Re: QASG last call comments: Normative redundancy

Le 03 févr. 2005, à 22:02, Gary Feldman a écrit :
> I disagree.  The formal language is far more useful for the humans who 
> have to create the tests or the code.  It's concise, precise, and 
> unambiguous.  It's often easier to do a reliable implementation from 
> the formal spec - recursive decent parsers (or any hand-written lexer 
> or parser), for example.
> I've always thought that the prose should focus on motivation and 
> decision processes, while avoiding anything that might resemble a 
> normative description unless there is really no choice.

Just as a side note. It's impossible to express every requirements with 
a formal language like XML for example.
Björn could give you plenty of examples about it.

The QAWG will address your question.

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

Received on Friday, 4 February 2005 12:35:27 UTC