Re: [SpecGL Draft] E. Good Practice: Write sample code or tests.

Hi Lynne,

Thank you very much for your comments. Much appreciated.

Le 14 juil. 2004, à 14:37, Lynne Rosenthal a écrit :
> At 02:28 PM 7/13/2004, you wrote:
>> Why should I care?
>>         When a technology is created, it is very tempting to add many 
>> features, because they really seem
>> appealing or useful. Though defining a new feature doesn't mean that 
>> the model is easily implementable or implementable at all.
>
> Don't understand this.

mmm My bad. I wonder what I was thinking when I have written this.

My point:
	When a WG is working group is developing a technology, all members of 
the WG are very tempted to add as many features as possible (because it 
will be cool, because it's needed here or there, because it's just an 
idea which popped up).
	But adding these features doesn't mean:
		- that it would be easy to implement them: Benefit/Cost for each 
individual features.
		- that it would be implementable: When you come to the implementation 
phase, you see that in fact, it doesn't make sense at all or it's 
impossible to implement.
		- that it doesn't fit in the bigger scheme of things: For example, 
two features in the spec that would have opposite or contradictory 
behaviour without mutual exclusion mechanism (The two features must not 
be present at the same time).

> Not sure how to rewrite this, but it does need improvement. Here is a 
> try:
>

You rock!!! All the rewriting you have suggested are a lot clearer. 
Thanks.

>> Techniques:
>>         1. Try to associate developers to the Working Group progress 
>> (Open source or commercial)
> Awkward.  I don't know how to rewrite this.

	1. Encourage developers (Open Source or Commercial) to work with the 
Working Group to get pre-implementation at the same time the technology 
is developed.


Lynne++ Lynne++ Lynne++


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