- From: David Kirby <david.kirby@rd.bbc.co.uk>
- Date: Fri, 16 Jan 2009 17:12:07 +0000
- To: public-tt@w3.org
Timed-text working group minutes 16 January 2009 Present: David Kirby (DK, co-chair, scribe) Sean Hayes (SH, co-chair) Geoff Freed (GF) Andrew Kirkpatrick (AK) Glenn Adams (GA) Philippe le Hegaret (PH) via irc Regrets: John Birch Frans De Jong Minutes from call PH: Awaiting implementations from AK and SH; this is starting to get in the way of progress. AK: sorting out our implementation has been delayed; hope to have it ready today. SH: should have the Microsoft implementation ready by end of Feb. GF: Some questions arising - bidi-override and direction - are these both needed? GA: Yes, bidi-override and direction are independent parameters and both are required for Hebrew and Arabic text. To be consistent with W3C translation of Unicode to XML we need both parameters. GF: For consistency with CEA 708, in earlier discussion said we need some features... AK: for example, borders on regions... GF: I'm undecided as to whether we really need to hold up the process in order to include these in DFXP. AK: I'm wondering about that too - does any manufacturer use this? SH: this goes back to the discussion as to whether we should be able to round-trip 708. It's not definitely needed in DFXP, but we should check that the lack of some of these features from 708 isn't going to be a problem. GF: region transparancy/fading falls into this category GA: for fading, if it really is needed, we have a way of doing it now [using set??] so nothing needs to be added. GF: Unsure about demand for some of these 708 features in US; possibly not really needed at the moment. GA: Looking at Q4 of issues questionaire - metadata already seems to be supported on regions. SH: metadata data attributes allowed there but not elements GA: OK DK: Re Q6, some of the test files appear to use ttm:descr rather than ttm:desc - is that correct? SH: Philippe fixed those earlier; they should be ttm:desc. DK: Looking at timing tests just emailed by Sean, at least one of them seems to use ttm:descr. SH: May have sent earlier versions to list; will check. DK: Re ttm:actor - this isn't included in tests yet as unclear on exactly how it is to be used. GA: ttm:actor links a real agent, e.g. Sean Connery, to a fictional one, e.g. "James Bond". Will send example to the list for discussion. DK: It seems to be conveying details that could be carried as an attribute of ttm:agent and this may be an easier way to do it. Will consider this more once example is available via list. SH: can Philippe add a text box to the questionaire, where it says "Needs discussion" GA: suggest this be a separate field to allow comments whatever the answer. [DK - request forwarded to PH via IRC just after the call - and done by Philippe before I managed to get these minutes out!] GA: report progress on editing of revised Rec - changes to tts:extent finished, work on default region is underway. Will also add "requiredFeatures" and those related elements, as previously discussed, as editor's notes for consideration. Updated spec should be available sometime next week. GA: Note that Q19 (throughLine or lineThrough) has already been fixed as it was really a typo. [Call ends]
Received on Friday, 16 January 2009 17:12:39 UTC