy and has caused injuries and even death to some of
specifica-on    issues?  
the birds. Another example, provided in [32], is a study on
the wearability of commercial tracking systems for cats. This
Document   yes  
issues   study reveals major issues of itchiness and irritation, measured
creates  just-­‐in-­‐ by the cat scratching due to the presence of the collars. The
,me  requirement  

DEVELOP  
authors note, e.g., that “while attempting to remove the device,
Build  
(implement)   the cat compromised his balance and risked falling off the tree
perch (160 cm high).”
To further illustrate the just-in-time vs. just-too-late re-
Test   Review  test  
quirements, we consider our own experience in a project

TEST  &  REVIEW  


for developing autonomous drones for interactions with dogs
creates  just-­‐too-­‐ ([33]). In this project we were following the AWAX model in
late  requirement   any  
yes   unforeseen   the following sense: we involved a ethologist for providing first
 issues?   requirements, reviewing the design and identifying possible
 must  address  all    
welfare  issues   no  
welfare concerns. One such concern, e.g., was that to avoid
stress to the animal the drone needs to maintain a distance
of at least 60 cm from the animal. We then applied rapid
LAUNCH  

Release  
prototyping to deliver a first working version of an autonomous
drone which was flying in a fixed itinerary 60 cm above
Finish   the dog. We then moved to the testing phase to allow the
animal expert to analyze the dog behavior, in order to make
recommendations on how to modify the behavior of the drone.
Fig. 2. The AWAX (Agile, Welfare, Animal Experts) iterative development
method for development of interactive animal technology.
Out of safety considerations, in our experiments we used only
small dogs and let the drone fly on a safe distance from them.
In the testing phase five small dogs were videotaped in a room,
issues which can only be taken care of at the next iteration. The interacting with the drone. One participant, however, exhibited
way to maximize just-in-time requirements (while minimizing behavior which was not anticipated by the animal expert: while
just-too-late ones) is by the explicit involvement of animal reaching out to the drone, it tried jumping towards it; combined
expert team, as shown in Figure 2. with the locomotion of the drone, and its acceleration, the
dog tripped and fell backwards. This just-too-late requirement
IV. T HE W ELFARE PARADOX : J UST- IN - TIME VS . was taken care of at the next iteration, as the drone was
J UST- TOO - LATE R EQUIREMENTS reprogrammed to fly 20cm higher, so that the dogs are not
The distinction between just-in-time and just-too-late re- tempted to jump towards it.
quirements deserves some further attention in the context of One possibility to minimize just-too-late requirements is
animal welfare, which is a driving value behind most ACI via synthetic requirements techniques, such creating a species-
products. The lack of explicit techniques for requirement specific knowledge base of expert knowledge on how particular
elicitation leads to rapid iterative prototyping and getting kinds of animals interact with particular classes of technology.
feedback from the user in a trial-and-error method, which in This knowledge base could have reusable fragments and
its turn may place the animal on which the artifact is tested patterns to guide ACI developers so as to avoid just-too-late
in unsafe environment, or induce unnecessary stress, thereby requirements.
negatively affecting its welfare. The difference between just-
in-time and just-too-late requirements in the context of ACI V. C ONCLUSION
is thus that just-in-time requirements are discovered in time The rapid advance of new technologies for animals, together
to guarantee the welfare of the animals involved in the devel- with ACI maturing as a scientific discipline make this a
opment process. Just-too-late requirements – while preventing timely moment to reflect on the methodological foundations
more animals from having their welfare impacted – are learned of designing for non-human users. This creates exciting op-
after we have unwittingly violated the welfare of the animal(s) portunities for extending, refining and adapting RE techniques
involved in the development process. to new types of users.
In this paper we took a step in this direction by proposing [15] H. Wirman, “Games for/with strangers-captive orangutan (pongo pyg-
AWAX, a development model for non-human users, which maeus) touch screen play,” Antennae, 2014.
[16] M. Carter, S. Webber, and S. Sherwen, “Naturalism and aci: augmenting
incorporates iterative prototyping, is driven by the value of zoo enclosures with digital technology,” in Proceedings of the 12th
animal welfare and is guided by active involvement of animal International Conference on Advances in Computer Entertainment Tech-
experts. It is our hope this can be a starting point for a cross- nology. ACM, 2015, p. 61.
[17] M. Wikelski, R. W. Kays, N. J. Kasdin, K. Thorup, J. A. Smith, and
fertilization between the ACI and RE communities, where G. W. Swenson, “Going wild: what a global small-animal tracking
RE can provide structured ways of improving the develop- system could do for experimental biologists,” Journal of Experimental
ment processes in ACI and contribute to the systematic and Biology, vol. 210, no. 2, pp. 181–186, 2007.
[18] P. Paci, C. Mancini, and B. A. Price, “Towards a wearer-centred
verifiable generation of reusable species-specific knowledge, framework for animal biotelemetry,” in Animal Computer Interaction,
thereby opening up re-usable requirements patterns, fragments, 2016.
and methods for further development. [19] R. Barbuti, S. Chessa, A. Micheli, and R. Pucci, “Localizing tortoise
nests by neural networks,” PloS one, vol. 11, no. 3, p. e0151168, 2016.
Extending RE techniques and development models to in- [20] A. Zamansky, A. Roshier, C. Mancini, E. C. Collins, C. Hall,
clude new types of users can also push the boundaries of K. Grillaert, A. Morrison, S. North, and H. Wirman, “A report on the
the RE discipline and shed new light on existing methods by first international workshop on research methods in animal-computer
interaction,” in Proceedings of the 2017 CHI Conference on Human
lifting the assumption our user can freely speak his mind. This Factors in Computing Systems, Denver, CO, USA, May 06-11,
can also serve to include other special types of users, such as 2017, Extended Abstracts., 2017, pp. 806–815. [Online]. Available:
people with disabilities, non-verbal children, etc. http://doi.acm.org/10.1145/3027063.3052759
[21] A. Zamansky, D. van der Linden, and S. Baskin, “Pushing boundaries
of RE: Requirement elicitation for non-human users,” in Requirements
R EFERENCES Engineering Conference (RE), 2017 IEEE 25th International. IEEE,
[1] “Samsung smart kennel,” http://www.telegraph.co.uk/lifestyle/pets/ 2017.
11448890/Watch-Samsung-invent-20k-hi-tech-luxury-dog-kennel-for-crufts. [22] C. Mancini, R. Harris, B. Aengenheister, and C. Guest, “Re-centering
html. multispecies practices: a canine interface for cancer detection dogs,” in
[2] S. Webber, M. Carter, S. Sherwen, W. Smith, Z. Joukhadar, and F. Vetere, Proceedings of the 33rd Annual ACM Conference on Human Factors in
“Kinecting with orangutans: Zoo visitors’ empathetic responses to Computing Systems. ACM, 2015, pp. 2673–2682.
animals? use of interactive technology,” in Proceedings of the 2017 CHI [23] C. L. Robinson, C. Mancini, J. Van Der Linden, C. Guest, and R. Harris,
Conference on Human Factors in Computing Systems. ACM, 2017, pp. “Canine-centered interface design: supporting the work of diabetes alert
6075–6088. dogs,” in Proceedings of the 32nd annual ACM conference on Human
[3] M. Lee and M. R. Lee, “Beyond the wearable hype,” IT Professional, factors in computing systems. ACM, 2014, pp. 3757–3766.
vol. 17, no. 5, pp. 59–61, 2015. [24] M. M. Jackson, G. Valentin, L. Freil, L. Burkeen, C. Zeagler,
[4] R. J. Thomas and R. B. Bonnell, “Pet triggered programmable toy,” S. Gilliland, B. Currier, and T. Starner, “Fido–facilitating interactions for
Jan. 8 2013, uS Patent 8,347,823. dogs with occupations: wearable communication interfaces for working
[5] D.-S. Kim and S. Y. Shin, “Wireless pet dog management systems,” in dogs,” Personal and Ubiquitous Computing, vol. 19, no. 1, pp. 155–173,
Sensor Networks and Configuration. Springer, 2007, pp. 423–438. 2015.
[6] C. Mancini, “Animal-computer interaction: a manifesto,” interactions, [25] F. French, C. Mancini, and H. Sharp, “Designing interactive toys for
vol. 18, no. 4, pp. 69–73, 2011. elephants,” in Proceedings of the 2015 Annual Symposium on Computer-
[7] C. Mancini, O. Juhlin, A. D. Cheock, J. van der Linden, and Human Interaction in Play. ACM, 2015, pp. 523–528.
S. Lawson, “Animal-computer interaction (aci): pushing boundaries [26] ——, “Exploring methods for interaction design with animals: a case-
beyond’human’,” in Proceedings of the 8th Nordic Conference on study with valli,” in Proceedings of the Third International Conference
Human-Computer Interaction: Fun, Fast, Foundational. ACM, 2014, on Animal-Computer Interaction. ACM, 2016, p. 3.
pp. 833–836. [27] S. P. Lee, A. D. Cheok, T. K. S. James, G. P. L. Debra, C. W. Jie,
[8] C. Zeagler, S. Gilliland, L. Freil, T. Starner, and M. Jackson, “Going to W. Chuang, and F. Farbiz, “A mobile pet wearable computer and mixed
the dogs: towards an interactive touchscreen interface for working dogs,” reality system for human–poultry interaction through the internet,”
in Proceedings of the 27th annual ACM symposium on User interface Personal and Ubiquitous Computing, vol. 10, no. 5, pp. 301–317, 2006.
software and technology. ACM, 2014, pp. 497–507. [28] C. Zeagler, C. Byrne, G. Valentin, L. Freil, E. Kidder, J. Crouch,
[9] A. Ferworn, C. Wright, J. Tran, C. Li, and H. Choset, “Dog and snake T. Starner, and M. M. Jackson, “Search and rescue: dog and handler
marsupial cooperation for urban search and rescue deployment,” in collaboration through wearable and mobile interfaces,” in Proceedings
Safety, Security, and Rescue Robotics (SSRR), 2012 IEEE International of the Third International Conference on Animal-Computer Interaction.
Symposium on. IEEE, 2012, pp. 1–5. ACM, 2016, p. 6.
[10] A. Bozkurt, D. L. Roberts, B. L. Sherman, R. Brugarolas, S. Mealin, [29] M. A. Ferrario, W. Simm, P. Newman, S. Forshaw, and J. Whittle,
J. Majikes, P. Yang, and R. Loftin, “Toward cyber-enhanced working “Software engineering for’social good’: integrating action research,
dogs for search and rescue,” IEEE Intelligent Systems, vol. 29, no. 6, participatory design, and agile development,” in Companion Proceedings
pp. 32–39, 2014. of the 36th International Conference on Software Engineering. ACM,
[11] A. Geurtsen, M. H. Lamers, and M. J. Schaaf, “Interactive digital 2014, pp. 520–523.
gameplay can lower stress hormone levels in home alone dogsa case [30] M. A. Ferrario, W. Simm, S. Forshaw, A. Gradinar, M. T. Smith, and
for animal welfare informatics,” in International Conference on Enter- I. Smith, “Values-first se: research principles in practice,” in Proceedings
tainment Computing. Springer, 2015, pp. 238–251. of the 38th International Conference on Software Engineering Compan-
[12] A. D. Cheok, R. T. K. C. Tan, R. L. Peiris, O. N. N. Fernando, J. T. K. ion. ACM, 2016, pp. 553–562.
Soon, I. J. P. Wijesena, and J. Y. P. Sen, “Metazoa ludens: Mixed-reality [31] B. Cid, R. d. C. d. Costa, D. d. A. Balthazar, A. M. Augusto, A. S.
interaction and play for small pets and humans,” IEEE Transactions on Pires, and F. A. Fernandez, “Preventing injuries caused by radiotelemetry
Systems, Man, and Cybernetics-Part A: Systems and Humans, vol. 41, collars in reintroduced red-rumped agoutis, dasyprocta leporina (ro-
no. 5, pp. 876–891, 2011. dentia: Dasyproctidae), in atlantic forest, southeastern brazil,” Zoologia
[13] P. Pons, J. Jaen, and A. Catala, “Developing a depth-based tracking (Curitiba), vol. 30, no. 1, pp. 115–118, 2013.
system for interactive playful environments with animals,” in Proceed- [32] P. Paci, C. Mancini, and B. A. Price, “Designing for wearability in ani-
ings of the 12th International Conference on Advances in Computer mal biotelemetry,” in Proceedings of the Third International Conference
Entertainment Technology. ACM, 2015, p. 59. on Animal-Computer Interaction. ACM, 2016, p. 13.
[14] I. Hirskyj-Douglas, J. Read, and B. Cassidy, “A dog centred approach [33] A. Zamansky, “Dog-drone interactions: towards an aci perspective,” in
to the analysis of dogs’ interactions with media on tv screens,” Interna- Proceedings of the Third International Conference on Animal-Computer
tional Journal of Human-Computer Studies, vol. 98, pp. 208–220, 2017. Interaction. ACM, 2016, p. 14.

View publication stats

You might also like