RE: Proposal for java-language-binding: use of collections framew ork

> -----Original Message-----
> From: Joseph Kesselman [mailto:keshlam@us.ibm.com]
> Sent: Monday, February 25, 2002 1:00 PM
> To: Stefan Wachter
> Cc: www-dom@w3.org
> Subject: Re: Proposal for java-language-binding: use of collections
> framework
> 
> 
> 
> Won't happen. Part of the point of the DOM is portability; 
> changing basic APIs at this late date would break that.

Uhh, it *has* happened.  Portability is "broken" in the Java/DOM world
because JDOM and dom4j have to be taken into account already. And the php
binding on top of the libxml DOM only vaguely resembles the DOM, and there
are various Python DOMs (some of which are quite strict, others quite
loose), and I'm sure I've only scratched the surface of DOM-like
implementations that more or less ignore the W3C when the native language
offers a more convenient way to do something.  

The question in my mind is "would an official W3C Java binding onto the Java
collections and iterators be just one more point of fragmentation, or could
it help unify things once again?"  It sure would be nice to say that there
is only one "DOM", but each language binding is a suitable place for
language-specific convenience methods, etc. that make it easy to work with
the native idioms.  In some ideal world, everyone would simply implement the
DOM as specified and get interoperability by strict conformance.  But that
hasn't happened, and the world looks like it is moving away from that even
as an ideal (e.g., JDOM is an official JCP activity now).

This is probably not a great job for the W3C, because a lot of this stuff
takes place outside the membership.  Nevertheless, there is something to be
said for initiating a public forum of some sort to encourage
dialogue/experimentation to see if some commonality of language-specific
convenience interfaces and methods at the *binding* level could help make
the DOM world less confusing.   I honestly don't know if this has any chance
of being useful, but it might be worth exploring the idea a bit.

Received on Monday, 25 February 2002 13:38:33 UTC