W3C home > Mailing lists > Public > w3c-wai-gl@w3.org > January to March 2000

Re: A "one size fits all" personalized web page?

From: Charles McCathieNevile <charles@w3.org>
Date: Thu, 13 Jan 2000 11:28:48 -0500 (EST)
To: William Loughborough <love26@gorge.net>
cc: Scott Luebking <phoenixl@netcom.com>, w3c-wai-gl@w3.org
Message-ID: <Pine.LNX.4.20.0001131112270.10010-100000@tux.w3.org>
So we get down to the question of what is the best solution to provide.

It is clear that users are happy to have a presentation that suits their
needs. It is equally clear that they are concerned about having something
whose semantic content is different from that supplied to someone else,
simply because that is how their needs are perceived. And finally I think it
is pretty clear that telling a user to go write their own style sheet is not
going to be a solution for the current generation of HTML, and is not likely
to become feasible until the so-called semantic web becomes a reality, and
machines are able to understand more complex metadata about a  page and
relate it to what the user has said they like and dislike.

Scott has suggested using dynamically generated web pages which are
customised to suit different sets of needs. In fact there is value in this
approach in circumstances where all else fails, but this case does not seem
to be one of those. (For an example of one, read
http://www.sidar.org/anima.htm - if you don't read spanish go to
http://babelfish.altavista.com and ask it for a translation - that is not
brilliant but makes some sense by the second read)

I would suggest instead encoding the information with all the known semantics
once, and making different stylesheets available. This is in fact why the
HTML specification reserved the value "alternate" for the rel attribute.
Unfortunately there are very few browsers around that fully implement HTML4 -
the only ones I know that make the claim are Ice and I believe Opera 4 will -
so it would be useful to provide a change-stylesheet link for people to swap
between different styles. This will provide everyone with the assurance that
they are getting the same content, and the semantics are all there if they
want to look into them, while at the same time allowing for customisation.
Further, it allows advanced users to enhance the customisation by adding
their own stylesheet.

I believe that a server-side provision of a stylesheet is likely to be less
burden than rewirting the entire pages model, improves the separation of
content from presentation to allow the latter to be adjusted as much as
possible, and incidentally reinforces to the author which is which, since
they need to determine the circumstances in which their presentation is
actually used to convey semantics (the order of things in a table is a good
example). It also provides all semantic information as a first-class object,
accessible through the web, rather than hiding some of it in a database
system that is invisible to the end user.


Charles McCN

On Wed, 12 Jan 2000, William Loughborough wrote:

  SL:: "Basically, user-side CSS is not a reasonable solution for the
  average blind user with limited computer technology."
  WL: It would seem that whatever server-side or required-of-author
  accommodation is fully as unreasonable in terms of probability of
  implementation. I don't think authors are as likely to undertake the
  sorts of things Scott is recommending as is implied by his proposals.
  The chance of getting semantics communicated is clearly: under the
  control of the author; doable by various strategies. If we urge the
  author to follow certain guidelines, it is possible for everybody to
  What I propose is that we find out if this is an unworkable strategy -
  so far I've seen nothing to prove that. Scott's letter to incoming
  freshman might be paraphrased as a letter to Web authors proposing that
  they must conform to guidelines/universal design vs. learn how to design
  particular variations for various situations. I'm sure the argument for
  people reading through a straw would differ from that for people who
  couldn't read - etc., etc. 

Charles McCathieNevile    mailto:charles@w3.org    phone: +61 (0) 409 134 136
W3C Web Accessibility Initiative                      http://www.w3.org/WAI
21 Mitchell Street, Footscray, VIC 3011,  Australia 
Received on Thursday, 13 January 2000 11:29:02 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 16 January 2018 15:33:31 UTC