Re: Definition of feature, interoperable and implementation

Le 15 mars 2005, à 07:13, Lynne S. Rosenthal a écrit :
> There are some words, (e.g., feature, function) that we may not want to
> define - that we purposely don't want to tie it down, since it is not
> consistently used, but everyone knows what it means within their 
> technology.
> If we do decide to define it, it will be difficult.

Lynne, yes I tend to agree in the sense that depending on the 
technology, you don't know what a feature is. So it's almost impossible 
to come with a generic definition of it.

But, from what I have read so far on the CDF mailing list, there might 
be an interest for a WG to define what they consider a feature in their 
technology, then when it's time to have a double implementation of 
every feature, there will be no (less) doubts that it has been 
realized.

So more than saying "A feature is", WG could come with a designation 
more than a definition:
	"This and that, and this one over there are features."


It's a bit like the double implementation requirement of the Process 
document. Let's say to have one implementation of "id" attribute in an 
XML Parser AND one implementation of "id" attribute in an XML schema 
doesn't mean there are two interoperable implementations.

Then we can identify that the notion of double implementations is tied 
to the Class of Products.



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

Received on Tuesday, 15 March 2005 15:10:15 UTC