W3C home > Mailing lists > Public > public-rif-wg@w3.org > May 2010

Re: Urgent: Issue with RIF-Core EBNF Grammar?

From: Dave Reynolds <dave.e.reynolds@googlemail.com>
Date: Wed, 12 May 2010 11:36:04 +0100
Message-ID: <4BEA8494.1050908@gmail.com>
To: Jos de Bruijn <jos.debruijn@gmail.com>
CC: Axel Polleres <axel.polleres@deri.org>, Christian De Sainte Marie <csma@fr.ibm.com>, RIF <public-rif-wg@w3.org>
On 12/05/2010 11:19, Jos de Bruijn wrote:
>
>
> On 05/12/2010 12:09 PM, Dave Reynolds wrote:
>> On 12/05/2010 09:36, Jos de Bruijn wrote:
>>> Axel, Christian, all,
>>>
>>> I raised my concerns about the RIF-Core spec in a separate email.
>>>
>>> Concerning facts about class membership: they are both in BLD and PRD
>>> (see [1]).
>>> Concerning class membership atoms in rule conclusions: I do remember
>>> that we explicitly forbade them in Core.
>>
>> That's my recollection too.
>>
>> Our official record of the decision [1] was to allow membership "in Core
>> facts and conditions".
>>
>> We did at one point have an EBNF that reflected that resolution.
>>
>> My memory [2] was that Gary on behalf of the PRD group later pointed
>> that asserting membership facts was just as problematic as concluding
>> them via non ground rules. The problem being that in object-based PR
>> implementations membership is hardwired in the external data model.

Let me re-emphasise that this is just memory (and my memory is fallible 
even for things that I care about - #/## are not in that category :) ). 
So if any one can turn up the evidence either way that would be great.

> Then I find it strange that PRD allows asserting class membership facts.

Quite. This came in later and it might be that Core should have been 
revisited in the light of the evolution PRD.

However, my understanding is that PRD has a notion of creating new 
individuals, assertion of membership facts is only allowed at creation 
time. [That's no doubt the wrong terminology.]  It's not intuitively 
obvious how to reflect a procedural notion such as "creation" accurately 
into Core but perhaps it would have been possible.

Dave
Received on Wednesday, 12 May 2010 10:36:41 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 12 May 2010 10:36:42 GMT