Re: ISSUE-406 + #lineBreak-uax14

On Mon, Sep 21, 2015 at 3:43 PM, Pierre-Anthony Lemieux <pal@sandflow.com>
wrote:

> HI Glenn,
>
> > I presumed that by listing feature designators, you were specifying
> whether the features so designated may be used in a document.
>
> Yes. This is the intent.
>
> > it appears you may in addition consider the presence of a designator in
> these sections as indicating whether
> > a feature can be required in a document specific profile over and above
> the use of an IMSC profile.
>
> No. That is not the intent.
>
> I was merely curious why, as opposed to other TTML parameter that
> impact processor behavior, #lineBreak-uax14 is not controlled using a
> ttp attribute.
>

I don't believe anyone ever suggested defining a parameter, since it would
not be something typically changed on a document by document basis (in a
given profile). Instead, we felt such a requirement was closely aligned to
profile definitions, and that a profile could specify this requirement or
not.


>
> > Given the focus on reproducible behavior, e.g., defining the use of
> specific glyph metrics, I would expect that UAX14 be mandated by IMSC1.
>
> Ok. To confirm, unless IMSC1 requires that processor implement
> #lineBreak-uax14, a processor may or may not do so.
>

That is correct. In other words, the default is that you have no idea which
line break algorithm is used. If you specify #lineBreak-uax14 as required
in a standard or per-document profile, then at least you know what
algorithm is used (provided the implementation heeds #profile).


>
> Best,
>
> -- Pierre
>
> On Mon, Sep 21, 2015 at 2:15 PM, Glenn Adams <glenn@skynav.com> wrote:
> > I was not sufficiently clear in my description. Of course, the
> > #lineBreak-uax14 designator can be listed in a document defined profile
> > based on IMSC1 text profile as the base profile.
> >
> > What I had apparently not understood was what was meant by "The Document
> > Instance shall conform to the following table:" in Sections 6.10, 7.4,
> and
> > 8.4. I presumed that by listing feature designators, you were specifying
> > whether the features so designated may be used in a document. In the
> case of
> > most features, there is a syntactic feature that corresponds to a
> > designator. However, that is not true for #lineBreak-uax14, which may be
> > used by an application to signal whether a particular algorithm is used
> or
> > not by the implementation, as opposed to whether a syntactic feature is
> used
> > in the document.
> >
> > From your question, it appears you may in addition consider the presence
> of
> > a designator in these sections as indicating whether a feature can be
> > required in a document specific profile over and above the use of an IMSC
> > profile.
> >
> > This doesn't seem a legitimate requirement. If an application of IMSC1
> also
> > wants to require an implementation to support a superset of IMSC1, then
> they
> > should be able to additional require features not required by IMSC1.
> >
> > Furthermore, I wonder if the fact that absent a generic requirement for
> > UAX14 line break support, then there is no guarantee of reproducibility
> of
> > line breaks even given the same font metrics, same resolution, and same
> > region extents?
> >
> > Given the focus on reproducible behavior, e.g., defining the use of
> specific
> > glyph metrics, I would expect that UAX14 be mandated by IMSC1.
> >
> >
> >
> >
> >
> >
> >
> > On Mon, Sep 21, 2015 at 9:35 AM, Pierre-Anthony Lemieux <
> pal@sandflow.com>
> > wrote:
> >>
> >> Hi Glenn,
> >>
> >> Couple of questions related to #lineBreak-uax14 and ISSUE-406 [1].
> >>
> >> - why can #lineBreak-uax14 only be signaled through a profile
> >> declaration and cannot be set through a ttp attribute?
> >>
> >> - what is the default value of #lineBreak-uax14 in absence of a
> >> profile declaration?
> >>
> >> Thanks,
> >>
> >> -- Pierre
> >>
> >> [1] http://www.w3.org/AudioVideo/TT/tracker/issues/406
> >>
> >
>