W3C home > Mailing lists > Public > public-html@w3.org > April 2007

Re: Level of specification detail (Was Re: Proposal to Adopt HTML5)

From: ryan <ryan@theryanking.com>
Date: Wed, 11 Apr 2007 11:07:34 -0700
Message-Id: <E13090CE-4112-4607-BC6F-54C9BBD6762A@theryanking.com>
Cc: Anne van Kesteren <annevk@opera.com>, Henrik Dvergsdal <henrik.dvergsdal@hibo.no>, public-html@w3.org
To: Dan Connolly <connolly@w3.org>

On Apr 11, 2007, at 8:24 AM, Dan Connolly wrote:

> There is a natural tension between writing the sort of precise,
> exhaustive specification that meets the needs of implementors,
> and writing for the audience of authors, who mostly need a bunch
> of examples to copy from and a bit of explanatory material.
> When we developed OWL, the Web Ontology Language
> (http://www.w3.org/2004/OWL/ ), we had enough
> writing resources to explain the language from 5 different
> perspectives:
>   Overview
>   Use Cases and Requirements
>   Guide
>   Reference
>   Abstract Syntax and Semantics
>   Test Cases

I don't think comparing HTML and OWL is very useful here. When you  
were working on OWL , (AFAICT) it was a new technology that no one  
had authored before. With HTML, we have a very different situation,  
we have thousands or millions of authors, hundreds of guides, entire  
sites dedicated to tutorials and instructional material, college  
level courses, etc., etc. People know how to author HTML as implemented.

What we lack is a specification for how to implement HTML user agents  
that is unambiguous, rigorous, complete and has a test suite.

I don't think this WG needs to worry about educating authors, instead  
it should focus, at least in the near term on bridging the divide  
between HTML "as specified" (HTML4) and HTML "as implemented" (which  
is what HTML5 is designed to do).

Just to be clear, I completely support the proposal that started this  

Received on Wednesday, 11 April 2007 19:02:56 UTC

This archive was generated by hypermail 2.3.1 : Thursday, 29 October 2015 10:15:18 UTC