- From: Jonas Sicking <jonas@sicking.cc>
- Date: Mon, 10 May 2010 17:51:17 -0700
- To: Sean Hayes <Sean.Hayes@microsoft.com>
- Cc: Philip J�genstedt <philipj@opera.com>, "public-html@w3.org" <public-html@w3.org>
On Sun, May 9, 2010 at 1:13 AM, Sean Hayes <Sean.Hayes@microsoft.com> wrote: > Regarding a) and b) It was you that compared the accessibility level to HTML, so I was curious as to what you meant by that, but yes in general making the format available as marked up text achieves the goals you state. My point was that while HTML is designed to mark up the semantic meaning of various pieces of content (headers vs. paragraphs vs. tables vs. navigation etc) and has been moving away from marking up style, it seems to me that TTML is much more designed around marking up just timing, as well as marking up style. However it does seem like TTML does allow marking up semantics as well, using the role attribute, which is better than my initial perception. > There is a trade-off between inventing new elements for every possible semantic role, and having an annotation mechanism; there are merits to both. On balance the TTWG felt that due to the large, and somewhat open ended number of possible roles, that an attribute value was preferable to convey this metadata, particularly as there are a number of different axes of metadata that may need to be conveyed. I agree having an element for every possible semantic role would likely not be good. However having an @role value for every possible semantic role. But I agree, everything is a trade-off and a judgment call. I just would have made different ones than the TTWG. >>>I would have thought that everyone that author TTML files, and that want to apply styling in the TTML, would have to learn XSL:FO? > > That seems to be a common misconception. The TTML specification is intended to provide adequate documentation of what all the style attributes do. It is not expected as I said that many authors will be hand crafting caption files. Those that do might find that knowledge of XSL:FO may be of help, but it would not be a requirement, Knowledge of CSS would probably also work, particularly for the attributes that are commonly employed like colour and font. The controls offered by TTML are far fewer than one would find in a typical HTML/CSS stylesheet, so it is much less likely that complex edge cases will be encountered. Sorry I wasn't very clear. I meant that it seems to me that everyone that author TTML files, and that want to apply styling to that TTML, would have to learn the parts of XSL:FO that TTML uses. / Jonas
Received on Tuesday, 11 May 2010 00:54:32 UTC