W3C home > Mailing lists > Public > public-rif-wg@w3.org > October 2008

Re: [PRD] PICK specification --> comment

From: Christian de Sainte Marie <csma@ilog.fr>
Date: Tue, 28 Oct 2008 11:14:04 -0400
Message-ID: <49072C3C.4060800@ilog.fr>
To: Paul Vincent <pvincent@tibco.com>
CC: Gary Hallmark <gary.hallmark@oracle.com>, RIF WG <public-rif-wg@w3.org>

Paul Vincent wrote:
> 
> My thoughts are:
> 
> A - use cases like MISMO are not inferencing, and indeed ruleflow-based
> BREs use of inferencing can be rare [*1], implying a CR strategy is
> irrelevant [*2] in most non-academic use cases for RIF. In other words,
> it could reasonably be ignored in PRD1.0 / deferred til such time as
> inter-BRE inference rule interchange became a proven PRD limitation [*3]

Hearing the discussions, it seems to me that ignoring the subject in PRD 1.0 might be a dangerous choice. But everybody seems to agree on having it simple.

> B - a classification scheme of CR strategies sounds dangerously like
> "R&D" to me, which should be a huge warning bell for W3C on this topic.

Hmm... Maybe :-)

> On the other hand, it is "useful" information for rule interchange. On
> the other hand again, I would expect CR schemes to be arbitrarily
> complex and/or proprietary.

We seem to be converging towards something like:
- PRD 1.0 will provide a simple way to indicate a CR strategy, using e.g key words at the ruleset (and maybe rule) level;
- if there is a limited number of basic schemes that everybody knows how to translate into their own language and that is close enough to (one of) their own basic schemes, PRD 1.0 will include standard syntax (e.g. keywords) corresponding to these basic CR;
- anyway, non-standard syntax sould be allowed (ee.g. non standard keywords), where making sure that the other side of the interchange understand what they mean is the responsibility of the producer.

> C - the proposed solution (aka/as far as I can tell: "lets have a simple
> classifier that provides some CR information to RIF rulesets") seems
> like a reasonable compromise. 

Indeed, this is where we seem to be converging to. Constructs to indicate more complex CR would definitely be left to future extensions.

> Comments:
> [1] Anecdotal only - I don't think anyone has done research on this and
> I doubt the BRE vendors measure this / want to make such data public

Maybe not the details of their internal working. But we are talking a couple (one or two) basic startefies that everybody has in common.

What I hear is that we might be ending up with one "forward chaining" standard strategy, which would be something like "refraction+priority+recency", if there is enough overlap in our various interpretations of refraction/repeatability; and maybe a "sequential" strategy, if there is enough overlap in the ways the different engines handle it.

> [2] If there is no inferencing, then a rule cycle will only complete
> once for any ruleset in scope. Ruleflow-defined rule systems only have 1
> ruleset in scope at a time. Therefore, CR is not required. Further,
> usually such rulesets will be exclusive (have only a single rule whose
> condition matches the current state of WM).

Interesting. Do you have evidence of that?

Cheers,

Christian                                   
Received on Tuesday, 28 October 2008 15:16:03 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 2 June 2009 18:33:57 GMT