Antw: Re: question on notation systems

[Thomas Hanke]

Franz Dotter wrote:
>The posting of Corrie gives me the opportunity to ask you the following:
>
>Many programs which are advertised as to "translate" written into sign
>language only perform a method of producing a signed word for every
>written word. Seen objectively, this is no translation at all, at the best
>we would get a variant of Signed Dutch or English, etc. Is that what the
>deaf community wants?
>
>I ask my question because
>1. the false advertisements of "translation" have a false impression of
>sign language in the public as  their consequence (a fact which I dislike
>very very much) and
>2. I do not know how the sign language community looks at the problem (and
>woul follow the majorities direction).
>>
>Best Regards
>Franz
>University of Klagenfurt

Dear Franz,
you may be right that some programs advertized as sl translation tools in
fact only do a word-by-word translation resulting in kind of Signed
English. However, this is not true for the project Corrie mentioned:
ViSiCAST is a major European collaborative research project funded by the
European Union, with translation into sign language as one of its major
activities. We invest a huge amount of time in carefully modelling the
grammars and lexicons of three European sign languages, and we think that
this project will substantially contribute to the young field of
computational sign linguistics. Obviously, with limited time and limited
resources, you cannot model all aspects of the language, not even all the
aspects you know about. As a consequence, at the end of the project we will
be satisfied if our system provides correct translations in a very
restricted domain. ("Correct" in that context obviously is defined by the
judgement of native signers.)

You now might wonder where HamNoSys fits into this context. First, HamNoSys
is one of the major building blocks for the description system that
interfaces between the translation process and the computer-animated
presentation of the sign language sentences. (The fact that HamNoSys was
designed to be usable across sign languages is essential for us as we have
three target languages.)

Secondly, the lexicon contains sign formation fragments also expressed in
HamNoSys that flow into the constraint-based (HPSG) language generation
process driven by a semantic representation (DRT) of the source language
utterance (English only for the time being). If you will attend the Gesture
Workshop 2001 in London, we will be happy to give you more details. If you
would like to see more background information on ViSiCAST, please visit
http://www.visicast.sys.uea.ac.uk.

Best, Thomas


--------------------------------------------------------------
Thomas Hanke
Institute of German Sign Language and Communication of the Deaf
University of Hamburg
Binderstrasse 34
20146 Hamburg - Germany
Phone ++ 49 40 428 38 6742
Fax ++49 40 428 38 6109
mailto:Thomas.Hanke at sign-lang.uni-hamburg.de
http://www.sign-lang.uni-hamburg.de
--------------------------------------------------------------