W3C home > Mailing lists > Public > www-dom@w3.org > April to June 1998

Re: iterators, etc. + Attribute question

From: Claude Zervas <claude@utlco.com>
Date: Mon, 01 Jun 1998 22:20:48 -0700
Message-Id: <>
To: www-dom@w3.org
At 09:34 AM 6/1/98 -0400, Mike Champion wrote:
>At 03:31 AM 6/1/98 -0400, Claude Zervas wrote:
>>Would there be any chance that the node tree traversal
>>methods in the Node interface be deprecated and sent
>>to some sort of work camp for bunk code patterns?
>This "bunk code pattern" is well-beloved by untold generations of SGML
>weenies ;~) who would prefer to deprecate the idea of used indexed arrays
>to represent document structures.   
>It's a dilemma for us -- there's no single acceptable paradigm for document
>structuring and navagation that we've been able to find, and we don't want
>to bloat the spec by supporting everything, and (while I happen to agree
>with you about the iterator design pattern) iterators are not supported in
>any of the legacy APIs from which we're trying to build.

Ok. I'm not really familiar with the legacy APIs and how much current
software depends on them so I was a little hasty in condemning them.
Thanks for the clarification.

But I would still like to know if the Attribute interface should
inherit from Node or not because Element.getAttributes() is very
handy but since it returns a NodeIterator (in the current spec)
I'm kind of confused. I'm assuming there is a typo either in the
declaration of the Attribute interface or in the declaration
of Element.getAttributes() but I can't know which. Kent Tamura's (IBM)
implementation of Attribute inherits from Node but this may just
be to facilitate the implementation. Am I missing something here?
The prose is pretty unhelpful on this, it mentions "attrubute node"
but does not explicitly state or show inheritance of Node.

Anyway, I'd like to thank all the folks at W3C for putting in
the work to get the DOM out!

Claude Zervas, UTL Co., Seattle
Received on Tuesday, 2 June 1998 01:13:52 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 20 October 2015 10:46:04 UTC