Re: towards a resolution on RIF BLD classification

Michael Kifer wrote:
> Dave Reynolds wrote:
>> Chris Welty wrote:
>>>
>>> RIF WG,
>>>
>>> We have been discussing this issue for a few weeks now, and while it is 
>>> a small minority who prefers not having rif:subclass in the language, 
>>> some probing revealed that some people who have stated their support are 
>>> "going along with it" out of indifference or without fully understanding 
>>> the issue.  I think we need some more community feedback.
>>>
>>> Also, discussion of this issue is delaying the production of a next 
>>> public WD, which I would like to release ASAP.
>>>
>>> My proposal is that we publish a WD with the current classification 
>>> scheme as it is in the editors draft on the wiki, suitably labeled as 
>>> "under discussion" with a specific call for feedback on that proposed 
>>> feature of RIF BLD, and a better design rationale description for the 
>>> feature, including the pros and cons.
>>>
>>> I would like to call this resolution at tomorrows telecon, if possible, 
>>> with actions on perhaps Michael and Dave to draft the pro and con sections.
>>>
>> I think we should resolve the question of whether and why we want to 
>> carry a data model at all first. Asking for feedback on the specific 
>> case of classification without presenting the rationale for this overall 
>> capability or the whole picture of the other things that would come 
>> along with it (domain/range stuff) seems unhelpful to me.
> 
> 
> Mixing the data model in rules is quite common in all F-logic based
> systems, because it is so convenient to define and query the meta model in
> such systems.

Sure, and in a f-logic RIF dialect I would expect such constructs.

BLD is not that dialect.

Dave
-- 
Hewlett-Packard Limited
Registered Office: Cain Road, Bracknell, Berks RG12 1HN
Registered No: 690597 England

Received on Tuesday, 28 August 2007 13:30:09 UTC