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: Fri, 03 Jan 2003 13:56:00 -0600
Message-Id: <4.1.20030103132706.018f2230(null)>
To: Ian Hickson <ian@hixie.ch>
Cc: www-style@w3.org

Hi Ian,

I've decided to reply to your private email on the list, because I think my
reply will fairly summarize our disagreement in a fundamental way.  I've
deleted what you wrote to me privately to give you the choice if you want
it public or not.

Apologies in that I violated my promise not to respond to you, but I think
this clear statement of our disagreement will clarify for the list all the
verbosity we did.

In your very first post you could have said that "my disagreement is that I
believe that semantics is 100% defined by specification".  Or at least
instead of asking me to make definitions when you already knew what your
definitions are.

That would have avoided a lot of pain and misunderstanding, which to tell
you frankly has made me angry at you at times in the debate.

Regarding that fundamental disagreement we have over what determines
semantics, I can assure you that implementation controls the semantics of
the web as much as specifications do.  It is a reality you might wish to
change, but you never will, because of the laws of nature.

I give you one last example.  If one renders a paragraph by dispersing the
characters randomly on the screen in random positions, no doubt that is a
presentation change, but I also argue it is a semantic change because the
paragraph can not be read any more.  It is quite different from changing
the font of a paragraph.

I understand your point that the semantics haven't changed in the sense
that just because the characters are rendered randomly dispersed, doesn't
change the meaning of what paragraph is _supposed_ to mean.

However, "_supposed_" is the operative word.  If 99% of browser clients
render the paragraph as dispersed, then you can be well sure that
implemention overrules the specification.

Would you call that a "bug" or would you say that it is the right of
presentation?

Regardless of what you call it, the reality can not be changed.
Implementation (interpretation of the semantic specification) plays a role
in the actual semantics that authors expect from markup.  The expected
meaning is the actual meaning.  As I said, specification is not the final form.

I realize we disagree.  You think specification is the final form.  But it
would be much better for the list if you had more directly gone to the
point of your disagreement.

You think specification is the final form.  This frees you to merge all
implementation (even what I consider semantic implementation) into the
style layer.   I strongly disagree with this and think it is better to keep
the semantic implementation above the markup parser, so that semantic
implementation (or semantic binding) is separate from style and
presentation implementation.  "Semantic implementation" (or binding) being
for example binding a new tag to implementation, as that determines
basically what the tag means.  One can separate the non-semantic specific
implementation (e.g. style) from the semantic implementation.  CSS does a
fairly good job of that.

Unless I have mistated our disagreement, I do hope this is end of our
debate.  We merely disagree on the reality of semantics and the need to
keep the markup semantic layer separate from the presentation layer.  Well
more accurately, we disagree on the defintion of what is the semantic
layer.  You say I am changing my terms, and I think you are.  It is because
we have an entirely different basis of understanding as to what really
constitutes semantics.  You hold a very pure view and I hold a realistic view.

But for sure you can agree that CSS group should not be involved in markup
semantic issues.  The thing is you and I do not agree on what markup
semantics is.  You say it is only specification and this provides broad
freedom to put all implementation in the style layer.  I prefer the CSS
group not try to swallow semantic properties that it is not expert at.

-Shelby Moore
Received on Friday, 3 January 2003 14:54:56 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 27 April 2009 13:54:19 GMT