1. Like HTML, the specification of the SGML Disabled Access attributes
is also evolving. The work of the European TIDE project (which
uses the ICADD work as the basic of a workstation for displaying
braille, computer voice and large type versions of documents based
on the SDA attributes) has resulted in the addition of a new
attribute to the SDA set. This one carries what is essentially
"help" information and provides a place to explain what something
is when a user queries (in places where the generated text doesn't
explain enough). There may be two or three places in HTML2.0 where
that is necessary. I will make these enhancements this weekend.
(This information is not used in the transformation process; only
in ICADD browsers which work from the DTD.)
2. While the final editorial work was going on, Jeff installed a
server for doing automated transformation of HTML2.0 conforming
files into ICADD-marked up files at UCLA. As of two weeks ago,
well over a 1000 files had been transformed, and some of these
were at a fixed time every day, suggesting that some people are
setting up batch style jobs for this service, which is wonderful.
But it's clear from that work that the heavy reliance on expert
intervention (implied by the heavy use of SDATRANS markers in the
transformed output) is less appropriate for HTML than it would be,
for instance, in the case of textbooks loaded with math and tables.
That is, we feel that we need to remove these and make the
tranformation into ICADD from HTML much more automated than might
be the case in a production shop. If services such as UCLA's are
to be useful, we feel, the transformation has to work for the
individual visually disabled person who simply wants to read any
of the HTML that's out there without expecting an expert to get
involved. If there are slightly misleading transformations some of
the time, they are a small price to pay for the service.
Jeff and I and other committee members have agreed on a cleaned
up set of attributes which I've proposed and I will add them to
the base line DTDs this weekend.
This may end up effecting less than ten of the SDA attributes
now in the spec.
My draft piece of accompanying documentation is being edited to
match by Bart Bauwens of the Catholic University of Leuven (who
was heavily involved in the enhancements to the SDA methods based
on the Leuven/Sensotec work in Belgium). I hope to have this
available by the end of the weekend too.
Yuri
---------------------------------------------------------------------
Yuri Rubinsky +1 416 239-4801
Member, ICADD (800) 387-2777 (from US only)
President, SoftQuad Inc. uucp: {uunet,utzoo}!sq!yuri
Suite 810 56 Aberfoyle Crescent Internet: yuri@sq.com
Toronto, Ontario, Canada M8X 2W4 Fax: +1 416 239-7105