On Mon, 11 Feb 2002, Joseph Kesselman/CAM/Lotus wrote:

> 
> >but DOM3 *is* the standard interface. It's the w3c rec and it's what
> >xerces-j 2.0 uses.
> 
> DOM3 is _NOT_ a standard yet. It's a Working Draft, not a RECcomendation.
> That means it's still subject to change.
> 
> However, there's nothing wrong with doing a prototype that follows the
> outlines of that draft, especially if you're willing to put in the work to
> track the draft as it continues to evolve.

I agree, this is what we have done with the XPath module. This way we can 
have feedback on the usefulness prior to REC.

> 
> So my recommendation would be (a) to try to wrap something similar to the
> DOM's API around your existing algorithm -- improving the algorithm, if
> necessary, to better follow the DOM3 sketch -- (b) to then treat that
> "prototype" as a nonstandard extension, and (c) to keep an eye on the DOM
> as new working drafts come out and update it appropriately.
> 
> If we're lucky and the DOM's behavior doesn't change too much, this would
> put us in a good position to go forward when DOM3 actually does become a
> standard. Meanwhile, it'd give us some implementation experience -- and the
> chance to tell the DOM folks which portions of their draft are unclear,
> and/or which of the proposed behaviors might want to be rethought.
> 
> Downside: If you're going to track the DOM, you may not be a highly stable
> platform for other folks to use. But that's the nature of experimental
> extensions anyway...
> 

Agreed, the XPath stuff changed at the end of last week so the module is 
now out of date. I'm hoping to assign some resources next week to get it 
back with the WD.


> 
> 
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [EMAIL PROTECTED]
> For additional commands, e-mail: [EMAIL PROTECTED]
> 
> 

Gareth


-- 
Gareth Reakes, Head of Product Development  
DecisionSoft Ltd.            http://www.decisionsoft.com
Office: +44 (0) 1865 203192



---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to