W3C home > Mailing lists > Public > www-html@w3.org > August 2005

Re: XHTML2: Proposal for total separation of semantics from structure

From: Orion Adrian <orion.adrian@gmail.com>
Date: Tue, 30 Aug 2005 15:23:56 -0400
Message-ID: <abd6c8010508301223e49dd2e@mail.gmail.com>
To: www-html@w3.org

On 8/30/05, Laurens Holst <lholst@students.cs.uu.nl> wrote:
> Orion Adrian schreef:
> >>> I disagree, for element names should carry the main semantics, and
> >>> prefer myself something like this:
> >>>
> >>>   <"some parent with only block children allowed">
> >>>     <code class="pre"/>
> >>>   </>
> >>>
> >> I disagree with that, class doesn't convey any semantic information, and
> >> if the code being preformatted is important to the clarity or even
> >> functioning of the code, it should use <pre> appropriately and not rely
> >> on the stylesheet.
> >>
> > Why doesn't class convey semantic information?
> Because it doesn't, per specification. Quoting:
>    The class attribute has several roles in HTML:
>        * As a style sheet selector (when an author wishes to assign
> style information to a set of elements).
>        * For general purpose processing by user agents.
> > Why aren't we styling
> > elements based on their semantic classification as opposed to string
> > that people just make up like "l76blue". I'd rather limit class to
> > semantic structures and just style those.
> >
> But that is not the case right now, in HTML 4, and existing content will
> contain nonsensical information in their class attributes, which makes
> them useless to process. A set of values for the class attribute could
> be standardised, but because there's so much 'noise', it can't be relied
> upon, which makes it quite unreliable and useless from a machine
> processing point of view (which is why they are added in the first place).
> I'd say that is the reason new attributes such as @role and @property
> are needed. Those can make a fresh start, and because they are only
> specified when they actually have semantic meaning, a user agent
> (including search engines) can do something with it...

Except both @role and @property are being used for something else.
@class is a field that lists the objects that meet the following

This element is-a object.
This element can-be-classified-as object.

@property is more open. Specifically it represents the relationship
with the contents providing the object value.

The fundamental problem with @property is that on elements that aren't
meta it doesn't allow for all the relationships that can exist for an
element. If you use it for an is-a relationship it doesn't cover the
whatever other relationships might pop up much less an additional
relationship other than is-a.

What we need is a way to represent triplets for an element (like meta,
but for any element). Class was useful, but what we need is a typed
version of class specifically for is-a relationships because they are
the most common for determining behavior and appearance.


Orion Adrian
Received on Tuesday, 30 August 2005 19:24:04 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 15:06:11 UTC