W3C home > Mailing lists > Public > public-tt@w3.org > December 2008

Re: xml:id and/or id? (Re: TT telecon - agenda for 5 Dec)

From: Silvia Pfeiffer <silviapfeiffer1@gmail.com>
Date: Sat, 6 Dec 2008 18:14:24 +1100
Message-ID: <2c0e02830812052314s41c849a0s3a2abd2cd0730b3f@mail.gmail.com>
To: "Philippe Le Hegaret" <plh@w3.org>
Cc: public-tt@w3.org

Ah, thanks - only just saw this email now.
I think that's a good decision - only one mechanism to specify identifiers.
In Gecko it may well be possible to simply map that to "id".

Thanks,
Silvia.

On Sat, Dec 6, 2008 at 3:23 AM, Philippe Le Hegaret <plh@w3.org> wrote:
> Hello Silvia,
>
> we discussed the topic of xml:id vs id again today.
>
> Since all existing DFXP players are either using xml:id or planning to
> do so, we concluded that the specification was fine as it was and it was
> better indeed to support only one way mechanism.
>
> Regards,
>
> Philippe
>
> ISSUE-14
>
>
> On Thu, 2008-12-04 at 23:08 +1100, Silvia Pfeiffer wrote:
>> Hi,
>>
>> If I may, I'd like to contribute some information to the xml:id vs id
>> discussion.
>>
>> I am aware that most players by now support xml:id, and that this may
>> be regarded as a reason to keep it in the spec.
>>
>> However, I have had feedback from another person looking to use DFXP
>> for subtitles in Web video. Here is what he wrote:
>>
>> "xml:id support in Gecko was attempted but it was pulled due to
>> performance issues. xml:id support has explicitly been decided against
>> in WebKit. The problem with xml:id on the general level is that it
>> breaks the assumption that an element can have only one ID attribute.
>> xml:id has already turned out to be more complex than expected for the
>> HTML5 Validator.nu. Currently, it seems that xml:id is falling out of
>> favor when it comes to browser-targeted XML vocabularies. (I'd prefer
>> DFXP to use an element called 'id' in no namespace instead of xml:id.
>> However, if at this point existing implementations support xml:id, it
>> may be better to stick to xml:id as the *only* way--since having two
>> ways is the problem--as long as xml:id isn't leaked to the
>> general-purpose Gecko DOM e.g. through a scripting API.)"
>>
>> I would support removing xml:id - in fact, I would support removing
>> the use of namespaces where possible. It will make implementation much
>> more lightweight. But that's a different discussion altogether.
>>
>> Best Regards,
>> Silvia.
>>
>>
>> On Thu, Dec 4, 2008 at 8:45 PM, David Kirby <david.kirby@rd.bbc.co.uk> wrote:
>> >
>> > Here's the agenda for tomorrow's meeting.
>> > David
>> > =============
>> >
>> > TT agenda for Friday 5th Dec 2008
>> > Time: 10am Eastern, 3pm UK, 7am Pacific.
>> >
>> > 1) Review action items
>> > www.w3.org/AudioVideo/TT/tracker/actions/open
>> >
>> > 2) Issues arising
>> >    www.w3.org/AudioVideo/TT/tracker/issues/open
>> >
>> > [Plus:
>> > a) Spec needs to say that user agent must define a default region (replaces
>> > action item 15)
>> > b) Spec needs to say that a dfxp file containing only timing and text will
>> > be displayed by a UA.
>> > c) tts:textDecoration should be "lineThrough" and "noLineThrough" (Geoff's
>> > email)
>> > d) allow both xml:id and id?
>> > e) Support for multiple languages (i.e. xml:lang on div?) ]
>> >
>> > 3) Proposed survey of issues
>> >
>> > 4) Progress with testing
>> >  - xml:space (clarify behaviour)
>> >  - Any other issues arising from the testing?
>> >
>> > 5) dynamicFlow
>> >
>> > 6) SMPTE
>> >
>> > 7) AOB
>> >
>> >
>
>
Received on Saturday, 6 December 2008 07:15:01 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 2 November 2009 22:41:39 GMT