Re: [SVGMobile12] more on data types

On Fri, 13 Jan 2006, Jim Ley wrote:
>
> "Ian Hickson" <ian@hixie.ch>
> > On Thu, 12 Jan 2006, Jim Ley wrote:
> > > 
> > > Certainly, however I fail to see the dis-advantage in having to have 
> > > attributeX in the new version if you want to extend the semantics
> > 
> > So you're saying you _do_ want well-defined error handling, but for 
> > only one case, namely unknown attributes? And you want that those 
> > "MUST be ignored"?
> 
> No, I want it left open to the user agent.  I don't see how the above is 
> confusing - the issue of extending the semantics in later versions is 
> orthogonal to handling invalid content.

You are asking for a spec where it is legal for the UA to do something 
with _every_ attribute. How then can you later introduce a new attribute?

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

Received on Friday, 13 January 2006 19:33:44 UTC