Re: The DOM is not a model, it is a library!

----- Original Message -----
From: David Brownell <david-b@pacbell.net>
To: Michael Champion <michael_champion@ameritech.net>
Cc: DOM Mailing List <www-dom@w3.org>
Sent: Thursday, October 07, 1999 3:40 PM
Subject: Re: The DOM is not a model, it is a library!


> I suspect you're right.  This then begs the question:  if people
> don't think DOM can become a lightweight API, how then will such
> APIs get created?  It doesn't need to be based on DOM, or done by
> the DOM WG (or even W3C).

> That being so, I'm more than comfortable saying that DOM should
> recognize its limits (and strengths!) and let other groups define
> truly light weight APIs, with no browser or legacy constraints.

I don't have an answer here.  I know there is a certain amount of interest
in the DOM WG in someday defining a "minimal DOM" for light clients,
high-throughput servers, etc.  I believe this did not make the initial cut
for Level 3, so anyone who thinks the W3C should do this in the year 2000
timeframe needs to lobby for it NOW.

As a representative of a company with a server-side focus and a strong
interest in ultra-light clients, I *do* want to see a standard here.  I
don't care ... and I doubt if my employer cares ... whether the W3C defines
the standard or something like the ad-hoc group that defined SAX does it, as
long as the "standard" is meaningful and widely adopted.  My intent in this
thread has been to suggest that DOM Level 2 put in some mechanism -- perhaps
but not necessarily the hasFeature() method -- by which an ad-hoc group
could define the subset, define (and perhaps register) the name that
identifies it, and still be "DOM compliant".  Then again, maybe your
suggestion that the DOM WG simply come to terms with its own limitations and
decide NOT to do this makes more sense.  In any event, I'll be glad to
participate in some other group if the W3C doesn't want to take on this job.

Mike Champion

Received on Thursday, 7 October 1999 16:10:52 UTC