Information Technology Reference
In-Depth Information
11.4.1.
Retrospective
Since its definition in 1999 [THE 99], plasticity is explored incrementally, in
three phases.
First stage
: in the first decade of the 21st Century, adaptation was studied in
terms of multi-targeting: it consisted, at the design stage, of generating different
versions of the UI for the different targeted contexts of use (typically, large screens
versus small screens). This work was stimulated by:
- development and maintenance costs induced by the production of as many UI
versions as contexts of use anticipated before the design stage; and
- the difficulty in ensuring ergonomic consistency between versions when
developments are carried out in a partitioned manner.
This sales pitch was in answer to the
varied
nature of the context of use
according to a system viewpoint: that of the designer confronted with multi-target
UI engineering [THE 01]. We label these works as the passage from
mono-targeting
to
multi-targeting.
Second,
the
variable
nature of the context of use is integrated into the research
agenda, thus making the leap from
multi-targeting
to
plasticity:
designing for
several key contexts of use identified in the phase before the design stage is not
enough. Changes in context of use must be addressed. The European CAMELEON
project (2001-2004)
2
covers these two increments (from mono-targeting to
plasticity) with a bary centre nonetheless on multi-targeting. In particular, user-
centered properties are not considered. The contexts of use and the changes in
contexts of use are identified at the early phases. As a result, UIs are prefabricated,
as for the Sedan-Bouillon demonstrator (section 11.2.2).
Third,
the
unpredictable
nature of the context of use is integrated into the
research agenda. From then on, it is no longer just a matter of perceiving the context
of use and switching to the most appropriate prefabricated UI but generating, if
necessary, an appropriate UI. The adaptation process then becomes more prominent.
It can be based on approaches from artificial intelligence, including automatic
symbolic learning techniques for decision modeling as in [HAR 08], [HAR 09]. The
following section illustrates this viewpoint in a scenario linked to the field of
transport
3
.
2 http://giove.isti.cnr.it/cameleon.html.
3 This scenario could be extended to be linked with elements of personalization, as described
in Chapter 3.
