WebAIM - Web Accessibility In Mind

E-mail List Archives

Re: Word documents and tables

for

From: Michael.Moore
Date: Oct 27, 2010 12:57PM

Hi Cliff,

Thanks for the questions, also I need to clear up one item that I got wrong yesterday.

Clearing up the mistake: If you have more than one table in work then you need to use "Title1, Title2..." I added a dot between the bookmark name and the number yesterday, which will not work. It is something that you do with cell names in Excel when you have multiple worksheets in a workbook. The differences between the programs can get confusing, which is why we have to test every document after we implement the hacks.

Now to answer your questions.

What's the difference between a hack and standard practice?

In my mind a hack is taking advantage of a feature or a bug to accomplish something that is not supported natively. Bookmarks and named cells which can be exploited to make JAWS recognize row and column headings in Word and Excel are intended to provide support for formulas and hyperlink anchors. A standard practice would be for Word and Excel provide a programmatic method of identifying table headers similar to the TH tag in HTML and PDF preferably with header and id support as well. In word styles are used to identify heading elements and this works well with most commercial screen readers and it makes semantic sense.

What would it take to get other screen readers to follow JAWS in this respect?

The other screen reader manufacturers would probably need to make sure that implementing that feature would not violate a Freedom Scientific patent or license the feature from Freedom Scientific. If you don't think this is an issue read article linked below from NFB. <http://www.nfb.org/images/nfb/Publications/bm/bm09/bm0911/bm091107.htm>; Even if this were implemented it does not get around the hack issue. Bookmarks and Cell names are just not semantically correct elements for creating row and column headers.


What would it take to get Word and other word processing applications to robustly support this technique?

Implementing the bookmarks and cell names in the same manner that they are implemented in Word would probably involve software patent issues. Perhaps someone from MS could speak up about this. Much better would be to extend the XML specifications for MS Office docs and Open Office docs to support semantic tags that differentiated between data cells and header cells and provide support for header and id attributes. Implementation for the AT developers would then be similar to support in PDF and HTML.


Mike Moore
(512) 424-4159