Re: QASG last call comments: Avoiding device-dependent profiles

Ian,

Le 20 janv. 2005, à 08:06, Ian Hickson a écrit :
>> Do you mean the Good Practice A in section 4.1 suggests profiling
>> specifications [...]?
>
> Yes.
>
>
>> Would the removal of such suggestions satisfy your comment?
>
> Removing all suggestions that specifications be profiled would be fine 
> by
> me too, yes.

* Trying to progress on this issue. I had an action from the QA WG to 
clarify with you what did you mean.

It seems you would like to remove any kind of references to profiles in 
SpecGL. As a personal comment, not the QA WG, I would like to stress 
the reality check about Specifications produced at W3C. Many of them 
are profiled and it's unlikely it will change given the nature of the 
Web.

I understand completely your desire to discourage the profiles, but I'm 
not sure it will happen. (Trying to not being like the 3 monkeys: not 
heard, not seen, not said)
And I don't think SpecGL encourages profiles, The whole document 
Variability in Specification is also about all problems regarding these 
issues.

So given that fact, how do people should deal with profiles, when they 
have some in their specifications. I think also we make a fair 
presentation on the benefits and drawbacks of variability in the 
section 4.
In the GP (Non normative) 4.1, we don't mention device-dependent 
profiles as well, except the example of CSS.

Should we:
	* modify the Good Practice?
		What does that mean.
	* modify the techniques?
	* stress a bit more in the table with benefits and drawbacks.

Which prose or at least topic you would like to be written. (Personally 
I would prefer the table).

-- 
Karl Dubost - http://www.w3.org/People/karl/
W3C Conformance Manager
*** Be Strict To Be Cool ***