SEKE 2012 Proceedings - Knowledge Systems Institute

• These characteristics are considered common in a
traditional interface, may not correspond to those
made for children. They have unique needs for their
age. Beyond these specific needs for the children users
the designers need to deal with the dangers that are
usually present in a web environment, such as
pornography and violent or racist content;
• The existence of elderly among the users should be
checked for these and needs met;
• People with special needs are another installment of
the user community of the system and need
adjustments in the system to operate in the
environment without difficulty.
For the CSLs we used the basis of the research article:
Project Tapejara from Souto [7]. The CSLs refer the subject’s
characteristic way of learning new concepts or even to
generate elaborations of prior knowledge. According to
Madeira, Wainer, Verdin, Alchieri, Diehl [11], the CSLs are:
Analogue-Analytical (AA), Concrete-Generic (CG),
Deductive-Evaluative (DA), Relational-Synthetic (RS) and
Synthetic-Evaluative (SA).
Users rated AA-style may require more time for learning
because when they were confronted with new information they
tend to get a considerable depth on the subject through intense
reflection. Users with CG-style tend to be pragmatic and
careful in their learning situation. The learning objectives,
evaluation criteria and feedback must be clear to this style
because then he can work towards the goals. The DA-style can
get to disregard a lot of concrete examples when they believe
they have already understood the logical pattern underlying
the new information. For the RS-style users are easy mental
work with pictures and they appreciate the use of charts,
diagrams and demonstrations. Efficient reading of charts and
mind maps [7].
We relate the CSLs with users with attention deficit
disorder and hyperactivity (ADD and ADHD). The people
with ADD and ADHD do not develop the scholar knowledge
as expected for their ages. The diagnostic in the scholar age is
common because in this period can be found the difficulties of
attention and remain silent as the studies of Siqueira and
Gusgel Giannetti [9], Poeta and Rosa Neto [10].
III. THE PROJECT´S GUIDELINES
The GuideExpert system was initially developed
containing in its knowledge base three hundred and twenty six
classified guidelines and 10 meta-guidelines, which is the
result of gr ouping guidelines, this allows you to search for
more targeted features of i nterface design that is being
modeled [5].
The system consists of four elements are: user interface,
expert system (inference engine and working memory),
knowledge base and database. Figure 1 shows the architecture
of the expert system which demonstrates the elements through
their organization in layers and modules.
Figure 1 – System Architecture Specialist GuideExpert
Layer 2 is the expert system, which, when executed,
accesses the layer 3 containing the knowledge base, and thus
carries the knowledge rules. The first module of the user
interface has been assessing the requirements of the HCI with
the designer, if executed. After obtaining the information, is
provided to the expert system and soon after is analyzed. The
database is accessed at layer 3, for the meta-guidelines
requested. The second module of user interface aids in
heuristic evaluation [5].
This research contributed to the tool GuideExpert
incorporating new knowledge items to it, enabling the groups
of users with different learning styles (CSLs) and users
suffering from attention deficit disorder (ADD and ADHD)
obtain special guidelines.
The metaguidelines were increased by more than 18
categories and 18 new rules are created according to the
surveys [9], [10], [11] and [12]. The selection rules help the
designer to automatically select the most appropriate
metaguidelines. For this project we raised through the
literature one hundred thirty-six new guidelines.
A. Representation of Guidelines
The knowledge basis of G uideExpert consists in the
“WHEN-THEN” rules, who help to select the appropriate
metaguidelines for the interface being designed. This study
adds to the base already built the 18 selected rules that were
595