- From: Ian Hickson <ian@hixie.ch>
- Date: Thu, 24 May 2007 19:37:24 +0000 (UTC)
On Fri, 30 Sep 2005, Anne van Kesteren wrote: > > I tested attribute parsing, createAttribute and setAttribute in Opera, Firefox > and Internet Explorer. > <http://annevankesteren.nl/test/dom/html/attr/001> > <http://annevankesteren.nl/test/dom/html/attr/002> > <http://annevankesteren.nl/test/dom/html/attr/003> > > It seems that Firefox lowercases all unknown and known attributes. It > seems that Opera and Internet Explorer lowercase all known attributes > unless they are created using createAttribute. Opera and Internet > Explorer do not touch unknown attributes. What Opera does is more complex than that actually. For example it turns this [1]: <!DOCTYPE html> <p tesT=a TEST=a TeSt=a test=a> ...into this: <p tesT="a" tesT="a" tesT="a" tesT="a"> ...which makes no sense. IE also returns four attributes for this, which would require the tokeniser to have element-specific knowledge, which I really don't think is a good idea. In the DOM, though, we want XML and HTML to be as close to each other as possible. I think the current rules in the spec are as good as we can get really. (Given the lack of interoperability here we have some flexibility.) [1] http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C%21DOCTYPE%20html%3E%3Cp%20tesT%3Da%20TEST%3Da%20TeSt%3Da%20test%3Da%3E Cheers, -- Ian Hickson U+1047E )\._.,--....,'``. fL http://ln.hixie.ch/ U+263A /, _.. \ _\ ;`._ ,. Things that are impossible just take longer. `._.-(,_..'--(,_..'`-.;.'
Received on Thursday, 24 May 2007 12:37:24 UTC