- From: <keshlam@us.ibm.com>
- Date: Thu, 1 Oct 1998 13:28:25 -0400
- To: www-dom@w3.org
>> * added text about HIERARCHY_REQUEST_ERR being raised >> by Node.insertBefore/replaceChild/appendChild when the >> specified node is one of this node's ancestors. >Oh, thank-you *very* much (not) ... bang goes O(1) insert. Well, it's "only" a performance hit and a relatively simple and not-horribly-expensive walk up the parents. As opposed to what's required to make getElementsByName work. On the other hand... One could argue that this should really be a quality-of-implementation issue, since it's a lot of work to guard against something that no properly-written application is ever going to do. I'd have argued to "fix it in pubs". Too late now, I presume...? BTW, I'm still getting pushback from users about the inconsistancy between attributes having no parent, and attributes not being sharable; they could understand either choice but the combination confuses them. And me, to be honest. ______________________________________ Joe Kesselman / IBM Research Unless stated otherwise, all opinions are solely those of the author.
Received on Thursday, 1 October 1998 13:28:43 UTC