Re: TTML similarity with SMILText profile of SMIL

Hello List,

In answer to Raphaël Hendricks question on the relationship between 
SMILtext and TTML, Philippe responded the TTML predated SMILtext. I 
think that he meant to say that work on TTML started before SMILtext 
went to Rec, but that TTML itself will -- for several good reasons -- 
come to rec a few years after SMILtext.

As someone closely associated with the SMILtext work (I wrote most of 
the spec), and as someone who has also served on the timed text working 
group, I think that simply saying one spec makes the other useless is 
not true. SMILtext was designed as a streaming text format with simple 
timing rules and a simple layout model. It also supports various forms 
of text motion as base properties of the text markup that can be 
directly embedded in an XML language or as a stand-alone content 
container. TTML was designed as an authoring format for an external text 
container that had a wider array of styling, layout and content markup 
options. For archival commercial systems or for systems that need to 
strictly adhere to a host of captioning formats, TTML is a language of 
choice. For more incidental streaming text (which is something that TTML 
doesn't address), SMILtest has advantages.

Personally, if we were to look for duplication, I am more concerned 
about the WebSRT effort than TTML. (But: this is my opinion -- no 
flames, please.)

There is a short TTML / SMILtext functional comparison in the SMILtext 
spec. Note that when this when to Rec, TTML was still called DFXP.

The bottom line: I agree that it doesn't make sense to stop TTML at this 
point.

Dick Bulterman
Co-Chair SYMM WG

 >Philippe Le Hegaret wrote:
> TTML predates the SMILText profile and is already deployed out there.
>
> So, imho, it doesn't make sense to stop TTML at this point.
>
> Philippe
>
> On Fri, 2010-10-01 at 12:18 -0400, w3c.fan@deadnet.endjunk.com wrote:
>> > To whom it may concern.
>> > The recent language TTML tries to provide a format for text inside a
>> > multimedia environment, especially for subtitles. This is however
>> > already achieved by the SMILText profile of the SMIL specification. In
>> > this context, the creation of TTML is likely to be redundant, this
>> > will make it more complex for implementors since they will have to
>> > support two formats, it will also be more work to maintain two
>> > specifications. It may be wise to see which, if any, features of TTML
>> > are missing from SMILText, implement them in the SMIL language, and
>> > add them to the SMILText profile instead of having to maintain two
>> > languages which provide similar features.
>> >
>> > Raphaël Hendricks
>> > w3c.fan@deadnet.endjunk.com
>> >
>> >
>> >
>
>

Received on Wednesday, 6 October 2010 17:15:25 UTC