Workpackages 2 and 3
The research & development that focuses on language was concentrated in workpackages 2 and 3. Together, they formed one of the two pillars under the AAC Internet Applications built in workpackage 6.
WP2: Concept Coding Framework, Concept Sets and Support Modules
Where symbols are used as an alternative or complement to character based texts, referencing to, and exchanging these alternatively coded messages over the Internet is difficult. This is due to the lack of standardised encoding schemes and common practises.
WWAAC strove for the introduction and adoption of a common, open and vendor neutral Concept Coding Framework (CCF) to overcome some of these difficulties. To spread the ideas about concept coding and involve other organisations in the further development of these ideas, main stakeholders and experts from several countries have been invited to workshops.
For more information see the concept coding framework website.
We promote a basic set of concepts as a contribution to a public standard for information encoding. The concept set is based on the symbol systems Bliss, PCS and Pictogram and on English, Dutch, Swedish, Finnish, Danish and Spanish wordlists.
The design and definition of a universal concept set and additional support mechanisms was accomplished in close co-operation with WP3. The results were usefull for both the components of WP3 and the Internet applications.
The activities in this workpackage could be characterised by Innovative Technology Development, contrary to the activities of WP3 which were more research focused.
WP3: Linguistic structures for e-mail
The challenge for this workpackage was, to make e-mail conversation possible between symbol users and people who are not acquainted with any symbol system.
Under the direction of KTH, support mechanisms have been developed. These mechanisms ensure that the symbol input and the text or speech output are as good and grammatically correct as possible. This has been done for English, Finnish, Danish, Dutch, Swedish and Spanish. Work to make the addition of new languages simpler in the future has also been carried out.
The workpackage took advantage of previously developed symbol-to-speech (and, in one case, symbol-to-text) rules. These rules can produce grammatically correct sentences, when a Bliss user provides the symbols in a grammatically correct "word" order.
These rules have been worked out for Dutch, Swedish, English, Danish and partially for Finnish and Spanish.
In the second half of the project more effort has been put into designing a writing support mechanism for users who are not able to provide symbols in a grammatically correct word order. Their symbol selection may be incomplete in some sense, perhaps telegraphic; so, the symbol sequence is likely to be misunderstood by the system and consequently by the reader.
The writing support mechanism guides the symbol user through the process of constructing correct sentences using a fixed number of correct syntactic structures. These structures have been obtained by constructing a large database of real e-mail messages and extracting the most common syntactic structures. Using synthetic speech, the user can listen to model sentences and check the intermediate results while substituting new symbols into the model sentences.