<!DOCTYPE article PUBLIC "-//NLM//DTD JATS (Z39.96) Journal Archiving and Interchange DTD v1.0 20120330//EN" "JATS-archivearticle1.dtd">
<article xmlns:xlink="http://www.w3.org/1999/xlink">
  <front>
    <journal-meta />
    <article-meta>
      <title-group>
        <article-title>Enhancing Collaborative Sketching with Adaptation Guidelines</article-title>
      </title-group>
      <contrib-group>
        <contrib contrib-type="author">
          <string-name>ACM Classification Keywords H.</string-name>
          <xref ref-type="aff" rid="aff0">0</xref>
          <xref ref-type="aff" rid="aff1">1</xref>
        </contrib>
        <contrib contrib-type="author">
          <string-name>User Interfaces: Prototyping</string-name>
          <xref ref-type="aff" rid="aff0">0</xref>
          <xref ref-type="aff" rid="aff1">1</xref>
        </contrib>
        <aff id="aff0">
          <label>0</label>
          <institution>Author Keywords Collaborative sketching; Context-aware adaptation; Collaborative design; Prototyping</institution>
        </aff>
        <aff id="aff1">
          <label>1</label>
          <institution>Vivian Genaro Motti, Ugo Braga Sangiorgi, Jean Vanderdonckt Louvain Interaction Laboratory, Universite ́ catholique de Louvain Place des Doyens</institution>
          ,
          <addr-line>1 - B-1348 Louvain-la-Neuve</addr-line>
          ,
          <country country="BE">Belgium</country>
        </aff>
      </contrib-group>
      <abstract>
        <p>Designing interactive systems for multiple contexts of use becomes a burden when several interaction scenarios, their characteristics and constraints, must be considered. Stakeholders face then two main challenges: they are not aware of what is significantly relevant to consider from contextual information, or how to appropriately adapt the user interfaces based on this information considered. Moreover, stakeholders cannot count on existing UI editors once they usually do not provide support for context-aware adaptation. Thus, adaptation is often ignored, resulting in user interfaces that are only suitable for static and conventional contexts of use. To support the design of user interfaces that are properly adapted to their target context, this paper proposes a novel methodology to enhance sketching activities by suggesting to the end user adaptation guidelines. This work aims at raising awareness about context-aware adaptation since the early stages of the UI design.</p>
      </abstract>
    </article-meta>
  </front>
  <body>
    <sec id="sec-1">
      <title>-</title>
      <p>
        INTRODUCTION
The contexts of use in which users interact nowadays
significantly vary in terms of user’s profiles, platforms and
environments. Due to such variety, it is challenging for stakeholders
to properly consider the specificities of dynamic contexts of
use while developing user interfaces. As such, usually
adaptation is ignored, either because stakeholders are not aware of
relevant characteristics and constraints of the context, or
because they are aware but unable to correctly address related
issues and thus to provide efficient adaptations. As a result,
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.
stakeholders often adopt a “one-size-fits-all” approach,
providing UIs that are suitable for one single conventional
context of use, with such an approach, accessibility issues often
raise, hindering or even preventing the user interaction.
Given that it is a challenge for developers to be aware of
possible adaptation issues and that adaptation must be considered
since the initial stages of the development process, this work
proposes the support to context-aware adaptation within user
interface editors, aiding stakeholders during all the
development phases, but specially starting from low fidelity
prototypes. This work also aims at raising awareness for
context peculiarities and to the importance of adopting
adaptation patterns and guidelines. To do so, a novel methodology
for electronic sketching is proposed, i.e. electronic sketching
activities must be enhanced based on the target context of use.
In previous works, multi-platform sketching has been defined
as the activity of drawing with an electronic stylus in
different devices while still having the same system running on
those different devices [
        <xref ref-type="bibr" rid="ref12">12</xref>
        ]. GAMBIT has been developed
as a multi-platform collaborative tool for user interface
design. GAMBIT enables to sketch UI’s and to simulate them
on different devices. Now we propose to augment sketching
activities by guiding stakeholders through the development of
adapted UI’s based on the target context. Three main
dimensions of context have been defined:
      </p>
      <p>Multi-platform: sketches can be done on the target device
itself, enabling prototyping and testing the UI’s on the very
device it is intended to run. GAMBIT has been built with
HTML5 and Javascript being capable of running on any
device with browsing capabilities, through a browser or
embedded into a native application;
Multi-user: since the user’s profiles vary in different
aspects, as expertise level (novice, intermediate, advanced)
and impairments (visual, cognitive, motor), also different
user interface requirements and needs must be considered;
Multi-environment: for varied constraints and
characteristics, different adaptation techniques must be applied. For
instance taking into account light and stability levels in
order to provide users appropriate interactors.</p>
      <p>This paper is organized as follows: the next section motivates
this work, presents and discusses related works; then Gambit
is presented, followed by the novel proposed methodology to
support context-aware adaptation.</p>
      <p>
        MOTIVATION AND RELATED WORK
The fact that the contexts of use significantly vary challenges
stakeholders while implementing interactive systems. The
profiles of users vary according to their age, expertise levels,
preferences and needs. The platforms have different screen
dimensions, interaction modalities, performances and
capabilities. The environments have different light, noise and
stability levels. Such heterogeneity impacts the requirements
for defining GUI’s for interactive systems, and the interaction
may be even prevented if certain characteristics are not
appropriately taken into account. Although some UI’s are already
adapted to a given context, it is not scalable for designers to
take each contextual detail into account in a way that several
different versions of the same system are implemented.
To support context-awareness during sketching activities we
propose the adaptation guidance to be incorporated into the
UI design, first by identifying the context of use, then by
suggesting related guidelines that can be adopted by the end user.
The existing tools that support electronic sketching are
usually not focused on context-aware adaptation. As such,
designers either ignore contextual information or explore
different alternatives in a very limited fashion. Nowadays, to
generate UIs that are usable in different contexts, designers must
be aware of the target context, and then properly know how
to apply adaptation, by following standard guidelines or
wellknown patterns. There exist already many tools available for
UI design and electronic sketching, however they currently
do not support context-awareness and guidelines. As such,
the resulting UI’s are not appropriately adapted to their
context, or several versions of the same UI must be implemented,
causing delay in the time to market, among other problems.
Concerning the UI design, sketching and prototyping, there
are several tools available in the market. MAQETTA, for
example, is a visual authoring tool for designing HTML5-based
user interfaces, it adopts a WYSIWYG approach. Because
MAQETTA is also a web-based application, no plugins or
downloads are required. Although it permits the design of
UI mockups, context-aware adaptation is not integrated as
a feature [
        <xref ref-type="bibr" rid="ref4">4</xref>
        ]. Besides MAQETTA, a series of alternative
graphical editors have been published to support activities of
sketching and prototyping UI’s in varied fidelity levels. Four
popular tools were selected as references; they are listed and
briefly described below.
      </p>
      <p>The most popular tools that are currently used for supporting
UI Sketching and prototyping are:</p>
      <p>
        BALSAMIQ [
        <xref ref-type="bibr" rid="ref1">1</xref>
        ] for sketching interfaces rapidly, and
communicate design ideas (www.balsamiq.com);
JUSTINMIND [
        <xref ref-type="bibr" rid="ref3">3</xref>
        ] a platform for defining prototypes for
web and mobile applications (www.justinmind.com);
MAQETTA [
        <xref ref-type="bibr" rid="ref4">4</xref>
        ] a visual authoring of HTML5 user
interfaces in the browser, open source, WYSIWYG
(maqetta.org);
SKETCHFLOW [
        <xref ref-type="bibr" rid="ref5">5</xref>
        ] a UI prototyping tool to create
interactive prototypes1.
1http://www.microsoft.com/silverlight/sketchflow/
The main shortcomings of such tools regarding context-aware
adaptation of user interfaces, can be summarized as:
Firstly, such tools are already not adapted, being implemented
to run in a conventional context of use of an able-bodied user
in a Desktop PC in a stable environment (e.g. office).
Although practitioners may be more familiar working in their
traditional workspace, because the design occurs outside the
target context, the results may be inaccurate. For instance, by
designing mobile applications in a smartphone, a more
realistic result may be achieved, specially concerning constraints
as screen dimensions and performance.
      </p>
      <p>Second, contextual information is not supported, or it is but
limited e.g. to screen sizes. Stakeholders are not aware of
which information to consider and how to consider. No
support is provided concerning usability guidelines and patterns.
Third, by being inflexible such tools force users to sketch and
to prototype one version of the same application dedicated
to each context. When various contexts must be covered it
becomes not scalable to properly ensure adaptation.
Finally, although users work usually in ’large’ teams, with
different profiles, such tools not always support
collaboration, thus either all stakeholders must be co-located or they
cannot simultaneously access and edit the project, resulting
in a pipeline of tasks. By matching different skills and
expertise simultaneously collaborating within a project, better
results can be obtained.</p>
      <p>
        Besides this, when the context is considered just during late
development phases, if the resulting UI’s are not satisfactory,
significant re-work may be needed. When usability tests and
experiments are conducted during later stages, results may be
incorrect once a high fidelity prototype is already available,
discouraging users to negatively evaluate the resulting UI’s.
ENHANCING SKETCHING WITH ADAPTATION
Sketching is considered to be a powerful tool for designing
UI’s. As the findings of [
        <xref ref-type="bibr" rid="ref8">8</xref>
        ] point out, the presence of
ambiguity in early stages of design broads the spectrum of
considered solutions and tends to deliver a design of higher quality.
Fostering creativity is considered to be important since
design is essentially a problem of wicked nature, i.e. the process
of solving it is identical with the process of understanding it
[
        <xref ref-type="bibr" rid="ref11">11</xref>
        ]. In wicked problems, in initial phases the designer does
not have a clear understanding of what to produce and has
only a vague goal in mind, therefore in order to support the
design activity with an interactive system, it is important that
the system does not get “in the way” between the designer
and the solution to be developed.
      </p>
      <p>
        GAMBIT, as Figure 1 illustrates, offers users a flexible setup
where they can choose their preferred devices for interacting
– i.e. devices that are more suitable for sketching input can be
used for pen interaction, while large displays can be used for
visualization. By observing UI design sessions on belgian IT
companies, three main sequential phases have been identified:
1. Sketching: one or more participants produce sketches to
express ideas.
2. Sharing: the participants normally share the drawings
using a big sheet of paper and use post-its. The sheets are
arranged as a storyboard on a wall for discussion.
3. Discussing: the participants refined the sketches based on
what was discussed and learned on the discussion.
GAMBIT provides a large (infinite) workspace where users
can add pictures and sketch. Parts of the workspace can be
shared among different platforms, composing a virtual
meeting room. The tool is a fundamental part of a research on
sketching user interfaces, its usage in current design practices
must also take into account the diversity of contexts of use.
In this sense, a designer can test an interface for a system on
the same device the system is intended to be used. GAMBIT’s
basic requirements list was presented in [
        <xref ref-type="bibr" rid="ref12">12</xref>
        ].
      </p>
      <p>Aiming to enhance the sketching activity currently supported
towards effective consideration of context-awareness, we
propose to augment it by integrating adaptation guidelines. To do
so, first the end user must provide the description of the target
context, then while sketching the tool will recognize the
layout of the UI elements and also propose relevant adaptation
techniques according to the UI definitions.</p>
      <p>
        Methodology and Design Decisions
For the context of use, it can be provided to the system by
means of an XML-based document, containing data about
the: user, platform and environment. Once parsered such
information will be used later on to search for corresponding
guidelines. For the sketch recognition, some algorithms have
already been investigated in previous works. UsiSketch [
        <xref ref-type="bibr" rid="ref7">7</xref>
        ],
for instance, presents an algorithm that automatically
recognizes sketches for graphical user interfaces. And
concerning the guidelines, an online repository as W3C guidelines
[
        <xref ref-type="bibr" rid="ref6">6</xref>
        ] can be accessed, a local repository integrating different
sources, or third-party web services on the cloud could be
connected to the system to provide the corresponding list
(based on the target context and the UI elements).
For the UI guidelines, three distinct dimensions of (graphical)
user interfaces can be considered, namely:
      </p>
      <p>
        Navigation: an interactive system is usually structured by
means of a task tree (being CTT [
        <xref ref-type="bibr" rid="ref2">2</xref>
        ] and HTA often
employed). Although one single task tree can enable the same
goals to be achieved in different contexts of use, there are
several groupings and hierarchies that can be adopted to
better accommodate tasks according to different scenarios.
For example concerning the screen dimension of the user,
if it is a small screen (smartphone) a reduced number of
tasks must be presented at a time.
      </p>
      <p>Presentation: the layout of graphical user interfaces can be
presented in different ways, the distribution of the contents
can follow a certain alignment, balance, or specific certain
re-molding actions.</p>
      <p>Contents: several formats of contents can be included in
the UI composition. Usually images, texts, videos and
UI interactors combined provide a task for the user.
Depending on the context, such contents can be: removed,
re-sized, replaced, etc.</p>
      <p>
        Such dimensions provide an overview about possible
adaptations, for a more complete list of techniques other references
can be accessed (e.g. [
        <xref ref-type="bibr" rid="ref10">10</xref>
        ]).
      </p>
      <p>According to the elements recognized in the UI, and the
context provided, the system can then automatically search for
and present relevant guidelines for the end user. A
semiautomatic approach can be applied to suggest changes for the
sketched UI, i.e. once the guidelines’ list is presented, the
end user is responsible for analyzing it, judging whether it is
valid, modifying the UI, and changing the guideline’s status
(e.g. as checked).
Given that design sessions often concern a specific
application domain, the suggested guidelines must cover a generic
purpose, following specific constraints of context of use and
UI elements, but still being domain-independent. Specific
requirements of the given methodology include:</p>
      <p>
        Context as an input: users must provide information about
the target context of use
UI Sketching: users design the UI defining its layout
Pattern Recognition: the system recognizes UI elements
Guidelines Suggestions: based on the element previously
recognized the system search for, retrieve and present
guidelines that match both context of use and UI element
User Feedback: the user provides his/her feedback about
the system suggestion, accepting, rejecting, evaluating
Figure 2 structures such steps as a sequence diagram. The
target context (message 1) is provided once, in the start of the
design session, sketching (2), recognition (2.1) and guidelines
(3 and 4) are repeated until the sketch is concluded.
Scenario
The case study consists in designing graphical user interfaces
for elderly users. It has already been proved that older adults
perform better and prefer touch screen interfaces. Actually,
many studies conducted in this domain, provided valuable
insight concerning guidelines for interaction, e.g. [
        <xref ref-type="bibr" rid="ref9">9</xref>
        ]. In our
study 3 GUI’s are needed to accomplish a user task: the user
is searching for a movie. First 3 main categories are presented
(e.g. drama, sci-fi, comedy), when one of those is selected a
list of movies is presented. Each movie contains an image and
a short description. For further information, the user needs to
click in a link available below the description. To navigate
through the list of movies, the GUI can be scrolled. Figure 3
illustrates this scenario. Observe that the scroll bar has been
detected as a potential issue for the UI. So the guideline points
it to the user, and suggests a new dimension or a new element
for the user, improving accessibility.
CONCLUSION
Many editors support UI design, but they often exclude UI
guidelines. Due to the technological and contextual
evolution, designers are unaware of which and how context
information to consider. Thus, most existing UI’s ignore
adaptation and target one context of use, while contexts vary and
evolve. We claim that by incorporating UI guidelines in UI
editors, stakeholders can enhance sketching, by being aware
of relevant principles, verifying and improving UI’s. This
paper proposes enhanced sketching activities, by means of a
novel approach, and illustrates the proposal with a case study.
ACKNOWLEGDEMENTS
Projects: QualIHM (Re´gion Wallonne, Dir. ge´ne´rale
ope´rationnelle de l’Economie, de l’Emploi et de la Recherche
DGO6) and Serenoa (E.C. FP7 (FP7-ICT-2009-5)).
      </p>
    </sec>
  </body>
  <back>
    <ref-list>
      <ref id="ref1">
        <mixed-citation>
          1.
          <string-name>
            <given-names>Balsamiq</given-names>
            <surname>Mockups: Rapid Wireframing Tool</surname>
          </string-name>
          ,
          <year>2012</year>
          . www.balsamiq.com.
        </mixed-citation>
      </ref>
      <ref id="ref2">
        <mixed-citation>
          2.
          <string-name>
            <given-names>Concur</given-names>
            <surname>Task Trees (CTT)</surname>
          </string-name>
          ,
          <year>2012</year>
          . http://www.w3.org/
          <year>2012</year>
          /02/ctt/.
        </mixed-citation>
      </ref>
      <ref id="ref3">
        <mixed-citation>
          3. JustInMind:
          <article-title>a platform to define web and mobile apps</article-title>
          .,
          <year>2012</year>
          . http://www.justinmind.com/.
        </mixed-citation>
      </ref>
      <ref id="ref4">
        <mixed-citation>
          4. Maqetta:
          <article-title>Visual Authoring of HTML5 User Interfaces in the browser</article-title>
          .,
          <year>2012</year>
          . http://dojofoundation.org/projects/maqetta.
        </mixed-citation>
      </ref>
      <ref id="ref5">
        <mixed-citation>
          5.
          <string-name>
            <surname>SketchFlow: Expression Studio Ultimate</surname>
          </string-name>
          ,
          <year>2012</year>
          . http: //www.microsoft.com/silverlight/sketchflow/.
        </mixed-citation>
      </ref>
      <ref id="ref6">
        <mixed-citation>
          6.
          <string-name>
            <given-names>Web</given-names>
            <surname>Content Accessibility</surname>
          </string-name>
          <article-title>Guidelines (WCAG) 2</article-title>
          .0,
          <year>2018</year>
          . http://www.w3.org/TR/WCAG/.
        </mixed-citation>
      </ref>
      <ref id="ref7">
        <mixed-citation>
          7.
          <string-name>
            <surname>Coyette</surname>
            ,
            <given-names>A.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Schimke</surname>
            ,
            <given-names>S.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Vanderdonckt</surname>
            ,
            <given-names>J.</given-names>
          </string-name>
          , and
          <string-name>
            <surname>Vielhauer</surname>
            ,
            <given-names>C.</given-names>
          </string-name>
          <article-title>Trainable sketch recognizer for graphical user interface design</article-title>
          .
          <source>In INTERACT (1)</source>
          (
          <year>2007</year>
          ),
          <fpage>124</fpage>
          -
          <lpage>135</lpage>
          .
        </mixed-citation>
      </ref>
      <ref id="ref8">
        <mixed-citation>
          8.
          <string-name>
            <surname>Goel</surname>
          </string-name>
          , V. ”
          <article-title>Ill-Structured Representations” for Ill-Structured Problems</article-title>
          .
          <source>In Proceedings of the Fourteenth Annual Conference of the Cognitive Science Society</source>
          , vol.
          <volume>14</volume>
          ,
          <string-name>
            <surname>Lawrence Erlbaum</surname>
          </string-name>
          (
          <year>1992</year>
          ),
          <fpage>130</fpage>
          -
          <lpage>135</lpage>
          .
        </mixed-citation>
      </ref>
      <ref id="ref9">
        <mixed-citation>
          9.
          <string-name>
            <surname>Kobayashi</surname>
            ,
            <given-names>M.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Hiyama</surname>
            ,
            <given-names>A.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Miura</surname>
            ,
            <given-names>T.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Asakawa</surname>
            ,
            <given-names>C.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Hirose</surname>
            ,
            <given-names>M.</given-names>
          </string-name>
          , and
          <string-name>
            <surname>Ifukube</surname>
            ,
            <given-names>T.</given-names>
          </string-name>
          <article-title>Elderly user evaluation of mobile touchscreen interactions</article-title>
          . In
          <string-name>
            <surname>Human-Computer</surname>
            <given-names>Interaction-INTERACT</given-names>
          </string-name>
          <year>2011</year>
          . Springer,
          <year>2011</year>
          ,
          <fpage>83</fpage>
          -
          <lpage>99</lpage>
          .
        </mixed-citation>
      </ref>
      <ref id="ref10">
        <mixed-citation>
          10.
          <string-name>
            <surname>Motti</surname>
            ,
            <given-names>V. G.</given-names>
          </string-name>
          , and
          <string-name>
            <surname>Vanderdonckt</surname>
            ,
            <given-names>J.</given-names>
          </string-name>
          <article-title>A Computational Framework for Context-aware Adaptation of User Interfaces</article-title>
          .
          <source>In Proceedings of the Research Challenges in Information Sciences Conference on - RCIS '13</source>
          (
          <year>2013</year>
          ),
          <fpage>12</fpage>
          .
        </mixed-citation>
      </ref>
      <ref id="ref11">
        <mixed-citation>
          11.
          <string-name>
            <surname>Rittel</surname>
            ,
            <given-names>H.</given-names>
          </string-name>
          <article-title>Dilemmas in a general theory of planning</article-title>
          .
          <source>Policy sciences 4</source>
          ,
          <issue>2</issue>
          (
          <year>1973</year>
          ),
          <fpage>155</fpage>
          -
          <lpage>169</lpage>
          .
        </mixed-citation>
      </ref>
      <ref id="ref12">
        <mixed-citation>
          12.
          <string-name>
            <surname>Sangiorgi</surname>
          </string-name>
          , U. B.,
          <string-name>
            <surname>Beuvens</surname>
            ,
            <given-names>F.</given-names>
          </string-name>
          , and
          <string-name>
            <surname>Vanderdonckt</surname>
            ,
            <given-names>J.</given-names>
          </string-name>
          <article-title>User Interface Design by Collaborative Sketching</article-title>
          .
          <source>In Proceedings of the Designing Interactive Systems Conference - DIS '12</source>
          ,
          <string-name>
            <given-names>ACM</given-names>
            <surname>Press (Newcastle Upon Tyne</surname>
          </string-name>
          , United Kingdom,
          <year>2012</year>
          ),
          <fpage>378</fpage>
          .
        </mixed-citation>
      </ref>
    </ref-list>
  </back>
</article>