W3C home > Mailing lists > Public > w3c-wai-ig@w3.org > July to September 2002

RE: SCOPE vs. ID/HEADER

From: Jim Thatcher <jim@jimthatcher.com>
Date: Thu, 12 Sep 2002 10:45:06 -0500
To: WAI IG <w3c-wai-ig@w3.org>
Message-id: <NDBBKJDAKKEJDCICIODLMEDOEJAA.jim@jimthatcher.com>

Al Wrote:

> What I don't understand is what the cases are where SCOPE adds
> value, between tables handled fine by the first-row, first-column
> heuristic and the cases where HEADERS markup is required to make
> the connections right.

You read the table of "tests" correctly. The best thing today is to put
headings in row one and column one as measured by assistive technology
support. I tried to be careful and say, "for programmatic identification"
use scope and TH. There are places where the scope "adds value" and one is
in the corner, like row 1 column 1. With TH there and without scope, you
can't tell whether that heading cell is a column heading or a row heading.

I hadn't seen Nakata's Power Point Slides before, Thanks. They (Doug
Wakefield and Ken Nakata) recommend the same as I said in my post - use TH
and scope for simple tables and when the table is complex (where TH and
scope are inadequate) then use headers/id.

Jim
Accessibility Consulting
http://jimthatcher.com
512-306-0931
Evaluation of web evaluation tools at http://jimthatcher.com/erx.htm


-----Original Message-----
From: Al Gilman [mailto:asgilman@iamdigex.net]
Sent: Thursday, September 12, 2002 9:48 AM
To: jim@jimthatcher.com; WAI IG
Subject: RE: SCOPE vs. ID/HEADER


At 10:19 AM 2002-09-12, Jim Thatcher wrote:

>I fear, Steve, that if you ask 5 different experts you'll get 5 very
>different answers.

The answers may not be so different as they may sound.

Let me hasten to say that what we were dealing with primarily
in the FedStats meeting would come under Jim's 'complex' category,
where HEADERS is the only way [at present] to ensure service
to the customer.

What I don't understand is what the cases are where SCOPE adds
value, between tables handled fine by the first-row, first-column
heuristic and the cases where HEADERS markup is required to make
the connections right.

>Below I have pasted a listing on how HPR, Window-Eyes and JAWS support
>various forms of table markup - these results are for versions a few months
>ago. The bottom line is that for programmatic identification of table
>headings, TH with scope is best because it is simplest and is supported as
>well as headers and id's. If the table is complex ("more than one Logical
>level of row and column headers") then your must use headers and id's
>because headings will not be properly identified with TH and scope.

>The chances of errors and the time required to touch each cell of a data
>table when using headers/id leads me to recommend that web developers
>revisit the design of the table; probably it can be simplified, possibly
>with multiple tables.

At the FedStats Workshop there were multiple vendors showing tools
that reduce the time and errors for this considerably.

>The versions for this test are JFW 4.01, HPR 3.02 and Window-Eyes 4.11.
>1. No special markup - row one and column one for headings: JFW,
>Window-Eyes, HPR.
>2. Scope attribute on either TH's or TD's - JFW and HPR.
>3. Axis attribute on TH's or TD's - JFW.
>4. Use TH for headings - JFW and HPR.
>5. Headers attribute with id's specifying heading cells - JFW and HPR.
>6. abbr attribute on heading cells - JFW and HPR.
>
>Jim
>Accessibility Consulting
>http://jimthatcher.com
>512-306-0931
>Evaluation of web evaluation tools at http://jimthatcher.com/erx.htm
>
>
>-----Original Message-----
>From: w3c-wai-ig-request@w3.org [mailto:w3c-wai-ig-request@w3.org]On
>Behalf Of Al Gilman
>Sent: Thursday, September 12, 2002 7:25 AM
>To: WAI IG
>Subject: Re: SCOPE vs. ID/HEADER
>
>
>
>
> >On Thu, 12 Sep 2002, Steve Vosloo wrote:
> >
> > >I've always used ID and HEADER attributes when creating accessible data
> > >tables, but after reading the Access Board's bit for 508 compliant
> > >tables, I might user SCOPE instead. From what they say SCOPE is as
> > >widely recognised by user agents and it is definitely easier to
> > >implement.
> > >
> > >Any thoughts on why we don't all simply use SCOPE and code our tables
in
> > >a tenth of the time?
>
>Use ID and HEADERS.  SCOPE is still not implemented widely enough to merit
>your relying on it at this time.
>
>This was the report of Ken Nakata and Doug Wakefield to the
>FedStats workshop of accessible tables as recently as the end of June.
>
>   http://workshops.fedstats.gov/Nakata_Fedstats.ppt
>
>   http://www.w3.org/2002/Talks/06/24-US_FedStatsWorkshop/slide1-0.html
>
>   http://workshops.fedstats.gov/FS508Workshop.htm
>
>The 508 materials you refer to are older, hopeful, and not yet valid as
>reflecting implemented technology.  You should regard what Ken and Doug
>reported to the FedStats workshop as a clarification with regard to the
>earlier technical information from the Access Board.
>
>Al
>
>At 04:31 AM 2002-09-12, Charles McCathieNevile wrote:
>
> >On Thu, 12 Sep 2002, Steve Vosloo wrote:
> >
> > >I've always used ID and HEADER attributes when creating accessible data
> > >tables, but after reading the Access Board's bit for 508 compliant
> > >tables, I might user SCOPE instead. From what they say SCOPE is as
> > >widely recognised by user agents and it is definitely easier to
> > >implement.
> > >
> > >Any thoughts on why we don't all simply use SCOPE and code our tables
in
> > >a tenth of the time?
> >
> >Browser makers implemented headers first, so that is what people started
> >recomending. When browser makers didn't implement scope people got
laughed
>at
> >for suggesting using it (understandable really) with the result that
people
> >who write sites got given the hard way first, and had to wait for the
>simple
> >solution to be useful.
> >
> >Unfortunately (but again understanably really) users don't change their
> >systems very often. So you don't get to take advantage of the newer
>features
> >(well in terms of implementation - these were all proposed in '96 or '97)
>to
> >make your life easier until users go through the process of upgrading,
>unless
> >you are happy leaving a whole lot of users in the cold.
> >
> >Chaals
Received on Thursday, 12 September 2002 11:46:47 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 19 July 2011 18:14:06 GMT