W3C home > Mailing lists > Public > public-rif-wg@w3.org > August 2006

FW: [RIF] Draft Minutes for July 25th RIF WG Telecon

From: Donald Chapin <Donald.Chapin@btinternet.com>
Date: Tue, 1 Aug 2006 16:08:41 +0100
To: <public-rif-wg@w3.org>
Message-ID: <001401c6b57c$602530a0$0200a8c0@DonaldChapin>
To correct mailing list


From: Donald Chapin [mailto:Donald.Chapin@btinternet.com] 
Sent: 29 July 2006 13:02
To: 'public-rif-wg-request@w3.org'
Subject: [RIF] Draft Minutes for July 25th RIF WG Telecon


 <http://www.w3.org/> W3C


RIF Working Group Telecon

25 Jul 2006

See also: IRC log <http://www.w3.org/2006/07/25-rif-irc> 



Hassan_Ait-Kaci, csma, MarkusK, ChrisWelty, +1.650.857.aaaa, FrankMcCabe,
Mike_Dean, David_Hirtle, Philippe_Bonnard, pfps, Sandro, Dave_Reynolds,
Leora_Morgenstern, Donald_Chapin, Axel_Polleres, StellaMitchell, Jos_De_Roo,
+43.512.507.9aabb, PaulaP, +1.441.224.aacc, jeffp, josb, johnhall, igor,
Harold, GiorgosStoilos?, Gary_Hallmark, MichaelKifer 


AllenGinsberg, MohamedZergaou,i FrançoisBry (apologies for the late notice
due to sickness), MichaelSintek


Christian de Sainte Marie 


Donald Chapin 


*	Topics <http://www.w3.org/2006/07/25-rif-minutes.html#agenda#agenda>

*	Summary
of Action Items 


 TOPIC: Admin


 Who will attend the August 1st telecon?

<Hassan> +1

<FrankMcCabe> i should be there 

<donald_chapin> +1

<pfps> +1

<AxelPolleres> -1

<MarkusK> -1

<DaveReynolds> -1

<PhilippeB> +1

<LeoraMorgenstern> +1

<mdean> +1

<DavidHirtle> +1

<trackbot> Sorry... I don't know anything about this channel

<PaulaP> +1

<Darko> +Darko


<ChrisWelty> RESOLVED: Accept minutes of July 11 telecon



    TOPIC: Liaison


<PaulaP> http://www.w3.org/2001/sw/grddl-wg/

<PaulaP> it is a new WG

Everyone should look at the charter of GRDDLL-WG to see if they have an
interest in it.



TOPIC: Use Cases & Requirements

Close Action 39

Action 59 closed, but discussion is going on about standardizing the
Semantic Web Layer Cake diagram - no decision



Action 61 done

Action 73 done


TOPIC: Discussion on Questionnaire


Axel - Can Francois reformulate the questions 3.1 & 3.2 about decidability -
agree that there is a problem

<csma> http://lists.w3.org/Archives/Public/public-rif-wg/2006Jul/0040.html

<PaulaP> +1 for the third suggestion of Francois

Chris w - 3.1 not confusing, 3.2 doesn't see problem, the new question is a
good one

Christian - propose add a discriminator on which kind of data format that
the language is able to consume

<PaulaP> no, it is not

<ChrisWelty> DaveR's message:

<ChrisWelty> RIFRAF questionaire:

Questionaire is strictly incremental. If you need a new discriminator,
simply ask for it to be added. Once it is there new languages can use it.

<MarkusK> +1 for Francois' remark on 3.1. Decidability of a rule language in
the sense of Computer Science really is not what we mean here. Decidability
should refer to some inference problem or similar reasoning/computation

<AxelPolleres> ok

Data format consumed can be multiple

<AxelPolleres> ok!

<AxelPolleres> ACTION: Axel to add to questionnaire question whether the
language has means to [recorded in

<AxelPolleres> access data in Web formats such as HTML, XML, RDF, OWL data.

ACTION Alex Add the data format discriminator

<JosDeRoo> http://www.w3.org/2002/09/wbs/38457/RAFQuestionnaire/

<Harold> Initially, we referred to Decidability of the QUERY problem of a
rule language.

<AxelPolleres> it should be about decidability of "entailment" of facts,

<Harold> E.g., Datalog has a decidable QUERY problem. Hornlog doesn't.

<MarkusK> 3.1 just is a formal issue

<MarkusK> that should be easy to fix

<MarkusK> just say what "decidability" refers to

<MarkusK> "decidability of a language" means deciding whether something
belongs to the language.

<AxelPolleres> I suggest that francois and markus both propose
reformulations for 3.1 and 3.2

<MarkusK> I will send an email regarding 3.1

<MarkusK> I have nothing to say about 3.2 so far ...

<JeffP> ACTION: MarkusK Write email with solution to 3.1 & 3.2 on the
questionaire [recorded in

<AxelPolleres> Can you formulate a question plus options for answers, Frank?

FrankeMcCable Need to add the kinds of 'types' supported as a discriminator

<AxelPolleres> Let's post all these suggestions on the mailing list under
the [RIFRAF] header.

We won't add Dave Reynolds email second point as a discriminator. If some
language needs it they will add it

Gary – Not all questions relate to every type of rules – not always possible
to know whether they relate or not – need reorganization or better guidance
on how questions relate to types of rules

<AxelPolleres> +1 to Gary, shall we split the questionnaire then?

<PaulaP> +1 to Gary's comment

Christian - Need to say not applicable for any question or a comment space
for any lack of clarity

Gary - Need a mapping of 'Prolog-like' languages to production rules

See the ILOG answer for questions on the mapping to Production Rules

Axel - Comments are switched on and can be used to say 'not applicable'

ChrisW - Can't uncheck / unanswer a question

Alex - Could use checkboxes throughout as they can be 'unchecked'

<Darko> it is possible not to answer a question and then you get a
notification: (1 response didn't contain an answer to this question)

<AxelPolleres> what about a checkbox: "not applicable (please specify)"



TOPIC – Questionaire is Not Ready for People Outside the Working Group to
Fill In


Dave Reynolds - Questionaries would need some work before it is used outside
the Working Group

<Harold> In
http://lists.w3.org/Archives/Public/public-rif-wg/2006May/0235.html I
responded to Axel's 'Syn' comment about types -- I agree, could be spliced
in thus: Typed vs. Untyped Variables (Types can reuse class definitions in
RDFS and OWL).

Chris W - The RIFRAF should be the place where the explanation of
discriminator belongs. People outside the RIF WG should be able to
understand the RIFPAF Wiki page.



TOPIC: - Visibility and Update Access to RIFRAF Questionnaire Wiki Page


<AxelPolleres> I set the questionnaire currently only visible to RIF WG

<AxelPolleres> ... at the moment.

Who should have access to the RIFRAF Wiki page?

Only RIF WG members can fill out the questionaire

The RIFRAF Wiki page questionnaire should be public because of our charter

Christian - Add a box at the end where people can suggest additional
discriminators together with their values

<DaveReynolds> +1

<PaulaP> +1

<AxelPolleres> ... and drop 1.1.3

<DaveReynolds> Depends how you are going to do the analysis of the form






TOPIC Items Moved to RIFRAF from Requirements


Chris W - The 'owners' of the items moved from Requirements to RIFRAF need
to take responsibility for creating any necessary discriminator for RIFRAF

Christian - move this topic to next week

<ChrisWelty> paula's message from this morning:

ACTION: 'Owners' propose new discriminators on email

Christian - Comment on email if your disagree with additions



TOPIC: Technical Design -- Semantics for Proposed RIF Condition Language


Harold - Email today with draft of these semantics which he describes.
Michael is working on the Wiki to make this really formal

Christian - The semantics page doesn't really state explicitly what the
semantics is


Harold - Current wiki semantics is still very general. Now this is being
more concrete on the Wiki - focusing on conjunctions only

Michael - Defining semantics by stating what the models are and what is the

<Harold> Positive condition semantics is at

<Harold> ... Given a condition formula phi(X1,...,Xn) with free variables
X1, ..., Xn and an interpretation M, define M(phi(X1,...,Xn)) as the set of
all bindings (a1/X1,...,an/Xn) such that M |= phi(a1,...,an), where
a1,...,an are elements in the domain of M. ...


TOPIC: Technical Design – How Many Semantics Should the Condition Language


Christian - Current semantics does not propose one semantic.

Michael - Already agreed that there would be different dialects

<pfps> As a point of interest, I would like to have a pointer to the
decision that the RIF will have different dialects.

<FrankMcCabe> there is no one semantics!

Christian - Is it not possible to have one semantics for the conditions and
that shared by all the dialects?

<sandro> sadly, pfps, I don't think the minutes of the most recent F2F ever
got done and approved, although they got very close. :-(

<pfps> Hmm. That is not a very happy situation. Were they not supposed to be
done quite some time ago.

Michael - Different semantics simply exist for the different dialects, so a
single semantic for the condition language is not possible

<pfps> Given that there does not appear to be a recorded decision that there
are going to be different RIF dialects, then it would probably be better not
to treat it has having being decided.

FrankMcCable - Thinks it is unrealistic that, except for possibly a very
small core, to have a single semantics. It could even be counterproductive
to do so.

<ChrisWelty> The decision was that there would not be "too many" dialects

<ChrisWelty> that does not preclude that there will be one

<Harold> In the Roadmap discussion, we had basically three dialects: FOL, LP
(Naf), and Production Rules.


TOPIC Technical Design – Para-Consistent Semantics


<AxelPolleres> para-consistent logics are nothing new... can you send a
pointer on this?

<MarkusK> +1 to send a pointer; there are so many approaches towards
para-consistency ...

<AxelPolleres> there are many different approaches to the issue of
para-consistency, right?

<FrankMcCabe> Carl hewitt's paper:

<Harold> Paraconsistency is a kind of scope-localized para-consistency, so
nicely fits to our scopes

<FrankMcCabe> I am not completely sure that this is the correct pointer

<Harold> I meant above: Paraconsistency is a kind of scope-localized
consistency, so nicely fits to our scopes.

<pfps> paraconsistency does not necessarily have anything to do with

<FrankMcCabe> right. Its my understanding that it refers to the shortest
proof of inconsistency. Any shorter proofs are OK


ChrisW - Everyone should record regrets on Wiki if they are not attending
Aug 15th meting

Summary of Action Items

[NEW] ACTION: Axel to add to questionnaire question whether the language has
means to [recorded in
[NEW] ACTION: John will add regrets to July 18 minutes and resubmit
[recorded in http://www.w3.org/2006/07/25-rif-minutes.html#action02]
[NEW] ACTION: MarkusK Write email with solution to 3.1 & 3.2 on the
questionaire [recorded in
ACTION: Will have a meeting August 1st and Sandro will chair [recorded in

ACTION Peter - Proposed a single dialect semantics - a kind of first order
logic semantics by Aug. 29 - Condition Language only

ACTION Harold make explicit the assumptions behind the semantics he and
Michael are documenting

ACTION - Christian will propose a single semantics for the Condition

ACTION Axel add 'no answer' to each question 

ACTION Axel to add box to suggest more discriminators

ACTION: Axel (and Sandro) - Fix questionarie so that it can be duplicated
for different languages

ACTION FrankMcCable will propose a list of type capabilities’ as a

ACTION: 'Owners' propose new discriminators on email

ACTION Alex Add the data format discriminator


 [End of minutes]


Minutes formatted by David Booth's scribe.perl
<http://dev.w3.org/cvsweb/~checkout~/2002/scribe/scribedoc.htm>  version
1.127 (CVS log <http://dev.w3.org/cvsweb/2002/scribe/> )
$Date: 2006/07/25 16:31:31 $ 

(image/gif attachment: image001.gif)

Received on Tuesday, 1 August 2006 15:09:16 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:47:40 UTC