W3C home > Mailing lists > Public > w3c-wai-hc@w3.org > October to December 1997

meeting happenings

From: Al Gilman <asgilman@access.digex.net>
Date: Fri, 14 Nov 1997 21:30:51 -0500 (EST)
Message-Id: <199711150230.VAA01937@access5.digex.net>
To: w3c-wai-hc@w3.org (HC team)
Cc: jbrewer@w3.org
A lot has happened in one short week thanks to the face-to-face
meetings in Austin.  On the whole, this is good.  Let me try to
bring the whole working group up to speed on what has been
happening.  There will be minutes from the HTML and Styles
Working Groups, and fresh drafts.  But let me try to outline the
major points.

First HTML.

The final meeting of the HTML WG as we have known it happened
Tuesday afternoon and Wednesday.  Tuesday morning included a
review of the WAI proposals.  Daniel and I represented the HC
team in this review with backup on policy and W3C issues from
Judy Brewer.  Gregory Rosmaita sat in and observed for part of
the day.

In the course of this review we agreed to some changes in the
proposals, but the overall level of capability is comparable in
benefit.

We reduced the OPTGROUP capability to support only one, not an
arbitrary number, of levels of grouping.  There will be OPTGROUPs
containing OPTIONs but not containing nested OPTGROUPs.  On the
other hand, the OPTGROUP syntax was retained as opposed to using
a GROUP attribute (analogous to AXIS) in OPTIONs.  The rationale
for staying with the OPTGROUP container syntax is that it is
believed to be easier for authors to understand and hence will be
more likely to get used.

The rationale for retreating to one-deep grouping is that the
vendors, led by Scott Isaacs, convinced us that they are eager to
get on with a more ambitious hierarchical-list proposal in the
near future but don't have time to work out the details in time
for 4.0.  This decision was also made easier by input from the
WAI representatives that having multi-level nesting capability is
not a show-stopper from an accessibility perspective.

Another change where at first blush one might feel as though we
gave ground is in the area of maps for OBJECTs.  We exchanged the
"shapes" attribute on OBJECT for the capability to put general
block-level HTML in a MAP.  This means that if an OBJECT contains
an image with sensitive regions, that one should use the USEMAP
to link to a MAP which gives the sensitive regions either by AREA
elements as at present or else by putting shape attributes on
Anchor elements in free-form HTML.  The structured-text
alternative capability of this design is very similar to what we
would get with "shapes" on OBJECT.  The sense of the meeting was
that this approach will be easier to explain to authors and to
get them to use.

The Proposed Recommendation is up now for public review and
member voting.  The final language of the specification is
amenable to adjustments by Tim Berners-Lee, the Director of the
W3C, in consideration of comments received during this voting
period.  The adjustments that we agreed on with the HTML Working
Group will go to Tim as part of a package of technical correction
comments from the Working Group.  You will not see changes in the
public Proposed Recommendation; there will be a Working Group
draft of the document as amended with these Working Group
comments that the HC team members can and should read to check
the details.

Those two areas are the only substatial changes.

As I mentioned in the GL discussion, the ACRONYM element was
re-insered in addition to adding ABBR.

Our position on the Q tag has been accepted as the HTML working
group position.  This was not discussed in the face-to-face
meeting as it is viewed as small enough to be nearly editorial in
nature.

Now, CSS2

FP caucus, Monday afternoon:

I reviewed a draft HC position with a breakout session held
concurrent with an Education and Outreach session.  The material
as it existed going into this caucus, with some annotations based
on the Monday caucus, is on the HC web site at

 CSS2 Status from a WAI(HC) Perspective
 http://www.w3.org/WAI/group/HC/css_austin.html

The Styles Working Group met Thursday and Friday, with
accessibility issues on the agenda Thursday.  Judy, Daniel and I
participated Thursday, up until the afternoon break when I left:

I presented the same transparencies as Monday afternoon with the
changes added by means of the oral discussion.

The upshot of the discussion was roughly as follows:

Media types:

Not discussed other than we said "change to be compatible with
the HTML specification."  

Generated text:

Generated text is going to be in CSS2 and not wait for CSS3
because of its important in accessibility scenarios.

One will be able to place content into the rendered document at
the start and/or end of an arbitrary element.  These inserts can
contain literal text and the text value of any attribute of the
current element, along with relevant numbers.  The question of
possible number representations was deferred to an
internationalization discussion late Thursday or today.  The
exact numbering capability was not nailed down as of the time
that I left but it seemed clear that there will be one.

The inserted text will be able to be styled in limited ways to
set it off from the author's text.  Style constructs that would
tend to break all association between the inserts and the other
content of the element they mark (such as "float" and fixed
positioning) will not be allowed.

Re-flowing content:

There is a concept of moving things like navigation bars around
the document in different media.  The "minimize" display
directive is targeted to this application, as well as a concept
of "positioning" in the audio stream.

As of the time I left it appeared quite clear that we will _not_
be able to write the SPEAKROW etc. functions in CSS2.  The
discussion on this point dealt with understandings about the
proper scope and processing difficulty for stylesheets.

The implementors feel that when the CSS work was started, a
commitment was made that styling an element would not require a
full tree walk but only excursions along the vertical thread from
the current element up to the root.

Another factor has to do with the CSS2 schedule.  It became
sufficiently convincing to Daniel and me that insisting on this
capability would force an unacceptable delay.  We followed a
position that it was more important to get CSS2 out by say,
February without the more advanced macro capability, than to
precipitate another quarter's slip.

There was some discussion to the effect that "this sounds like a
DOM function" which we will have to follow up on in the PF
context.

A new CSS2 internal draft may be available as soon as Monday.  We
definitely need to keep reading the drafts and considering how
they would be used in Braille and other accessibility scenarios.

Based on how things were going in this meeting, however, it would
seem we can expect to have functionality comparable to what I
have described here and that this meets the most critical needs
we have identified.

-- Al Gilman
Received on Friday, 14 November 1997 21:31:10 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 19:56:12 UTC