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

Scott,

Can you please provide a reference to Tim's statement to which you keep
referring. It seems to me that the statement is not really correct, at least
in the context here. It is possible to extract sematic inormation from
properly-coded HTML, since it contains almost nothing but semantic
information.

I think you are misunderstanding my point about the web, which is that it is
not a good solution to meet one group of user's needs by providing a differnt
presentation and at the same time ignoring the needs of most users. There is
less conflict in user needs than you seem to be making out. The conflict is
between user needs and designers who are suficiently obdurate to ignore them
for the sake of what they think is visually appealing, which is another
matter entirely.

I am not arguing against pages that can be presented according to user
preferences, I am simply arguing that the approach which says it is alright
to have a page that is badly designed and then have an alternative page for
some people is not a good approach, and saying itis preferable is to say that
it is a good idea to keep something which prevents or reduces access. In
particular where there is a provsion for one group of disabled users biut
others are then comletely ignored, it does not seem to even be an effective
approach to accessibility.

Cheers

Charles McCN

On Wed, 29 Dec 1999, Scott Luebking wrote:

  Hi,
  
  A limitation of HTML, as Tim Berners-Lee has pointed out, is that
  it is hard for software to extract semantic information out of the HTML
  once the HTML is created.  This means that any desired transformation
  based on semantic information will be very difficult.
  
  When you're saying that you believe "the opposite is true", I believe you
  are thinking in terms of non-semantic transformations.  The problem is that
  semantic-based transformations will often can give better results in
  a number of situations.  However, since HTML often cannot be transformed
  on a semantic basis, using HTML as for information representation
  has its limitations.
  
  I believe you might have misunderstood my point about buildings.  I'm saying
  that user needs in architecture are often less at a conflict.  So, it
  is easier to have a "one size fits all" in a building than in  a user
  interface.
  
  While I agree that the decoupling is a very useful tool, HTML is not the only
  way to decouple.  A database of HTML fragments is another way to decouple.
  XML is another way.  It is important to keep in mind the range of ways to
  achieve decoupling.  (Abstracting is another term to call decoupling.)
  
  Also, I'm more concerned about imposing the "one size fits all"
  approach on such web pages like personalized web pages, search engine
  results, etc, where the web pages are generated dynamically.
  While the "one size fits all" approach has good technical reasoning for
  stored web pages, there is not the same strength of reason for dynamically
  generated web pages.
  
  Scott
  
  
  > > The analysis that "one size fits all" means "not quite right for anyone" is
  > > exactly right.
  > 
  > But surely this is why HTML tells us to mark up our content honestly
  > for what it is, rather than how we hope it to be presented; and
  > delegates the presentation to the client agents (browsers etc.), with
  > or without suggestions from stylesheet(s).
  > 
  > >  Here's where information engineering and building architecture
  > > are different.  In architecture, the concept of "one size fits all"
  > > is more likely to succeed
  > 
  > I'd be inclined to say the exact opposite.  With the WWW, a reader is
  > presented with logical content, and they have at their disposal a
  > whole range of client agents to help them to present the content in
  > ways that are congenial to their requirements[1].
  > 
  > > since the needs of the various users are less likely
  > > to be in conflict or can be resolved.
  > 
  > Why then are we "normals" not supposed to use the same building
  > facilities as the disabled users?
  > 
  > >  Information engineering and HCI
  > > are very different than architecture.  To impose an approach from architecture
  > > onto information engineering and HCI and expect the same results is
  > > probably not reasonable.  It overlooks the many subtleties of how people
  > > interact with information.
  > 
  > In a WWW context I've come to believe that there are many, many issues
  > that favour the decoupling of honest content-based markup from (one or
  > more) specific presentation suggestions, i.e in the WWW situation via
  > optional stylesheet(s) (which may be actioned, or ignored in part or
  > whole, at the reader's discretion).  The different physical abilities
  > of various users are one set of issues.  The "disabling" effect of
  > browsing situation (to take only one example: a vehicle driver who is
  > rightly forbidden to examine a visual display while controlling their
  > vehicle) leads to similar conclusions as far as authoring is
  > concerned.
  > 
  > > Is the stance of the guidelines that "one size fits all"? 
  > 
  > I don't believe so, but logical content has to stand for itself. Some
  > kinds of content are inaccessible to some users, whether one likes it
  > or not.  Reducing General Relativity (which I don't comprehend, so I'm
  > taking this as a real example) to terms which every inhabitant of
  > Earth could handle, and upon which each of them could base new and
  > testable predictions, would be a rare talent.
  > 
  > [1]Which is not for a moment to say that authors could not make some
  > extra effort to assist accessibility to unusual browsing
  > situations.  But with all too many web pages today, we see that the
  > author has gone to considerable _extra_ effort to sabotage the
  > accessibility that was designed into the WWW from the start.  And then
  > they complain that it would be too expensive to add
  > accessibility.  [Expletive deleted] of course it's doubly expensive if
  > you start with expensive effort to make something impossible, and then
  > add more effort to restore what had been there in the first place.
  > 
  > seasons greetings to all
  

--
Charles McCathieNevile    mailto:charles@w3.org    phone: +61 409 134 136
W3C Web Accessibility Initiative                    http://www.w3.org/WAI
21 Mitchell Street, Footscray, VIC 3011,  Australia (I've moved!)

Received on Wednesday, 29 December 1999 21:14:43 UTC