- From: Jim Larson <jim@larson-tech.com>
- Date: Sun, 26 Oct 2003 06:57:08 -0800
- To: ishida@w3.org, duerst@w3.org
- Cc: w3c-i18n-ig@w3.org, www-voice@w3.org
Minutes of the telecon between VB and Internationalization on November 24, 2004. Richard Ishida Dan Burnet Jim Larson Martin Duerst Reviewing old items and confirm the status of each old item: Item 22: havent heard from Luk yet. Item 2: Martin is going to elevate issue to Philips HCG group Item 20: will discuss later Item 51: VBWG will rewrite; will review with final text review Item 54: Withdrawn Now we move to the new items: Item 46: Can be postponed to V3. This may introduce some incompatibilities if adopted. Martin: Try to keep text out of attributes as much as possible in future revisions. Dan: We want to keep text that is to be spoken as content and move other stuff to attributes. Item 42: Concern with openness of alphabet set and difficulties in portability. Dan: because of current state of the art and because different venders use different alphabets, we can not do this today. Even within alphabets, there may be different subsets that are implemented. VBWG examined several alternative solutions: We discussed several possibilities, with each having different disadvantages. There is no good W3C mechanism for reserving name spaces. Qualified names are extremely long. Full URI are also to long. Registered ID was disliked by all. Dan: Currently we believe there will be self-policing in this area. Our final belief is that we did not want to make a specific decision. Dan: Do you like the idea of each vender using their company name as part of the name of the character set? Martin: What if a new company wants to use a new label? How to avoid collisions? Dan: I will take the idea of using a short identifier, such as x-company name back to the VBWG. Dan: Will this pass all the way through the W3C chain, or will we get push back. Martin: It will probably be OK Dan: Dan Connelly will likely push back. Martin: Id try to sell it this way. Dan: we would use and example using more than just the company name: x-company name optionally followed by more stuff. Martin: I think the internationalization group will be OK with this. Bottom line: If the VBWG agrees with this approach, then is approved by the internationalization group. Go back to item 20: Martin sent e-mail about this item: Martin: only two examples: (1) this is the base with everything in one language (2) text is English, alt text is German, audio is German Martin: Wants to indicate the language of the WAV file so that specific audiences can avoid certain languages. Dan: We have never tried to do this in SSML. There is nothing that can be done to the file, so there is no need to indicate its language. Richard: There is a more general situation. Assign a language to a large chunk of language and then have exceptions. Dan: We could not think of any use cases for wrapping a voice file with a language. Dan: When does the content of the description element need to be a different language from the audio element? Martin: That would be a weird case. Richard: Use XSL to pull all German files, possibly in a language training application. Dan: You would like language indication on audio files. Richard: yes Dan: Use the voice element in SSML to do this. Martin: Use XML:lang on the description element Richard: SSML lesson for teach yourself language course. Create a smaller file with just a single file. Extract only the files for a specific file. Dan: Is this an issue in SMIL? Dan: Is this enforceable? There is no way to enforce developers putting a voice element around the audio file. Richard: It should be optional. General comments: Dan: Item 44: If someone uses something outside of the current language, it may not be supported. For example, an African click within English. Dan: Item 1: We dont want example in spec that most venders can not relate to. It is NOT that we dont want Hebrew and Arabic. Some languages need context in order to determine how the word is spoken. I.e. read. We believe that we can likely complete these discussion in the next telecon.
Received on Sunday, 26 October 2003 09:57:06 UTC