Re: Supporting propriety "Extensions"

On Tue, 22 Feb 2005, Andrew Fedoniouk wrote:
> 
> I understand "how" but do not understand "why"... What is the purpose of 
> introducing namespaces using such strange method and notation?

To discourage their general use. Authors should not be using 
vendor-specific extensions. The entire issue is only a problem if you 
assume that vendor extensions will be in common use. If that were to 
happen, then it would be bad.

Note that we need a prefix because when a company uses a property without 
one, it causes all kinds of problems. For example, one of the main reasons 
the entire prefix idea was introduced was Microsoft's 'filter' property. 
Because they introduced it without a prefix (it was before the idea of 
prefixes), the W3C is basically no unable to introduce 'filter' in CSS 
without causing conflicts with existing stylesheets.

Same with 'behavior' (we're having to use 'binding' instead); same with 
many of the vertical text properties (we are being constrained to using 
the same values as Microsoft in order for the standard to be compatible 
with the implementations that exist).

The only problem that -foo-x properties introduce is the fact that 
identifiers can't be prefixed with a unary minus operator. The CSS working 
group, as I mentioned earlier [1], considered this, but did not think it 
likely that CSS would be extended in that way, and considered the hyphen 
prefixes on identifiers to be fine.

Note that -foo-bar is better than \-foo-bar and _foo_bar from an aesthetic 
point of view, and that in any case it has already been interoperably 
implemented. The ship, as they say, has sailed.

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'

Received on Wednesday, 23 February 2005 13:10:09 UTC