W3C home > Mailing lists > Public > w3c-wai-gl@w3.org > April to June 2005

Definition of "semantics"

From: Joe Clark <joeclark@joeclark.org>
Date: Sun, 12 Jun 2005 21:53:53 +0000 (UTC)
To: WAI-GL <w3c-wai-gl@w3.org>
Message-ID: <Pine.LNX.4.60.0506122118350.20743@aristotle.multipattern.com>

In a recent phone call, I volunteered to come up with a definition of 
"semantics" as used in the domain of Web standards.


First of all, I don't want to hear any more objections that "semantics" is 
already in use in general English and in the domain of RDF, hence we 
cannot use the term ourselves. Terms can be polysemous, that is, have 
multiple meanings, and polysemous terms are actually quite common in W3C 
documents. The UAAG glossary lists three senses of "content" and two of 
"user agent," for example. You'd be surprised how many polysemous terms 
are listed in the W3C Glossary and Dictionary (which actually exists).

Next, I strongly urge the Working Group to cease dismissing the outside 
expert opinion I go to some length to collect for you. Almost nobody in 
this esteemed Group knows the first thing about Web standards, which is 
such an established trend in Web development that a stack of books, 
including mine, have been published about it. I'm sure you don't like 
hearing that, but it is nonetheless true. (And one of the members who does 
understand Web standards is quitting in three weeks.)

Since the Working Group has a years-long track record of *immediately* 
putting through whatever certain preferred persons propose but ignoring 
whatever I propose, I thought I would avoid the whole issue by canvassing 
the opinions of outside experts whose stature is not in dispute outside 
the Working Group.

And you pretty much ignored that, too. The following is an accurate 
paraphrase from the mini-telecon Gregg, Becky, and I had on May 26:

 	ME: I sent the advice of outside experts to the list.
 	GREGG: And the group didn't accept it.

I suppose it's hypothetically possible that a group of established experts 
could be wrong about their own work and their own disciplines while the 
Working Group might be right, but that isn't where I'd place smart money.

So: To fulfill my action item, I passed around a pad of paper to a group 
of nearly a dozen standardistas who were in London for the @media2005 
conference. (We were having dinner along the banks of the Thames.) These 
are seriously qualified people, including Jeffrey Zeldman, Andys Budd and 
Clarke, Molly Holzschlag, and Derek Featherstone. You may not have heard 
of them, but that is a problem a moment's Googling will solve. (You could 
also buy their various books.)

I asked them to define "semantics" or "document semantics" as they 
understood the terms. There were a couple of humorous entries (including 
two that were drawn instead of written!), which I retain for the Working 
Group's amusement, but you can see a general clustering of definitions:

>   Doug Bowman
>         1. Described as appropriately and accurately as possible
>         2. Meaning and purpose
>   Jeffrey Zeldman
>          Line drawing shows smiling young man with sweat flying off his
>          face. The word `Tag' sits atop an arrow pointing to the man,
>          who utters the word `Meaning?' The illustration is signed with
>          a Z in a box
>   Carrie Bickner-Zeldman
>          Meaning
>   Derek Featherstone
>          The art of selecting the most appropriate tag to provide
>          meaning to content, and extending and redefining that meaning
>          when it doesn't already exist
>   Patrick Griffiths
>          42
>   [Unsigned]
>          A nervous twitch triggered by a discussion of acronym and abbr
>   Molly Holzschlag
>          The meaning of something. In the context of markup and CSS, the
>          meaning of the element or property in relation to the content
>          which it describes
>   Andy Budd
>          The ability to add meaning and structure to a document
>   Andy Clarke
>          The most appropriate use of markup elements (tags?) with a
>          focus on using markup to describe the structural architecture
>          of a document without reference to presentation
>          Drawing shows spiky-haired man smoking a cigarette and the word
>          balloon `Buggered if I know!'

Onstage two days later at the conference, the definition of "semantics" I 
proposed was approximately the following:

 	The choice of markup that corresponds to the purpose,
 	intent, or meaning of the content

(later misquoted on one Weblog)

A definition was already provided to this esteemed list by reference:


which referred to:

(by Molly Holzschlag, which I guess you didn't read)

which states: "in markup, semantics is concerned with the meaning of an 
element, and how that element describes the content it contains."

I'm not providing one single proposed wording, because you'll just shoot 
it out of the sky (Jason White especially). I'm proposing wording from a 
half-dozen serious experts. We can use that as feedstock to produce a 
finessed definition later.

It would be unwise to dismiss expert advice of this kind.


     Joe Clark | joeclark@joeclark.org
     Accessibility <http://joeclark.org/access/>
       --What's wrong with top-posting?
Received on Sunday, 12 June 2005 21:54:12 UTC

This archive was generated by hypermail 2.4.0 : Thursday, 24 March 2022 21:07:40 UTC