FW: Some comments on the requirements

-----Original Message-----
From: Bert Bos [mailto:bert@w3.org] 
Sent: Tuesday, February 04, 2003 11:50 AM
To: www-tt-tf@w3.org
Subject: Some comments on the requirements


The timed-text requirements document is old and probably no longer up
to date anyway (requirements have a tendency to age very quickly), but
I thought I would send some comments anyway. They should still be in
time for the first working draft...

I Architecture

2 Have a valid XML representation

    Apart from the fact that this contradicts the first requirement
    ("simple and easy"), I think that limiting the syntax up front is
    not very useful. Especially so, because of the timed-text formats
    (and proposed formats) that I have seen so far, those that were
    not based on XML were by far the most elegant. (E.g, those of
    Quicktime and Mplayer.)

8 Allow the language of the text to be identified using xml:lang

    A corollary of the previous comment. The language should indeed be
    identified somehow, but it could be by other means than XML.
    Indeed, it could be outside the document itself, e.g., in an HTTP
    header.

11 Have a default UNICODE font

    Is it really a requirement to have a standard font? What font
    would that be? Arial? Maybe the intention was to require that the
    format supports all Unicode characters, rather than a specific
    font?

18 Allow motion [...]

    I think this requirements makes the format too complicated and
    makes it approach SMIL and SVG too much. Those formats already
    exist. What's needed here is something simple, as requirement 1
    says.

19 Use SVG, MathML, XHTML or other language for complex font displays

    I agree with this requirement as such (as I said in the previous
    comment), but its appearance here seem to suggest something more,
    viz., that UAs that support timed text also have to support SVG,
    MathML and XHTML. That seems rather pointless to me. If the UA
    already supports those three, then there is no need anymore for
    timed text. I thought the purpose of timed text was precisely to
    have something that is simpler than those formats and still does
    the job in 95% of the cases.

22 Adopt SMIL 2.0 as a base language.

    Like requirement 2 ("XML"), this puts too many a priori
    restrictions on the format. Besides, unless you subset SMIL so
    much that it doesn't look like SMIL anymore, it would not be
    simple and easy, as required by requirement 1.

23 Be no less functional than EIA-708 [...]

    Interesting, but since EIA-708 is not available on-line, I don't
    know what this requirement means :-(

II Display

12 Allow other ways to display text; for example, via text balloons.

    Good idea, but what exactly is required? That such other ways are
    not made impossible by the format? Or that the format itself
    contains provisions to make text balloons? I think some basic
    styling (font, color, justification...) should be available within
    the format (based on the CSS formatting model and properties
    catalogue, I assume), but if somebody wants text balloons, that
    seems to me to be a separate system, which only uses documents in
    the timed-text format by reference.

III TIMING

4 Define text and timing markup in two separate modules in the
specification.

    Why do the requirements talk about the format of the specification
    itself? Maybe it turns out that there are a handful of separable
    modules, but requiring that they exist seems rather unnecessary.
    Indeed, I think this specification will be so small that it is
    even not necessary to split it into chapters. I'd be tempted to
    add a requirement: the specification should not contain more than
    10000 words...



Bert
-- 
  Bert Bos                                ( W 3 C ) http://www.w3.org/
  http://www.w3.org/people/bos/                              W3C/INRIA
  bert@w3.org                             2004 Rt des Lucioles / BP 93
  +33 (0)4 92 38 76 92            06902 Sophia Antipolis Cedex, France

Received on Tuesday, 4 February 2003 21:18:07 UTC