W3C home > Mailing lists > Public > www-style@w3.org > January 2003

Re: XBL is (mostly) W3C redundant, and CSS is wrong W3C layer for semantic behavior *markup*

From: Shelby Moore <shelby@coolpage.com>
Date: Wed, 01 Jan 2003 17:11:04 -0600
Message-Id: <4.1.20030101163858.0387b4d0(null)>
To: Ian Hickson <ian@hixie.ch>
Cc: www-style@w3.org

At 06:23 PM 1/1/2003 +0000, Ian Hickson wrote:
>Before going any further with the other points of this discussion, we
>have to clear up the term "semantic binding", as it seems to be the main
>point which we disagree about.
>I don't understand what you think the difference is between a semantic
>binding and a non-semantic binding.
>Could you explain this?
>A straight forward definition of "non-semantic binding" first, followed
>by a straight forward definition of "semantic binding", would be most

Okay I respect you for responding this way.

Hopefully you will try to get at the "meat" of what I write below, and not
just twist some fuzzy aspect of communication.  Please argue against the
central themes I am asserting.  Which is what you are apparently doing with
this response.  That is productive.  Thanks.

Before I attempt a definition, let me also point out that one aspect of my
goal is to keep the semantic dependencies out of the CSS layer, so that you
and others here can continue doing your excellent work on presentation
hints, without accidentally getting CSS bound up in restrictive or brittle
knots later.  So it is very discouraging when I try to make a point and
people think I am threatening their career or status.  If anything, I have
always helped the people around me have better careers and more success, or
had no impact.  I am rarely involved in a product that failed in market
(only Art-o-matic because no market for 3D).  Truth always prevails in the
end.  As an example, I got lamblasted about 2 - 3 years ago because I told
Rasmus (the creator of PHP) that missing the OO class destructor was
crucial to making PHP useful in OO applications.  You will see that Zend 2
is in alpha now and addresses those concerns "because it was necessary to
support GTK+ and DOM in PHP".  I tend to say things that people are not
ready to hear. And I tend not to get credit later when my assertions are
proven true, but that is okay for me as long we get progress we need.

I do not disrespect your status or contribution here at W3C.  But if
someone disrespects the truth, then I will usually fight for the truth.
When I am wrong, I admit it.  For example in the sentence spacing thread, I
admitted to you (and list) that I was wrong about the ability to __easily__
parse "what is a sentence?".  You won that aspect (you were correct) and I
also made useful points about possibilities forward.  The one thing that
will really burn my blood is a hipprocrit (I am not saying you are).
Usually the people pointing the finger, are guilty of the same thing they
are accusing the other person.  And I am also guilty of that here in this
thread, by not ignoring some emotional things.


The way a semantic markup element is bound to the implementation which
defines it's semantics.

There will other bindings to semantic markup which do not define semantics.
  For example, there are many implementation (presentation) details (CSS)
which do not affect the semantics of an element.  For example, the font of
a <P>.  That is not "semantic binding" because the "hints" do not affect
the semantics of the element.  If ever we make a ::sentence property in
CSS, then that will not affect the semantics of a <P> tag.  Those are
presentation issues that are orthogonal (separable) from semantic details.

And there is a basic core set of implementation that defines the semantics
of an element.  For example, subclassing an <A> tag as you did, such that
it no longer links arbitrary markup, but assumes that it's content is a
date.  That is a significant change in the semantics of the content of an
<A> tag.  Before it was content "as marked up", then it is changes to
content is "a date no matter what is marked up".

The major reason for keeping semantic binding orthogonal to presentation
hints, is so that "least common denominator" effect is avoided.  We do not
want semantics reduced to the level of presentation, because then the power
of meaning is mitigated.  Your <A> subclassing is an example.  The semantic
layer lost it's ability to have markup knowledge within the content of the
<A> tag.  Again I *sincerely* thank you for making an example which made
the point so easy to explain.

Received on Wednesday, 1 January 2003 18:17:10 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 27 April 2009 13:54:18 GMT