Re: report on coverage issue

All,

While we were discussing about that issue (the definition of 
"coverage"), Chris offered that the real, precise, formal definition 
would actually be the technical specification itself, and that, until 
then, we could, and probably should, keep by an informal, intuitive 
definition. Sandro and myself found the argument convincing.

Sandro stressed that all we could say in UCR had to relate the notion of 
coverage to the users' requirements. Basically: we say that a rule 
language is covered by RIF if RIF enables the rule interchanges that the 
users of that language are interested in. And, thus, a rule language 
coverage by RIF is not all or nothing: a rule language may be more or 
less covered, depending on the kind of usage (use cases) you are 
considering. And all rule languages to be covered will not be equally 
covered within the same time frame, as requirements will be prioritised, 
and this is the purpose of RIFRAF (analysing rule languages features in 
order to help us prioritize coverage).

One benefit of this kind of definition is that we do not have to 
distinguish different kinds of coverage (like wrt expressivity or 
translatability, as proposed by Allen [1])

Issue 22 will be discussed at the telecon tomorrow.

Christian

[1] http://lists.w3.org/Archives/Public/public-rif-wg/2006Sep/0046.html

Received on Monday, 2 October 2006 15:54:04 UTC