Towards the design of intelligible object-based applications for the Web of Things

Abstract

As more and more things, sensors, appliances and devices are getting connected to the Internet, researchers of the Web of Things community have recently been exploring the use of the World Wide Web as a platform for smart objects. Such technology enables the creation of object-based applications mixing real-world objects that embed tiny Web servers with existing Web resources. To ensure the adoption of such applications, which potentially modify the behaviors of objects, it is needed that the overall system or architecture supports the mental models developed by users. In this paper, we propose a definition of object-based applications based on the literature review and present a protocol aiming at better understanding users' perception of smart environments. We then present our preliminary results and highlight the need of creating intelligible systems and tools.

Full text

Towards the Design of Intelligible Object-based
Applications for the Web of Things.
Pierrick Thébault 1, 3
1 Alcatel-Lucent Bell Labs France
Route de Villejust
91620 Nozay, France
pierrick.thebault@alcatel-
lucent.com
Dominique Decotter 1, 2
2 Université Paris Descartes, EA
LATI
71, Av Edouard Vaillant
92100 Boulogne Billancourt
dominique.decotter@alcatel-
lucent.com
Simon Richir 3
3 Arts et Métiers Paristech, LAMPA
2, Bd du Ronceray
49000 Angers, France
pi.laval@ensam.fr
ABSTRACT
As more and more things, sensors, appliances and devices are
getting connected to the Internet, researchers of the Web of
Things community have recently been exploring the use of the
World Wide Web as a platform for smart objects. Such
technology enables the creation of object-based applications
mixing real-world objects that embed tiny Web servers with
existing Web resources. To ensure the adoption of such
applications, which potentially modify the behaviors of objects, it
is needed that the overall system or architecture supports the
mental models developed by users. In this paper, we propose a
definition of object-based applications based on the literature
review and present a protocol aiming at better understanding
users’ perception of smart environments. We then present our
preliminary results and highlight the need of creating intelligible
systems and tools.
Categories and Subject Descriptors
H.5.m [Information interfaces and presentation (e.g., HCI)]:
Miscellaneous.
General Terms
Design, Experimentation, Human Factors.
Keywords
Web of Things, Object-based Applications, Physical Mashups,
Mental Models, User Perception, User Study, Intelligibility, Smart
Objects, Smart Environments.
1. INTRODUCTION
The Web-enablement of physical artifacts (e.g. tagged or visually
marked things, networked sensors, appliances and devices
embedding web servers) today opens up the possibility for small
applications to be built on top of real-world objects (RWO). By
exposing these smart objects as accessible and addressable
resources of the World Wide Web, researchers of the Web of
Things community demonstrated that popular Web technologies
(e.g. HTML, Javascript, Ajax, PHP) could be used to create
applications mixing RWO and existing services of the Web
[3,4,7,17]. Following the trends of Web Mashups (i.e. ad-hoc
applications) and participatory services [9], they promise to
provide developers and proficient users with new tools to
compose resources not only from the digital but also the physical
world. Such technologies would allow end-users to customize
their RWO, to augment their capabilities and to shape new
behaviors that will lead to the creation of smart environments.
So far, projects and initiatives have mainly focused on building
system architectures, defining semantic descriptions of RWO
capabilities or status and offering human-understandable
representations. If several mobile browsers providing users with a
new means of visualizing digital counterparts of RWO and to
interact with them have recently been presented [3,5,16], the
concept of object-based applications has just started to be
explored and is not yet clearly defined. Researchers often assume
that users will most likely want to create or pull applications on
top of their RWO and arbitrarily recombine them according to
their needs [10]. We argue that such a vision can only be achieved
if the overall system matches users’ mental models. The way
object-based applications are functioning and their impact on
RWO’s behaviors need to be intelligible for people if we want
them to use, manage and interact with smart environments.
In users’ mind is there such concept as object-based applications?
How do non-technical users perceive interconnections between
RWO and Web resources? How consistently do they consider
these applications as being part of a RWO or the environment?
Although researchers and designers may represent applications in
a principled, logical fashion, user perception of “disappearing
systems” is rarely systematic [15]. By blending in a seamless way
into user’s environment, charging RWO with additional
functionalities and making them act in a proactive way, these
applications are indeed modifying our perception of inner
systems. It is therefore time to consider the type of representations
people make without regard for technical concerns.
To ensure the user adoption of the Web of Things approach, we
propose to investigate people’s perception of interconnected
RWO and Web resources. To do so, we asked a group of users to
draw a schematic representation of a smart environment (i.e. an
ecosystem of smart objects interconnected with Web Resources)
and conducted a collaborative evaluation of their productions.
After reviewing the related work and trying to define object-based
Permission to make digital or hard copies of all or part of this work for
personal or classroom use is granted without fee provided that copies are
not made or distributed for profit or commercial advantage and that
copies bear this notice and the full citation on the first page. To copy
otherwise, or republish, to post on servers or to redistribute to lists,
requires prior specific permission and/or a fee.
WoT 2011, June 2011, San Francisco, CA, USA.
Copyright 2011 ACM 978-1-4503-0624-9/11/06…$10.00.

applications, we present in this paper the protocol and the first
results of our qualitative research. We argue that this work
contributes to Web of Things research by better defining object-
based applications and providing insights into users’ mental
models. Our results will help researchers to build systems that
support these models and designers to create intelligible user
interfaces.
2. DEFINING OBJECT-BASED
APPLICATIONS
Inspired by Kindberg’s work on “web presence” [8], the Web of
Things community managed to expose RWO as resources of the
Web and created a new type of application built on top of the real
world. In this section, we review the related work on object-based
applications and then highlight their capabilities. We finally
identify three types of applications that could be presented in
different ways to users.
2.1 Related work
Wilde first pointed out the advantage of fully integrating “things”
into the Web with a RESTful approach [17]. He tells us that Web-
exposed objects can be reused in different contexts or applications
as any accessible Web resources and facilitate the creation of
applications linked with the physical world (e.g. a backend
infrastructure to monitor and manage products of shop or
warehouse). Guinard and Trifa later refer to these applications as
“physical mashups” [6] that can be categorized in two groups:
“physical-virtual mashups” presenting a user interface (e.g. a
dashboard monitoring the energy consumption of household
appliances) and “physical-physical mashups” enabling machine-
to-machine interactions (e.g. a lamp that changes its color
depending on the energy consumption level).
In recent work, Guinard presented two “mashup editors” [5]
offering people to program their RWO according to specific rules
or events (e.g. augmenting the home temperature when a user is
coming back home). Benoit et al. describe such applications
composing RWO or Web resources that are “functionally not
compatible” as “complex” [2]. As opposed to chains of objects
based on objects’ input / output capabilities, “complex
applications” embed the computational logic that is required to
bridge any RWO with another (e.g. as illustrated on Figure 1). In
a similar approach, Thébault et al. finally use the generic term
“object-based applications” [16] to refer to any application
interconnecting a RWO with other objects or Web resources that
have been designed to support a specific task in an environment.
2.2 Capabilities of applications
Based on the related work, we propose to highlight the potential
capabilities of such applications. Developers and proficient users
can use Web of Things technologies to leverage the mechanisms
described in the following paragraphs.
Capacity to deliver Web content or media through RWO
whose output capabilities are compatible. Data stream can
whether be queried from Web resources (e.g. using the LCD
screen of an alarm clock to display Tweets) or others RWO (e.g.
redirecting a video stream from a laptop computer to a video
projector). Information can potentially be converted or
reformatted by the application to be conveyed through the chosen
RWO (e.g. a text can be read if no display is available).
Capacity to publish information related to RWO’s states on
Web Resources. Data related to RWO’s daily uses or capabilities
can be logged for personal use (e.g. generating graph from a
user’s TV watching time) or processed to be shared on other Web
resources (e.g. posting on LastFM all songs that are being played
on my hi-fi system). Publishing rules can eventually be
implemented in order to avoid user-sensitive information to be
broadcasted on social network platforms (e.g. posting a message
on Facebook only when I lose weight).
Capacity to trigger RWO’s capabilities based on RWO’s
states. RWO can be turned on/off or fine-grained controlled by
applications in order to automate the home (e.g. setting up the
heaters and shutters according to temperature and light sensors) or
stress events of the physical world (e.g. making a lamp blink with
a different color when a user receive a phone call). Commands
can also be chained and be sent after a specific user’s interaction
with a RWO (e.g. automatically reducing the light intensity and
declining calls when a user is turning his DVD player on).
Capacity to modify the behavior of RWO based on Web
resources. Multiple data streams can be queried and processed
from Web Resources in order to augment the awareness of RWO
and leverage an ambient intelligence. The inner working of
objects can be bypassed to anticipate unexpected events (e.g. a
digital video recorder can be reprogrammed if a TV show has
been rescheduled to another time slot) or warn the users of
important matters (e.g. displaying a customized alert on a TV
when a family member is sending a real-time message to a user).
2.3 Types of applications
By combining the mechanisms described in the previous part,
developers and proficient users are able to create more or less
complex object-based applications. Another way to highlight the
variety of these applications is to consider them according to their
impact on the physical world. We argue that they can whether
enable users to monitor, augment or orchestrate their RWO. We
propose three categories that are described hereafter.
Applications that monitor RWO. Most likely designed for
computer or mobile displays, these applications provide users
with a comprehensible overview of a small or large scale RWO’s
ecosystem. They facilitate the aggregation, the storage and the
presentation of RWO’s states and may offer remote control of
their capabilities.
Figure 1. A mobile “object browser” allowing users to
download object-based applications on their RWO.

Examples: social networks for “things”, smart metering tools
related to energy consumption or daily life activities (i.e. life
logging), inventory and tracking systems facilitating asset
management, etc.
Applications that augment RWO. Specifically designed for a
type or model of RWO, such applications offer users to add one or
several capabilities to an object. They enhance the inner system of
RWO by enabling a new means of interoperability with RWO and
Web resources and allow users to customize their product’s
experience.
Examples: facilitating data circulation among devices and Web
resources, adding a social dimension to RWO, enhancing RWO’s
awareness with open data, suggesting short interactions with Web
services (e.g. postponing a meeting or sending a “templated”
email from a RWO), etc.
Applications that orchestrate RWO. Involving multiple
interconnections of RWO, this type of application aims at
augmenting the environment. They allow multiple RWO of a
physical space to be automated according to a given context.
Spatiotemporal and social information, often provided by Web
resources, allow users to define events that will trigger or adapt
the behaviors of RWO.
Examples: automation of household appliances and activities (i.e.
domotics), assisted living systems for elderly people, RWO
reconfigurations for ambience sharing or communication
purposes, personalization of content delivery according to social
preferences and presence, etc.
3. INVESTIGATING USER’S MENTAL
MODELS OF OBJECT-BASED
APPLICATIONS
By mixing RWO and Web resources, object-based applications
open up the possibility for physical artifacts or environments to be
augmented and orchestrated. These applications, which modify
the behaviors of RWO, raise new issues regarding users’
understanding of complex inner systems. As researchers are
working towards the design of new tools allowing people to
create, deploy and manage object-based applications, it is needed
that their architecture supports users’ mental models. In this
section, we present the protocol and the preliminary results of an
experiment that aims at capturing the mental models of a smart
environment.
3.1 Overview
Rouse et al. [13] proposed a comprehensive definition of mental
models, which has also been quoted by Schmitt et al. in their
research on “disappearing systems” [15]: “Mental models are
mechanisms whereby humans are able to generate descriptions of
system purpose and form, explanations of system functioning and
observed system states, and predictions of future system states”.
In our research, we especially seek at better understanding users’
structural mental representations of object-based applications,
described by Preece [12] as the internal working of a system. We
are indeed interested in better knowing how users perceive the
interconnections between RWO and Web resources and envision
such applications.
As smart objects and object-based applications have not yet been
introduced in users’ life (i.e. except from Nabaztags, smart
meters, televisions or radios have not hit the French market yet),
we propose to build a protocol that will allow participants to
project themselves in a simulated reality involving well-known
RWO or Web resources. By asking them to draw a schematic
representation of a non-existing but functioning smart
environment (i.e. depicted in a provided use case scenario), we
expect to extract their mental models. Since people generally do
not develop the same mental model of a system [11], we will
focus on measuring users’ understanding of the different
representations or schemas. This will allow us to identify the
different strategies used by users.
In order to enable successful interactions with computer systems,
Sasse [14] tells us that designers and engineers need to create and
communicate a user model that can be adapted by people. It is
therefore necessary to collect users’ mental models in an early
design phase.
3.2 Participants
The sample of participants included six students between the ages
of 23 to 27 years (M=23.6, SD=1.7). All of them were undergoing
a full time post-graduate master’s degree in a French design
school where they have just started to work on smart objects,
mainly from a product design perspective. All the students
accepted to freely participate in our research program, which has
been conducted with the agreement of the school representatives.
During the same day, participants took part in a co-creation
session aiming at combining RWO and Web resources to create
new applications. They were therefore introduced to the overall
concept of object-based applications but were not aware of
research initiatives conducted in the academic field. We argue that
their limited background on smart objects and their interest in the
topic should not be considered as a bias but as a prerequisite to
apprehend such a concept. Pre-tests of this experiment showed
that “regular” users are most likely not able to project themselves
in a world where objects and Web resources are combined to
create applications.
Figure 2. Participants sketching a smart environment

3.3 Materials
Each participant was provided with a printed use case scenario
describing a smart environment. Eight well known RWO and
eight existing Web resources were arbitrarily chosen to build
object-based applications that aim at supporting users’ daily life
activities. These applications vary in their level of complexity and
rely on different types of interactions. In order to investigate
users’ perception on object-based applications, we only described
their behaviors in the use case scenario. Instructions were reported
as followed:
• Every morning, my alarm clock plays a song from my
Deezer playlist (i.e. a music on demand platform) to wake
me up.
• When I press the snooze button, a public message is
automatically published on my Facebook profile.
• The alarm time will be automatically delayed if snowfalls
have caused serious traffic jams.
• The bathroom heater and the coffee machine will be
automatically turned on before the alarm of my alarm clock
is fired.
• My multi-color lamp will turn green if I received some letter
in my mailbox.
• My digital photo frame will display Facebook pictures of
my friends when they are at my place (i.e. their position is
retrieved from Google Latitude).
• Every morning, during weekdays, my lamp blinks with a red
color to warn me that I have to leave now if I do not want to
miss the metro. The schedule is fetched from RATP’s
website (i.e. Paris public transport).
• Shutters and lights of my place will be automatically
triggered to simulate a presence when I am on holydays
somewhere else (i.e. the dates are retrieved from Google
Calendar).
A questionnaire based on a six-point Likert scale (i.e. from 1
corresponding to totally agree to 6 which indicates that they
totally disagree) was given to participants. They had to answer
these four questions:
• The schema is clearly understandable.
• I clearly see the links between RWO and Web resources.
• I understand how to activate or deactivate an object-based
application.
• The schema will help me to manage my applications.
Paper (i.e. A3/ledger size), glue, scissors, colored pens and cards
representing the RWO and Web resources mentioned in the
scenario were finally made available to users during the session.
3.4 Procedure
This experiment was divided into two parts. The first one focuses
on sketching the given smart environment and the second one on
evaluating the schemas produced by all participants.
During the first phase, participants had 45 minutes to draw the
interconnections of RWO and Web resources that were described
in the provided use case scenario. To make sure that they come up
with a personal representation of the smart environment, they
were asked to work on their own. They were notified that there
are neither good nor bad representations and that they could either
draw items or use the printed materials.
During the second phase, schemas were self-evaluated by all the
participants. Each representation was anonymously displayed on a
white board and examined by users with a questionnaire.
3.5 Results
As questionnaire criteria were evaluated in a different way, we
discriminated positive evaluations from negative evaluations. We
considered all ratings lower than 3.5 of the Likert scale to be
positively perceived and ratings higher than 3.5 to be negatively
perceived. To facilitate comparison and analysis, all positive
evaluations have been highlighted. In table 1, we present the
results that allowed us to identify the schemas that have been best
perceived globally or according to a specific dimension. All
schemas are shown on figure 3 and available for download [1].
Observations are detailed as followed.
Clarity of schemas. We observe that, apart from schema 3, all
representations of the ecosystem are considered as clearly
understandable. Schema 4 has the best score regarding the
perception of clarity.
Clarity of links. Interconnections between RWO and Web
resources are well understood in all visual representations. In
schema 4 and 5, links are nevertheless considered as the clearest.
Understanding how to activate or deactivate an application.
According to the participants’ point of view, none of the
representation clearly indicates how to control object-based
applications.
Helpfulness for application management. Schemas 1, 4, 5 and 6
received positive evaluations, but schemas 1 and 4 are considered
as more relevant to manage a collection of applications.
Global evaluation. Participants globally best perceive schemas 4
and 5. Even though they received the same average evaluation for
the link clarity, schema 4 is considered as more intelligible and
helpful for application management.
Users’ feedback tells us that the third criterion was not well
understood by participants. If most of them were able to
apprehend the interconnections of objects with Web resources, the
concept of user-managed applications that can be turned on and
off seems not to be envisioned. This can be explained by the fact
that users did not explicitly represent object-based applications in
their drawings, as discussed in the following section.
Table 1. Status for mean evaluations of schemas
Schema
Schema
clarity
Links
clarity
Activation/
Deactivation
Helpfulness for
application
management
1
3,3
2,2
3,5
2,5
2
3,3
2,7
4,2
3,5
3
3,5
2,2
3,7
3,8
4
2
1,8
4,2
2,5
5
2,2
1,8
3,7
2,8
6
2,3
2,7
4,5
3

Figure 3. Schematic representations of a smart environment involving RWO and Web resources.

4. DISCUSSION
Out of the six participants, four came up with an abstract
representation where RWO or Web resources are connected by
links. In most cases, RWO and Web resources can be considered
as nodes of small graphs whose orientation depend on the number
of interconnections they have with each other’s. By observing the
way such graphs or clusters are directed, we found out that
participants make a difference between interactive systems that
rely on a specific RWO to operate and the ones that are based on
information provided by Web resources. In schemas 1, 3 and 4,
users for example considered Google Calendar, Latitude and the
subway platform as the starting of point of a chain of interactions
whereas the alarm clock and the lamp seems to articulate others
RWO and Web resources. It confirms that people are likely to
“attach” behaviors on RWO when interconnections aim at
augmenting RWO’s capabilities and reinforcing their role in the
ecosystem. As the orchestration of RWO often relates to a
particular situation or context, other behaviors have less
opportunities of being embodied in physical artifacts and might be
considered as “floating” in the environment. Especially in schema
5, a house was drawn to convey that idea.
Although participants did not mention the term “application”
during the exercise, we can nevertheless assume that such concept
is inherent in their representations. Schema 3 is for instance
organized with three labeled clusters that support users’ daily life
activities (i.e. prepare to go to work, receive friends, go on
holidays). If applications that orchestrate RWO appear to be easier
to apprehend, those that augment RWO remain fuzzy. Participants
are indeed likely to arbitrarily decide how interactions with RWO
and Web resources can be grouped or split. For some participants
(i.e. schemas 1, 3 and 4), each interconnection adding capability
to a RWO can be interpreted as a subpart of a unique application
(e.g. the one dedicated to the alarm clock). For another one (i.e.
schema 5), each of them can be considered as ad-hoc applications
that can be “pulled” on RWO like mobile applications can be
downloaded on smart phones. This leads to small variations about
the way applications are “anchored” on RWO (e.g. which RWO,
from the mailbox and the lamp, should depend on the other?).
Even if object-based applications are not clearly perceived by
participants, it is encouraging to see that most of them were not
reluctant to merge the physical and digital worlds in their
representations. Schema 6 tells us that separating the private,
semi-private and public spheres is not a proper strategy to
represent the interactions between RWO and Web resources.
Splitting objects and Web resources into different categories make
the links between each of them very hard to retrace. Instead of
linking resources all together to create a global ecosystem (e.g.
schema 4), some participants emphasized the separation between
clusters or grouped them in a way that relates to their personal life
(i.e. the different phases of a day). In schema 5, user draws meta-
clusters (e.g. every morning, week day, absence, social) of what
can be interpreted as object-based applications. In this case, the
same RWO appear several times in different categories. It is
worth mentioning that this representation is considered by
participants as one of the easiest to understand, despite the fact
that it does not provide users with any detail on the type of
interactions (i.e. descriptions, flow charts and link directions were
not used). We argue that such temporality-based or activity-based
categorization may not only help users to have a global
understanding of a smart environment but also to manage and
navigate through their collection of applications.
5. CONCLUSION
In this paper, we presented an overview of initiatives leveraging
Web of Things technologies to create new types of applications
bridging the physical and digital worlds. To better define these
object-based applications, we highlighted their capabilities in
terms of interactions with RWO and Web resources and presented
them in three categories: applications that monitor, augment and
orchestrate RWO. While the first type is most likely to be
accessed on a computer or a mobile for management purpose, the
two others enables developers and proficient users to modify the
behavior of existing RWO, whose user interfaces have not yet
been designed for capability reconfiguration or augmentation.
Such applications have therefore an impact on users’ perception
of inner systems that may influence users’ adoption. Whether
researchers work towards creating programming tools or
designing browsers for application instantiation, it is needed that
their systems support the mental models users are likely to
develop.
In order to investigate users’ understanding of a smart
environment that could be created using Web of Things
technologies, we conducted an experiment whose protocol is
based on users’ drawings of a given set of resources and self-
evaluations of produced representations. We argued that this
procedure would help users to project themselves in a simulated
reality and help us to understand their perception of RWO and
Web resources interconnections and interactions. In a first step of
our research, we intended to capture mental models and observe if
the concept of object-based applications is underlying them. We
found out that applications can whether be “attached” to a RWO
or be “floating” in the environment. In the first case, applications
add capabilities or rely on a system event to trigger other RWO.
In the second case, interactions are triggered among several RWO
according to a spatiotemporal context often brought by a Web
resource. Although participants did not explicitly mentioned the
term “application”, the way they represented interconnections and
grouped RWO tells us that such concept make sense for them.
In future work, we plan to re-conduct our experiment with
different panels to validate our findings. We also plan to add a
qualitative evaluation to our experiment in order to better analyze
users’ production and investigate their understanding of the
object-based application concept. We expect to collect insights
that will help us to better understand how people are likely to
consider and use such applications. We argue that researchers
need to considerate users’ mental models to create user-driven
systems or architectures that will support these models. Designers
also need to leverage this knowledge to create intelligible
navigation or management tools that will help users to understand
how their smart environment is orchestrated. We expect to
contribute to the Web of Things community by providing
guidelines for object-based application creation, instantiation and
configuration by end-users.
6. REFERENCES
1. Wothings. http://www.wothings.com/publications.
2. Benoit, C., Boussard, M., Lu, M., Pastor, A., et Toubiana, V.
Web-of-things vision: things-as-a service and interaction
patterns. Bell labs technical journal, (2011).
3. Boussard, M. et Thébault, P. Navigating the Web of Things:
Visualizing and Interacting with Web-Enabled Objects.
Leveraging Applications of Formal Methods, Verification,

and Validation, (2010), 390–398.
4. Duquennoy, S., Grimaud, G., et Vandewalle, J.J. The web of
things: interconnecting devices with high usability and
performance. Embedded Software and Systems, 2009.
ICESS'09. International Conference on, (2009), 323–330.
5. Guinard, D. Mashing up Your Web-Enabled Home. Adj.
Proc. the International Conference on Web Engineering
(ICWE 2010), Vienna, Austria, (2010).
6. Guinard, D. et Trifa, V. Towards the web of things: Web
mashups for embedded devices. Workshop on Mashups,
Enterprise Mashups and Lightweight Composition on the
Web (MEM 2009), in proceedings of WWW (International
World Wide Web Conferences), Madrid, Spain, (2009).
7. Guinard, D., Trifa, V., et Wilde, E. A Resource Oriented
Architecture for the Web of Things. Proc. of IoT, (2010).
8. Kindberg, T., Barton, J., Morgan, J., et al. People, places,
things: web presence for the real world. Mob. Netw. Appl. 7,
(2002), 365–376.
9. Murugesan, S. Understanding Web 2.0. IT professional 9, 4
(2007), 34–41.
10. Newman, M.W., Sedivy, J.Z., Neuwirth, C.M., et al.
Designing for serendipity: supporting end-user configuration
of ubiquitous computing environments. Proceedings of the
4th conference on Designing interactive systems: processes,
practices, methods, and techniques, (2002), 147–156.
11. Payne, S.J. Users’ mental models: The very ideas. HCI
Models, theories and frameworks: Toward a
multidisciplinary science, (2003), 135–156.
12. Preece, J., Rogers, Y., et Sharp, H. Interaction design:
beyond human-computer interaction. (2002).
13. Rouse, W.B., Morris, N.M., et RESEARCH,
G.I.O.T.A.C.F.H.S. On looking into the black box: Prospects
and limits in the search for mental models. (1985).
14. Sasse, M.A. Eliciting and describing users’ models of
computer systems. Unpublished PhD Thesis, School of
Computer Science, The University of Birmingham, (1996).
15. Schmitt, F., Cassens, J., Kindsmüller, M.C., et Herczeg, M.
Mental Models of Disappearing Systems: Challenges for a
Better Understanding. Proceedings of the Sixth International
Workshop on Modeling and Reasoning in Context, CEUR-
WS.org (2010), 61-72.
16. Thébault, P., Boussard, M., Lu, M., Mivielle, C., et Richir,
S. EnvB: An Environment-based Mobile Browser for the
Web of Things. Workshop on interacting with smart objects,
Proceedings of the 15th international conference on
Intelligent user interfaces, (2011).
17. Wilde, E. et others. Putting things to REST. School of
Information, UC Berkeley, Tech. Rep. UCB iSchool Report
15, (2007).