W3C home > Mailing lists > Public > w3c-wai-ua@w3.org > July to September 2000

Re: Comments on presentation/structure related to issue 297.

From: Al Gilman <asgilman@iamdigex.net>
Date: Fri, 21 Jul 2000 22:15:18 -0400
Message-Id: <200007220210.WAA3218382@smtp2.mail.iamworld.net>
To: Ian Jacobs <ij@w3.org>, w3c-wai-ua@w3.org
At 08:31 PM 2000-07-21 -0400, Ian Jacobs wrote:
>   I would also note that this discussion sounds a little like the
>   discussion about content meant for humans v. content meant for
>   machines. Refer to Al's comments on this topic [7]:
>     "The distinction between data (raw content) and meta-data 
>     (markup) is an artifact of the view assumed by the author.  
>     There is no fundamental semantic difference between what is 
>     called data vs. metadata.  They both play the same role as 
>     bearers of information  Semantically, it is all just
>     one class of data.  This is a little-understood fact of 
>     information science."
>   Al, will you make the same comment about an assertion that
>   presentation v. non-presentation (or structure)? 

[not promising to have read the rest of the message...]

No.  These concepts are well-posed enough, even if HTML elements don't sort
neatly along these lines.

In fact, I would try to clarify my earlier remark.  The syntactic
distiction between markup and marked-up content is perfectly clear.  What I
was trying to say is that some people think of metadata as semantically a
different class of stuff from data.  That is what I was claiming is
illusory.  The meta-distinction is a byproduct of the view or packaging,
not in the semantic meaning of the data.  

In the case of HTML some markup [counting style rules in a stylesheet as
markup, loosely] is clearly presentational and some, such as DIV is clearly
structural.  Now, of course, there is markup like FRAMESET and TABLE which
are both.  This markup has both layout and structural semantics in one
element type.  So at least in HTML it is possible to talk about markup with
presentational semantics and markup with structural semantics, but it is
not possible to partition the set of element types into two groups which
are purely one and purely the other.

Hope this is responsive to your question.

Received on Friday, 21 July 2000 22:10:17 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 14:49:27 UTC