Re: Answer to Ian Hickson: Formal vs prose language normativity

Hello Ian,

Le lundi 16 mai 2005 à 17:27 +0000, Ian Hickson a écrit :
> Whatever solution the QAWG is advocating for contradictions between two 
> different parts of the prose would IMHO be acceptable as a solution for a 
> contradiction between the prose and formal language.

I don't know if the previous messages in this thread made it clear what
we intended to change in SpecGL to accomodate your request:
* instead of saying "define which from formal language vs prose has
priority", we would say "if the WG has a position on which takes
precedence, make it clear which it is"
* and adding that taking such a position doesn't relieve the WG from
dealing with any discrepancies as errata as defined in the process
document.

> My point is mainly that I think it makes no sense to establish priority 
> rules for one class of contradiction arbitrarily. What is special about 
> formal/prose contradictions as opposed to prose/prose (or formal/formal) 
> contradictions?

Contradictions are more frequent in formal vs prose than prose vs prose;
this good practice was added as a result of ambiguities in various
specifications, e.g. produced by the HTML WG, where they had indeed a
position with regard to which was "more normative", but didn't put it in
the specification.

Does clarifying the fact that errata are still the normal road
alleviates your objection? Is there any middle ground acceptable short
to remove the said sentence?

Thanks,

Dom
-- 
Dominique Hazaël-Massieux - http://www.w3.org/People/Dom/
W3C/ERCIM
mailto:dom@w3.org