October 2006 Volume 9 Number 4

From this definition, we figured out alternative mechanisms to (i) indicate which pages compose the learning
object and which content elements compose each page, (ii) verify these content elements express all the required
information and (iii) add metadata associated to the learning object. We have provided many facilities to sustain
contents creation also by authors which are non-expert in technologies and standards concerning accessibility
and usability issues. In particular for accessibility, we supply authors with MSWord styles which are accessible
with a custom toolbar. Each style is meant for a specific role in the final document; the role could regard a
particular required accessibility constraint or a predefined semantic position in the text. The correct use of styles
according to their implicit semantics is managed by some macros which alert the author about lacks of
information. Moreover, the author can use a Word embedded form to fill in specific metadata.
An ISA-BeL author, in fact, writes his/her contents by using a word processor, and he/she assisted by three
different tools, specifically designed to cover these three aspects:
An authoring toolbar which offers a fast access to him/her main activities such as defining presentational
aspects and structural elements, inserting accessibility related information and so on. A screenshot of the
ISA-BeL toolbar is depicted in the following Figure 2.
A verification toolbar which runs automatic controls over the respect of accessibility and universality
constrains. For example, some exploited controls are: have any image an alternative description? Has any
table a summary? Are titles used according to the correct hierarchical structure? Other manual controls are
made by the post-production team (about foreign language words, acronyms, abbreviations, and so on); if
the LO does not respect such constrains then authors have to correct and control contents again, before
starting the production phase.
A set of forms to collect SCORM Metadata. As expected, a key role is played by the SCORM metadata:
while some of them can be derived by the system (last-saved date, file size, language, version number and
so on), other information have necessarily to be provided by the authors.
A step-by-step description of author activities, from creation to production is detailed in the following Section 4.
Production
Figure 2. The ISA-BeL authoring toolbar (in MS Word)
The production process is performed by an ad-hoc application, ISA-BeL which is composed of two modules:
ISA (Immediate Site Activator): a conversion tool which actually transforms document from the word
processor format into an intermediate XML representation, enriched by all the necessary metadata. An
introduction about ISA architecture and functionalities can be found in (Vitali, 2003).
BeL (Backed e-Learning): a stand-alone application which gathers all the information stored in the
intermediate XMLs, creates the SCORM structures (in particular the tracking scripts and the manifest file)
and merges the content into a single .ZIP file, by processing the output of ISA. BeL also integrates into the
LO a (multimedia) recorded accessible video lecture, which is automatically transcoded through a different
line of the LO production. More information about this transcoding process can be found in (Salomoni et al.,
2005).
The production process is based on a set of templates and configuration files which are used to define structural
aspects as well as layout and graphical aspects of the automatically produced LO.
Delivery
The system has been used for the above mentioned A 3 project. Contents which were produced by the automatic
process are loaded on the e-learning platform by means of standard (for SCORM compliant contents) import
procedures. Our choice has been made between the large number of open-source platforms and after a deep
8