Mash-Up Personal Learning Environments - E-Infrastructures and Technologies for Lifelong Learning - page 137

learner interactions: the LISL code of the current
activity page and the output of the LISL command-
line interpreter. The tab 'code' shows an inline
editor for editing the LISL code (not shown in the
figure). By switching the view to the tab 'preview'
(open by default and highlighted with a red rect-
angle in Figure 3), learners see the web application
mash-up space generated from the executed LISL
code. The code itself is created either by logging
traces of the interactions with the web-based
widgets shown on the MUPPLE page or by
manual scripting which is supposed to be done
by technically experienced learners only.
On executing a LISL script, the activity model
is built up, and the web-based applications are
launched within designated areas of screen. Now,
learners can work with the web-based control wid-
gets, whereby their interactions are materialized
by appending new LISL statements to the page.
On returning to a page, its last state is restored.
Furthermore, any part of the learning activity is
subject to adaptation by the user (facilitators and
learners) to ensure scrutability. Technically, LISL
is similar to AppleScript (http://www.apple.com/
applescript), but it does not limit the user to auto-
mate interactions with one application. Instead, it
enables learners to reflect their learning processes
and to collaborate in learning networks.

As stated in Mödritscher et al. (2008b),
LISL has been developed according to relevant
language design principles given by traditional
software development, user-related principles
for modelling, and young streams like end-user
development. Due to the possibility that (ex-
perienced) learners might prefer scripting their
environment instead of constructing it through
the web-based widgets, the focus has been laid
upon usability issues like learnability, efficiency,
simplicity, readability, uniqueness, seamlessness,
reversibility, and supportability. More technical
issues, e.g. security, fast translation, or efficient
object code, are considered to be of subordinate
importance. With respect to these principles, we
decided to engineer a close-to-natural language
for learning environment design. Subsequently,
our learner interaction scripting language (LISL)
is described in detail.
The lexical and syntactical structure of LISL
is kept simple, as shown with a code example in
Figure 4. The tokenizer splits each line of code
into tokens separated by white spaces. Single or
double-quoted constructs are considered to be one
token. Moreover, tokens are case-sensitive, and
identifiers are restricted to alphanumeric strings
(beginning with a letter) and may be framed by
quotation marks.

Figure 4. Example LISL script