- From: Boris Zbarsky <bzbarsky@MIT.EDU>
- Date: Tue, 06 Mar 2007 18:00:33 -0600
- To: Daniel Beardsmore <public@telcontar.net>
- CC: www-style@w3.org
Daniel Beardsmore wrote:
> I didn't say that it acts similarly, Boris did. But it would seem
> better, if such a feature is defined in CSS (which it is), to at least
> try to tie in one's implementation to real CSS
Only if your implementation actually follows the spec. If your implementation
is known to be seriously buggy, as far as the spec is concerned, giving it the
spec name is a bad idea. It means people can't use the spec version in their
pages (for UAs that support it) because it'll break the page in your buggy UA.
It's standard procedure for the browser vendors who care about the health of the
web to have vendor prefixes on the parts of their CSS support that's not really
per spec yet, as well as on their extensions to the CSS standards ("-o" for
opera, "-khtml" for KHTML, "-moz" for Gecko, etc).
In any case, in this case the value wasn't there in CSS2 and the implementation
in Gecko predates its addition to CSS2.1. So it was added with a vendor prefix,
being an extension to the then-existing CSS standard. Since CSS2.1 came on the
scene, the implementation hasn't really been revisited to check whether it
complies with what CSS2.1 says, so to be safe it's staying as a -moz- property.
> This is, though, how most browser development goes anyway :)
I should hope that "most browser development" doesn't mean that you take your
CSS extensions that happen to do something similar to a standard feature and
just give them the same name!
See also https://bugzilla.mozilla.org/show_bug.cgi?id=261081 for what it's worth.
-Boris
Received on Wednesday, 7 March 2007 00:00:49 UTC