RE: FPWD of Additional Requirements for Bidi in HTML

Hello Maciej,

> From: Maciej Stachowiak []
> Sent: 07 March 2010 03:10
> To: Richard Ishida
> Cc:;
> Subject: Re: FPWD of Additional Requirements for Bidi in HTML
> It's a little quirky to deliver feedback on the HTML WG's deliverables in
> the form of a Working Draft developed elsewhere, and it might have been
> better to make us more aware of this effort ahead of time. (I, for one,
> surprised to see  seventh Working Draft published with ours and was
> that I hadn't seen it before.)

Yes, I was as surprised as you to see the 6 HTML documents included in the
announcement of the i18n doc ;). I assume that was a decision made by the
W3C Comm team. It was pure coincidence that all those documents were
published on the same day.

The document we published was mentioned on the HTML list previously [1] and
I have discussed it with Paul Cotton on a couple of occasions, but until now
this has been an independent initiative developed by Google using a Google
list of 40 bidi experts (though I did offer the use of the W3C I18n Wiki,
since I felt that would help transition it to the W3C eventually).  The
author recently became an i18n WG participant, and publication of this
document and establishment of a new public-i18n-bidi list at this time are
the first steps in formally bringing this work into the W3C.  We've been
trying to move this forward as fast as we can.

> That being said, as long as the HTML WG gets the feedback in the end, I'm
> not too concerned with the process of developing it.
> From reading over this draft, it seems to me that most of this feedback is
> ready to be delivered to the HTML WG right now. I see many specific points
> that identify a specific problem in great detail, outline why the current
> state of the spec doesn't work, and propose at lest one workable solution.
> That's more than enough data to go into a bug report.
> I would expect that bug reports on these issues would most likely resolved
> expeditiously to everyone's satisfaction, and without any great
> controversy. The only potential problems I see are with details of
> syntax(*). I think those are best resolved within the HTML WG. I also
> suspect some of the comments may be issues for CSS, not HTML, for
> example
> the treatment of list markers. HTML completely defers to CSS on list
> rendering. Either way, it would be good to identify those kinds of issues
> ASAP rather than continuing to develop in a silo.

We have already had some discussion on that point in the group and with
Fantasai, and have in mind to more carefully consider ramifications for CSS
and XSL-FO in the near future.  Aharon wanted to focus on HTML for this
particular document in the interest of getting the information to you asap.

> If the I18N WG would like these issues addressed before Last Call, I
> strongly recommend delivering the feedback to the HTML WG as soon as
> possible, ideally in the form of bug reports, one per distinct issue.

I think we will be happy to submit the problem statements and proposals in
the form of bug reports.  We need to still confirm the expectations for some
of those proposals with the intended user base, and there are some
outstanding issues to resolve that also depend in particular on the
requirements of the user base, and it is on ironing out those points that we
expect to be busy during the near future.  I had assumed that it would help
the HTML folks to clarify those ideas before asking you to look at it,
therefore removing unproductive cycles, but we can consider raising the bug
reports sooner rather than later.

Note that we have not *just* produced bug reports because we see the value
of the document as being wider than just HTML.  Although I have not pushed
for the document to widen its current clear coupling to HTML (partly because
of it's recent origins and partly so that we can intersect more quickly with
HTML), I see a great deal of relevance to other specifications such as SVG,
Widgets, VoiceXML, OpenDoc, DITA, etc, and in fact XML markup in general
(and as such, a useful addition to any future ITS Rec). As you mentioned,
there is similar work to be done wrt style sheets too.

> Regards,
> Maciej
> * - Examples of potential syntactic quibbles: (1) It would probably be
> better for the "bdi" attribute to act like a normal HTML boolean attribute
> where only presence or absence is relevant, not the value; (2) the name
> "bdi" is a bit obscure for a global attribute and may be prone to typos;
> "submit_dir" does not match the usual conventions for HTML attribute
> I raise these not to suggest changes to the draft but rather to point out
> that the HTML WG needs to review these issues and should do so ASAP.

I hope that make things clearer for you.



Received on Monday, 8 March 2010 10:53:26 UTC