W3C home > Mailing lists > Public > w3c-wai-gl@w3.org > October to December 2001

Re: [SCRIPT-TECHS] DOM-2 request for links or content

From: Jim Ley <jim@jibbering.com>
Date: Sat, 17 Nov 2001 14:17:28 -0000
Message-ID: <022901c16f84$b80228c0$963c70c2@7020CT>
To: <w3c-wai-gl@w3.org>
"jonathan chetwynd":
> I expect DOM-2 needs a whole large section.

The techniques involved in DOM2 are no different to the techniques
involved in any other DOM, there's no way of identifying if a browser does
support DOM 2 (in theory there is - there should be the implementation
object hasFeature method, but as the only implementations of this give
incorrect results you still need to do all the same object detection to
overcome the areas where it's not been implemented, despite what
hasFeature says.)

There's a case for DOM 2 being encouraged over proprietary DOM's, but then
simple "maximum exposure" of your script already does that, the only
multibrowser multiplatform approach to most scripting is using DOM, either
0,1 or 2, but DOM 2 alone isn't enough, for example, how you get to the
Document object (for example.) is not standardised anywhere other than the
browser vendors mutual agreement.  I see no differing techniques involved
in authoring accessible scripts by using DOM2 over proprietary ones, all
the issues are the same, we just need to encourage use of DOM 2 which we
already do via the "author to appropriate standards" checkpoint.

> Please could members post me their suggestions?
> I've started the ball rolling with a 'whitespace' issue:
> http://www.learningdifficulty.org/develop/w3c-scripts.html#dom-2

The whitespace is a non issue, unless you're iterating over all the
elements, and this isn't something you often need to do - I certainly
wouldn't've thought those needing to read a techniques document would need
to. It is most inefficient and poorly supported by browsers, using
getElementById, or getElementsByTagName is a lot more efficient and makes
a lot more sense, especially as it avoids the problems you're addressing
there - Rather than  go for complicated workarounds (which result in parse
errors in older standards compliant browsers.) you just avoid having to
iterate over objects, unless you really have to. (The nodomws.js script
incorrectly defines whitespace and for no reason other than making the
Regular expressions more complicated that I can see.)

Received on Saturday, 17 November 2001 11:26:35 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 16 January 2018 15:33:39 UTC