- From: Karl Dubost <karl@w3.org>
- Date: Thu, 24 Aug 2006 19:23:00 +0900
- To: David Woolley <david@djwhome.demon.co.uk>
- Cc: www-html@w3.org
Le 24 août 06 à 15:18, David Woolley a écrit :
>> - These semantics of meaning are useful
>> - These should not be element but role/property attribute values.
>
> I'd say that was in conflict with the original SGML concept, and
> actually
> represents a move towards a presentational language.
The original SGML specification is in front of my eyes right now and
nothing is said in this sense. But let's clarify a bit more, because
it's definitely not a move towards a presentational language.
XHTML 2.0 Specification abstract says it
[[[
XHTML 2 is a general-purpose markup language designed for
representing documents for a wide range of purposes across the World
Wide Web. To this end it does not attempt to be all things to all
people, supplying every possible markup idiom, but to supply a
generally useful set of elements.
]]] -- XHTML 2.0
http://www.w3.org/TR/2006/WD-xhtml2-20060726/
Wed, 26 Jul 2006 20:04:34 GMT
I will avoid to use word "semantics", because it has too many
understanding and will dilute the message. Let's keep this in mind:
"general-purpose markup language"
I think we can agree that the fields of *meaning* for text and
applications are wide, very wide. Each profession, each domain has
its own set of meaning, with sometimes overlaps.
HTML = Hyper Text Markup Language
Hyper stands for the link capability
Text stands for the text.
Since the beginning of HTML, we can classify the elements in 3
categories
- Structure: p, li, table, etc.
- Functionnal: a, form, link, etc.
- Meaning: code, quote, samp, address. etc.
The proposal here is not to remove "Meaning" but to move it from the
element to the "property/role" attribute, which have been conceived
for this purpose. Creating a specification which would cover all
possible meanings even only the most important, would be tiresome,
not achievable in a reasonable time and highly unsatisfying because
of different cultural needs and professional needs.
The block/inline model is part structural par presentational
(inheritage of the codex as in conveying structure by a specific
layout), CSS covers it which is right.
Keeping "Meaning" elements leads to problems of presentation
specifically:
blockcode/code
blockquote/quote
Duplication of "meaning" elements because of presentational aspects.
Keeping the "structure" element and at a certain degree "functional"
element meets exactly the general purpose nature which is stated at
the beginning. More than that, it makes it the ideal vehicle to
create "meaning" application of all kind without too much trouble.
> I don't really like role, because it seems to be an abdication of
> responsibility, which could result in every web site having its own
> definition of what role names are and what they mean.
Yes and that's good. Right now, an author has the choice of the terms
to be used in "class" attribute values. Standards are a community and
social process where a community agrees on the needs of this community.
Sets of values for "role" and "property" can be defined exactly in
the same way than the semantics of an element, either being "meaning"
or "structure". It's a question of defining them.
The proposal is to ship with XHTML 2.0 what was before "meaning"
elements has a standardized set of modules defining the values for
the attribute role and property. Exactly what WAI will do, and
exactly what XHTML 2.0 already does for certain values.
It is not an abdication at all, it is a proposal to give more
flexibility, to have specifications and associated schemas a lot
easier to maintain.
> I think I would much rather have first class elements, but if the
> elements
> aren't considered general purpose enough, there should be standards
> profiles defined, which, for example, had different profiles for say;
> e-commerce sites; corporate sites; informational sites; search
> engines,
> leaf node pure documents, etc. Each profile could be made of several
> modules, but authors would be expected to use a complete profile, not
> make their own selection from modules.
That is the proposal. Even better a professional domain could say we
need a standard set of values for our profession. Can we do it at
W3C? Yes! Do it.
> I'm not convinced this will work, though, because most authors seem to
> want to confuse CSS, EcmaScript, object models, and Flash into a
> single
> thing that they call HTML.
Most authors don't edit their page by hands, the authoring tools
(forms, converter, or HTML authoring tools) do. It's a lot easier to
add set of defined values than having to deal with adding new
elements for each version of the language.
I hope I made the proposal clearer.
--
Karl Dubost - http://www.w3.org/People/karl/
W3C Conformance Manager, QA Activity Lead
QA Weblog - http://www.w3.org/QA/
*** Be Strict To Be Cool ***
Received on Thursday, 24 August 2006 10:24:05 UTC