Re: Progress event spec

On Mon, 5 Feb 2007, Jean-Claude Dufourd wrote:
> 
> Such an extreme message does not help the discussion. Besides, in the 
> paragraph 2 of the section you quote, I read: "All chapters are 
> normative except for specific sections marked as being informative. " 
> The sentence I quoted was in an unmarked, hence normative paragraph. At 
> best, there is a contradiction in the SVG spec. But it is *not* as clear 
> as you both pretend.

Oh, well, if we assume that the SVG specification might have 
contradictions, then we would just be debating working group intent, which 
we could do til the cows come home. Fortunately, if this is indeed an 
error, then I imagine the working group will fix it. Until it is fixed, 
however, it continues to say what it says, which is to say, in this 
instance, nothing of substance (since there are no normative testable 
conformance criteria in the cited sentence).

One has to work on the assumption that a specification in CR is correct, 
since if we work on the assumption that it's wrong, it will be basically 
impossible to obtain two interoperable implementations (which is what we 
need for the spec to advance to REC). If implementors assume the spec is 
wrong, and just guess at what was intended, instead of following the spec 
to the letter, then we'll just end up in the same Tag Soup mess that HTML 
has ended up in because of HTML4's vagueness.

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

Received on Monday, 5 February 2007 21:25:52 UTC