RIF Telecon 23 Sep 2008 -- 23 Sep 2008

<sandro> f

yes, I can scribe

<ChrisW> Scribe: LeoraMorgenstern

Admin

<ChrisW> http://lists.w3.org/Archives/Public/public-rif-wg/2008Sep/att-0111/00-part

<ChrisW> PROPOSED: Accept last week's telecon minutes

Above, minutes from last week's telecon

<ChrisW> RESOLVED: Accept last week's telecon minutes

<StellaMitchell> double quotes in PS

Agenda amendments? Harold may have something --- possibly to be determined later in the call, since neither Harold nor chris remembers

Liason

Dave Reynolds had started to do some work looking on OWL rule profile

Dave: Translation ought to be possible, but there are two issues:
... First issues: data type mismatch
... But can restrict to datatypes that both OWL and RIF support
... so there is a way to generate safe RIF Core rule set

<DaveReynolds> http://www.w3.org/2005/rules/wiki/OWLRL#Datatypes_supported

Above link gives types that are and are not supported in OWL RL

Publicity

<sandro> (excellent work Dave, thank you.... thinking about the issues you raise, and how to progress with them.....)

Chris: Good number of public comments

Sandro: probably okay. (11 public comments)

Chris: There are a couple of new comments, for which I've set up wiki pages; I've sent 3 responses out; Also have emailed a draft of response to Peter Patel-Schneider
... responses are accessible from rif wiki home page

Action Review

Action 576 on Leora: continued until tomorrow night

<trackbot> Sorry, couldn't find user - 576

Action 573 on Sandro to ask for review from OWL wg on RDF & OWL

<trackbot> Sorry, couldn't find user - 573

Sandro's action 573 is continued

<Hassan> yes

Sandro's action 565 depends on Adrian sending ssh key: will straighten it out at f2f

actions 563, 564, and 565 are continued

F2F11

chris: pending review: some prd issues still under discussion: hopefully will be discussed at f2f

same for ??

f2f11 starts this Friday.

Core task force debrief

Harold's update on Core task force (telecon on Monday, Sept. 22)

Harold: We went through all open issues, most of which came from Dave.
... Much of it had already been resolved by David's previous emails.
... In our notes can be found current state of resolution to proposed issues.

<ChrisW> http://lists.w3.org/Archives/Public/public-rif-wg/2008Sep/att-0140/2008-09-22-rif-core-minutes.html

<DaveReynolds> Issue 76 is actually special case of Issue 71 which is also agreed within the task force

<ChrisW> PROPOSED: RIF Core will include member (#) but syntactically restricted its use in rule bodies.

<DaveReynolds> There was an extra clause in the PROPOSED member resolution.

Harold will have action to list in email proposals on which there is consensus.

<ChrisW> ACTION: harold to summarize proposed resolutions from Core task force [recorded in http://www.w3.org/2008/09/23-rif-minutes.html#action01]

<trackbot> Created ACTION-578 - Summarize proposed resolutions from Core task force [on Harold Boley - due 2008-09-30].

chris: are there any that are still contentious?

<DaveReynolds> Issue 72 is still open

harold: existing contentious issues include: haven't worked out how inclusion of or in rule bodies affects safety.

dave: issues 74, 71, 76 are prety clear.
... issue 48 --- membership issue --- still needs discussion, but within core group, there is compromise resolution. Issues 70, 72 still need discussion

<ChrisW> PROPOSED: Core will have unrestricted equality in conditions, closing ISSUE-76

dave: issue 76 was special case of issue 71, so worth rephrasing and then closing 71
... (that is, 71 was rephrased)

<DaveReynolds> PROPOSED: Core should keep unrestricted equality and external function and predicate calls in rule bodies and keep external functions calls in rule heads.

dave: above covers both 71 and 76

chris: but this is more extensive than what was in agenda today.

<DaveReynolds> ok

chris: Let's wait on this, and pass it at f2f.
... any discussion about this broader resolution, which covers external calls as well as equality?

no questions, no discussion about this.

chris: this looks pretty good. No problems; we should be able to pass it at f2f.

PRD task force debrief

<csma> http://lists.w3.org/Archives/Public/public-rif-wg/2008Sep/0116.html

<csma> http://www.w3.org/2008/09/16-rif-prd-minutes.html

csma: we have a (quasi) frozen draft, but it doesn't have everything that we want to have in it.

gary: we'd need someone to show us how FLD can do aggregation.

me/ gary, you need to know how fld can do aggregation in order to share condition language?

me/ gary, or in order to share something else?

<Harold> Aggregation in DLV, SPARQL, etc.

<Hassan> Aggregation as used in PRD is operational - see: http://lambda.uta.edu/monoid.ps.gz

mkifer: rather complicated to do this in FLD.
... complicated in logical languages because of recursion.

Hassan: Should not be that complicated because aggregates as used in P
... in PRD are operational, as posted above.

michael: probably what is being asked for will be feasible.

Gary: Need to work on conflict resolution issues on PRD

Test Cases

[Note: conflict resolution is issue 64]

<ChrisW> *PROPOSED:* Each RIF test case MUST, eventually, be provided in RIF XML, and MAY be provided in some syntax for which a translator-to-XML has been promised. The translator SHOULD be available for use by the WG to check the translation.

<csma> to clarify: I am not against normative PS, for dialect where the stakeholder want one, and provided that a PS is not normative accross dialects. But even for BLD, where people seem to want a normative PS, I think very strongly that it is not a priority for this WG

chris: this does depend on Hassan finishing up the translator.

hassan: it is somewhat more difficult than expected to work on the P
... on the P
... on the PS to xml translator, because the spec/requirements aren't clear
... I had a prototype of translator with strict PS syntax
... because tokenizing was sound: everything was put into double quotes, and you didn't have to worry about weird syntax
... but currently, there is some inconsistency which needs to be fixed.
... we need t have an understanding of what a token means.

chris: people do understand this --- just must iron out errors

hassan: yes, I am confident that it will be straightened out.. Due to my schedule (traveling through month of October), will be hard to totallly finish up, but am trying to get something working.

dave: Jeff Pan has a student Stuart Taylor, who had an implementation of early PS syntax, and is now trying to work on updates to PS syntax; perhaps those efforts can be merged, but possibly not.

hassan: Good. Perhaps we should have a task force on this issue.

chris: Okay, we'll send out msg to find out how many are interested in participating

csma: real priority is finding test cases that we recommend ... rather than working on an implementation.
... we will finally converge, but I'm afraid it will take some time.

hassan: we could use the original PS to express test cases. For that, the tool is already available.

csma: let's take whatever Hassan already has, and just use it.

<Gary> we don't need a tool to write more test cases. Just to execute them, if at some time someone has something that can execute BLD-RIF-XML.

Hassan: but admittedly, the examples are hard to write in original PS

csma: but easier to write the examples in original PS than in xML.

Hassan: I've posted the documentation and results; I can also provide and package the code.
... this takes some time off other thing.

<StellaMitchell> yes, this sounds good

<Harold> q

<Gary> I'd rather hope that we'll have a tool for the APS

I am afraid that fewer people will submit test cases if they know they must submit them in original PS.

As it is, most cases that have been submitted are not in original PS.

<StellaMitchell> leora, they don't have to submit in original PS, we will fix them up for them

Or at least, were not first submitted in original PS.

<sandro> Sandro: We have to have names for these two different versions of the PS, so they can be named in the Test Case

stella, yes, I realize.

<StellaMitchell> sandro, don't you think that will be more confusing

No, I agree with Sandro; it's good to make the distinction clear.

mkifer: Original PS is not subsumed by new PS

<ChrisW> in "abridged" PS: "aaa" is equivalient to "aaa"^^xsd:string

<Hassan> ok

chris: so we move forward in parallel: we continue working on test cases, and as presentation syntax improves, changes are incorporated into test change document, and in parallel, we work on implementation of translation using new spec.

<Hassan> :-)

<StellaMitchell> yes

okay

<sandro> The question is how many of the current test cases are in which syntax?

Sandro, Stella is pretty much the only one who has been careful to work in original PS.

The rest of us have been much sloppier.

<Zakim> csma, you wanted to ask about keeping open the possibility to submit in any language provided someone commits to a translator

So we'll need to rework our cases.

<StellaMitchell> leora, most of the test cases are in ps

<sandro> I guess when we have Hassan's code we can see which test cases actually parse. :-)

<Hassan> Suspense ... :-)

chris: no need for resolution, this will just be our mode of operation.

<sandro> (we could call the the PS's PS1/ and PS/2 :-)

<StellaMitchell> it just makes it harder for reviewers

<StellaMitchell> but ok with me

Stella, It makes it harder for us.

<StellaMitchell> right

Not just the reviewers.

<StellaMitchell> we are part of the reviewers

Right, but we're also the translators.

<csma> satella, why does it meke it more diffivcult for you?

The reviewers will presumably be getting the translated cases?

<csma> language->xml->ps

<StellaMitchell> csma, if we have the reverse translator, it's ok

<csma> hassan just said we have

<DaveReynolds> I think Stuart's old tool worked both ways, might be able to adapt that.

<sandro> ACTION: Sandro implement RIF-XML to PS translator [recorded in http://www.w3.org/2008/09/23-rif-minutes.html#action02]

<trackbot> Created ACTION-579 - Implement RIF-XML to PS translator [on Sandro Hawke - due 2008-09-30].

<ChrisW> PROPOSED: Each RIF test case MUST, eventually, be provided in RIF XML, and MAY be provided in some syntax for which a translator-to-XML has been promised. The translator SHOULD be available for use by the WG to check the translation.

<Gary> sandro, the only "tricky" part is divining the Prefix() to make CURIEs look good

<StellaMitchell> the discussino has covered it

Chris, the resolution seems fine.

<StellaMitchell> seems fine

It's the mode of operation --- working in original PS --- that makes life somewhat harder for Stella, Adrian, and me.

<StellaMitchell> no, leora

<ChrisW> RESOLVED: Each RIF test case MUST, eventually, be provided in RIF XML, and MAY be provided in some syntax for which a translator-to-XML has been promised. The translator SHOULD be available for use by the WG to check the translation.

<sandro> Gary, it's not hard if you don't care how good they look. :-) (seriously, I've written that kind of code 20 times.)

<DaveReynolds> Gary, Jena has a "localname" algorithm that splits off the NCName end off a URI just a few lines of code.

AOB

<Hassan> +1 on the telcon

<csma> +1 to adjourn

- DRAFT -

RIF Telecon 23 Sep 2008

23 Sep 2008

Attendees

Contents