[sw-l] SWDB - Linguistic SignWriting Database

[rocha at ATLAS.UCPEL.TCHE.BR]

Sandy,

  Now you are right! There is no z-data (layer information) in SWML (yet,
at least :-)

  The original SWML format was modeled after the internal representation
of signs in the SignWriter program, and that information was not present
there
either.

  That is, SignWriter had its own internal way of dealing with the z-data,
that was not refleted in the files that stored the texts and
dictionaries.

  It was clear from the begining that this was an issue that had to be
solved, but I decided (maybe wrongly) to follow the same track and assumed
that each application would deal with layers in its own way.

  The option was, at least, consistent with the SignWriter program,
because in SignWriter the symbols are transparent, so the z-data are not
present visually.

  But I had not noticed, until now that you mentioned it, the discrepancy
between SignWriter and the manual.

  Perhaps we should ask Valerie to think about the issue and let her
choose the best way (if symbols should be transparent or not), and then
decide if the z-data are important for SWML.

  All the best,

  Antônio Carlos




> Antônio Carlos,
>
> Ah, I see it now - I missed the little 1..infinity things  :)
>
> Thanks for explaining!
>
> Looking further into this, I do have another concern about SWML, though.
> In
> the "Lessons in SignWriting 2002" manual, the example signs often overlap
> opaquely. However, examining the SWML DTD version 1 and the SWML-S DTD I
> can't see how the z-layer is handled. How is the fact that one hand is on
> top of another expressed?
>
> I notice that the gifs used for the Puddle dictionaries are completely
> transparent, so this couldn't be expressed in Puddle anyway. Is it
> unimportant or do we really need three-colour gifs so that one colour can
> be
> designated as transparent for use by software?
>
> Hopefully I'm wrong again, but I still need an explanation!
>
> Sandy
>
>> -----Original Message-----
>> From: owner-sw-l at majordomo.valenciacc.edu
>> [mailto:owner-sw-l at majordomo.valenciacc.edu]On Behalf Of
>> rocha at ATLAS.UCPEL.TCHE.BR
>> Sent: 22 October 2004 18:08
>> To: sw-l at majordomo.valenciacc.edu
>> Subject: RE: [sw-l] SWDB - Linguistic SignWriting Database
>>
>>
>> Sandy,
>>
>>   Gladly to us, you read the schema wrongly :-)
>>
>>   Translations can be in any number of sign languages (and any number of
>> oral languages too :-)
>>
>>   And signs are identified sequentially by numbers. That is, they are
>> not
>> identified by gloss, not even for basic search: signs have to be search
>> as signs.
>>
>>   Sign languages treated as languages in themselves in context of
>> computational systems has been my major concern even before I met
>> SignWriting. By the way, that is the reason I started to search for
>> something like SignWriting, before I knew it :-)
>>
>>   All the best,
>>
>>   Antônio Carlos
>>
>>
>> > Antônio Carlos,
>> >
>> >>    We are calling it SWDB - Linguistic SignWriting Database (but we
>> are
>> >> not sure yet this is the best name).
>> >>
>> >>    The SWML format, called SWDB, has been added to the SWML site
>> >>
>> >>       http://swml.ucpel.tche.br
>> >>
>> >> so you can check it. It's an XML schema, not a DTD, but it is
>> shown as a
>> >> diagram. The diagram is in the PDF format.
>> >
>> > It's possible that I'm reading the schema wrongly, but if I'm reading
>> it
>> > correctly I have a serious objection to it.
>> >
>> > It seems to assume that the sign language will have an oral language
>> > translation.
>> >
>> > Why shouldn't there be a linguistic database with translations to sign
>> > languages?
>> >
>> > For example an ASL database for BSL users who have poor English or
>> don't
>> > want to work with English as a go-between for the two sign languages?
>> >
>> > This is something I've thought of before - at the moment glosses for
>> use
>> > by
>> > editors and dictionary software are always in oral languages
>> but it should
>> > be possible to supply sign language glosses for those who want to work
>> > directly from one sign language to another.
>> >
>> > This makes sense to me, as oral languages are quite "foreigh" to sign
>> > languages and only being able to work between a sign and oral language
>> > when
>> > trying to learn a foreign sign language seems to me like an English
>> > speaker
>> > trying to work with Dutch but only having a Dutch-Japanese and
>> > English-Japanese dictionary.
>> >
>> > In fact it's worse because imagine I wanted to access Brazilian Sign
>> > Language - I'd have to go through English and Portuguese to get at the
>> > Brazilian Sign Language instead of just going direct from British Sign
>> > Language to Brazilian!
>> >
>> > While not many people might want to produce an authoritative
>> > British-Brazilian SL dictionary, signers who want to learn or teach a
>> > foreign sign language should at least be able to set up a database
>> like
>> > this
>> > for their own or classroom use. And how about researchers comparing
>> sign
>> > languages?
>> >
>> > I hope either I've read the schema wrongly, or you or Juliano find
>> this
>> > suggestion helpful!
>> >
>> > Sandy
>> >
>> >
>>
>>
>> -----------------------------
>> Antônio Carlos da Rocha Costa
>> Escola de Informática - UCPel
>>
>>
>>
>>
>
>


-----------------------------
Antônio Carlos da Rocha Costa
Escola de Informática - UCPel