RE: Comments on Working Draft "Timed Text (TT) Authoring Format 1.0 Use Cases and Requirements"

Dear Guido,

Thanks for your further comments.

Regarding your follow-on comments to our respones, I will ensure
they are more fully discussed, however I expect the response to
be something as follows:

GG-09-01 #1: On integration with W3C Technologies

Consider adding requirement that explicitly indicates that if
TT-AF-1-0 is used as a distribution format, then it should be
capable of being referenced as a SMIL text media object (but
not necessarily textstream media object).

GG-09-01 #2: On Common Authoring and Distribution Formats

Respectfully disagree. Current consensus of group is to define
authoring format first in order to support transformability to
existing distribution formats. Current text does not preclude
use of TT-AF as a simple, non-streaming distribution format.

GG-09-01 #3: On Basic Profile

This *may* be possible, but may end up being not very useful.
If defined, I would anticipate such a profile to be extremely
general, such as: "TT-AF Basic Profile includes the following:
text content, timing, and style information. The conformance
requirements on content are as follows: (1) if author wishes 
to include text content, then he should do so; (2) if author
wishes to include timing information, then he should do so;
(3) if author wishes to include style information, then he
should do so. The conformance requirements on processing are:
don't throw away information unless you intend to do so."
Get the idea... Such a profile would not make statements about
what facilities should or should not be supported in a user agent.

GG-09-01 #4: On Disallowing Device Dependence

Respectfully disagree. If an author's intention is to make
content device depenedent, e.g., to predetermine available
geometries in order to precompute line breaks, then the
author should be able to do so.

********************

The above should be considered my personal opinion at this time
until further corroborated by the TTWG.

I would welcome more direct involvement by Nokia in the consensus
process of the TTWG. The work and outcome of the TTWG, like other
W3C WGs, is effectively based on volunteer activity by its members.
The extent of volunteerism serves as a limiting factor on what
requirements a WG will commit to accomplish. In the case of
TT-AF-1-0 requirements as presently documented, I believe there is
sufficient commitment by members to accomplish this work. However,
I believe there is NOT sufficient commitment to accomplish this
work AND the possible work of defining a distribution format
IN PARALLEL. That is why the TTWG has decided to defer the work
of defining a distribution format. We are following a tried and
true principle of engineering: piece-wise refinement.

Regards,
Glenn


> -----Original Message-----
> From: guido.grassel@nokia.com [mailto:guido.grassel@nokia.com] 
> Sent: Friday, January 09, 2004 4:24 AM
> To: Glenn A. Adams; public-tt@w3.org
> Cc: Art.Barstow@nokia.com
> Subject: RE: Comments on Working Draft "Timed Text (TT) 
> Authoring Format 1.0 Use Cases and Requirements"
> 
> 
> Dear Glenn, dear TTWG members,
> 
> Thank you for considering our comments and sending a detailed 
> response.
> 
> GG 09-01: While reviewing your response I noticed that a 
> reference to the TTWG charter is missing. In fact, I can not 
> find it from the TTWG public Web pages either. Reference to 
> the charter is important because requirements need to reflect 
> what the TTWG has been chartered to produce. Searching the 
> W3C site I found a document at 
> http://www.w3.org/AudioVideo/TT/ttcharter20020901.html In the 
> following, I assume this doc is the charter of the TTWG. I 
> will refer to it as "the charter".
> 
> Feedback to your disposition of comments below, marked as "GG 09-01:".
> 
> BR
> - Guido
> -----Original Message-----
> From: ext Glenn A. Adams [mailto:gadams@xfsi.com]
> Sent: Thursday, January 08, 2004 8:10 PM
> To: Grassel Guido (Nokia-NRC/Helsinki); Barstow Art (Nokia-TP/Boston)
> Cc: public-tt@w3.org
> Subject: RE: Comments on Working Draft "Timed Text (TT) 
> Authoring Format
> 1.0 Use Cases and Requirements"
> 
> 
> 
> Dear Guido,
> 
> The TTWG has reviewed your comments on [1] and provides the
> following responses (inserted inline below).
> 
> [1] http://www.w3.org/TR/2003/WD-tt-af-1-0-req-20030915/
> 
> We greatly appreciate Nokia's efforts to review and comment, and
> understand that this requires valuable resources to accomplish.
> 
> Regards,
> Glenn Adams, Chair, for TTWG
> 
> > From: <guido.grassel@nokia.com> 
> > Date: Mon, 8 Dec 2003 14:41:28 +0200
> > To: <public-tt@w3.org> 
> > Cc: <guido.grassel@nokia.com>, <Art.Barstow@nokia.com> 
> 
> > NOK-1: The TT AF should not duplicate functionality that can already
> > be found from existing or upcoming W3C Recommendations. Instead it
> > should adopt useful functionality from W3C languages such as SMIL,
> > XHTML, SVG or CSS. We request adding such a requirement to section
> > 4.1. 
> 
> We interpret this requirement as stating essentially the following:
> "don't invent something new [unless you have a good reason to do so],
> and in case you don't invent, then adopt an existing W3C technology".
> 
> Although we view this as a very basic requirement under which we have
> been operating, we have no objection to documenting it as such;
> therfore, we will add a new general requirement to this effect.
> 
> GG 09-01: OK
> 
> > NOK-2: Use of the TT AF in combination with SVG and SMIL are very
> > important and should be mentioned as a requirement. For instance, it
> > should be possible to use the TT AF as a 'textstream' media 
> object in
> > a SMIL presentation.
> 
> The current work (and requirements) document focuses solely on the
> specification and use of an authoring content format, and not a
> distribution format. 
> 
> Since SMIL is effectively designed to work with distribution formats
> for use by its media objects, we believe it is best to not explicitly
> define such usage in the TT Authoring Format Requirements.
> Nevertheless, we recognize the need for such a usage, and intend to
> address that as possible future work of the TTWG. In this regard, we
> would solicit Nokia's participation in helping define such a
> distribution format.
> 
> Note that the current requirements document does not exclude the
> use of the authoring format as a distribution format; so, it would be
> possible, though perhaps not desirable, to use the authoring format
> directly as content referenced by a text media object in SMIL. Keep
> in mind, however, that it is unlikely that the current authoring
> format will be expressed in a format that is suitable for streaming,
> particularly not suited for arbitrary stream entry points.
> 
> 
> GG 09-01: The definition of the scope of the TTWG in the 
> charter reads (1st bullet): "Develop a new Timed Text format 
> that integrates well with other W3C technologies.". I think 
> this is a clear requirement to the TTAF. See also comments on NOK-3.
> 
> 
> 
> > NOK-3: Use of the TT AF as a distribution format is insufficiently
> > represented in use cases and in requirements. It appears that the TT
> > AF is primarily intended as an authoring format that serves as input
> > to a transcding process into a proprietary distribution format. Use
> > of the TT AF as distribution format should be at least equally
> > important as serving as an authoring format.
> 
> Your observation is correct. And we do view a distribution format as
> important as an authoring format. However, we do not agree that use of
> the authoring format directly as a distribution format is desirable.
> Further, we have determined that the group should take up the formal
> definition of a distribution format only after we have completed an
> authoring format.
> 
> One of the primary requirements driving an authoring format is the
> existence of many distribution formats, with none of these being
> sufficient as an authoring format interchange standard. As a
> consequence, our current focus is on satisfying this need rather than
> adding one more item to the already large set of distribution formats.
> 
> As a side-bar, we have already noted that SVG would be a reasonable
> distribution format.
> 
> GG 09-01: This plan to make two specifications does not 
> become clear to the reader of the reviewed requirements 
> document. Furthermore, the charter talks about specifying one 
> format not two.
> 
> It is not clear to me why a TT format can not serve both 
> authoring and distribution formats. Why is TT different from 
> other Web technologies? The TT group should provide clear 
> evidence that specification of two formats has advantages 
> over the specification of one format. 
> 
> There are use cases where a person wants to first read a TT 
> document and modify (re-use) this content afterwards. This 
> use case can likely not be supported with separate authoring 
> and distribution formats. Person-to-person messaging in the 
> mobile domain, e.g. Multimedia Messaging (MMS) is one example 
> for this use case. As of today, the highest use of 
> synchronized multimedia and SMIL is in MMS.
> 
> In summary, we disagree with the specification of two 
> separate formats, one for authoring and another one for 
> distribution, instead of one format that can serve both purposes.
> 
> > NOK-4: A "Basic" language profile of the TT AF should also 
> be defined
> > that is suitable for distribution to constraint embedded 
> devices such
> > as mobile terminals. We request adding such a requirement to section
> > 4.1. 
> 
> We have discussed the issue of defining profiles for the authoring
> format and have determined that the axis for determination should be
> around authorial usage scenarios, e.g., subtitling versus captioning,
> visual presentation versus aural presentation (via text to speech),
> and so on.
> 
> When the TTWG does take up the definition of a distribution format,
> then it is expected that device capabilities will be a determiner in
> profiling the distribution format.
> 
> 
> GG 09-01: A "Basic" profile is needed for a distribution format only.
> 
> 
> > NOK-5 It must be possible to author TT documents in a device
> > independent way. We request adding such a requirement to 
> section 4.1. 
> 
> Because we have focused on an authoring format rather than a
> distribution format, the current approach is effectively device
> independent, since we are expressing authorial intention and not
> expressing device behavior. However, if you should have specific
> ideas about some features being device dependent, then please
> let us know.
> 
> 
> GG 09-01: Disagree. A format expressing author intention is 
> not necessarily device independent. Author intentions are in 
> many cases highly device specific. Some authors only have one 
> specific (set of similar) devices) in mind when creating 
> their content. This is a situation the Web needs to get away from.
> 
> Therefore, I encourage the TTWG to include this quite mild 
> requirement. A more stringent requirement on DI would be the 
> following: "The format should only allow to author TT 
> documents that are device independent. The format should 
> prevent that authors from creating documents that dependent 
> on specific device properties."
> 
>