RIF F2F13 -- 17 Apr 2009

<AxelPolleres> http://lists.w3.org/Archives/Public/public-rif-wg/2009Apr/0091.html

<AxelPolleres> (and thread following)

<AxelPolleres> In the light of the new information that owl:real is disjoint from float and double, I'd prefer to drop owl:real as a datatype in RIF.

<ChrisW> calling

<ChrisW> dave, can you hear?

<cke> scribe: cke

<ChrisW> zaim, MIT-G631 contains ChrisW, csma, sandro, Harold, mkifer, AdrianP, cke, AxelPolleres, josb, GaryHallmark, StellaMitchell

Discussion about this morning's agenda

quick resolutions - lists - breakout sessions, etc.

<csma> PROPOSED: Close ISSUE-97 (earlier resolution removed its object).

<ChrisW> issue-97

<ChrisW> issue-97?

<trackbot> ISSUE-97 -- Shoudl Core safeness be restricted to Eiter-Schindlauer safeness -- CLOSED

<trackbot> http://www.w3.org/2005/rules/wg/track/issues/97

chris: we closed it but not resolve explicitely

<csma> RESOLVED: Close ISSUE-97 (earlier resolution removed its object).

next one

<csma> PROPOSED: accept the minutes of April 7

<ChrisW> http://www.w3.org/2009/04/07-rif-minutes.html

<csma> RESOLVED: : accept the minutes of April 7

next one

<csma> PROPOSED: remove owl:real from RIF primitive data types.

<sandro> Axel: Bijan and the owl Wiki confirm that owl:real is disjoint from float and double.

real subsumes rational, which subsumes decimal

<csma> (See Axel's email and following thread: http://lists.w3.org/Archives/Public/public-rif-wg/2009Apr/0091.html)

<sandro> sandro: We should have isNumeric

sandro: I see no reason to keep owl:real

chris: what's the disadvantage? Why drop?

axel: it's confusing. For example ...

<DaveReynolds> Bijan said: "Of course, this would make owl:real harmless to support in RIF, perhaps. [snip] So I don't think it really matters one way or the other, technically. Socially, I think it helps not to have it since having things which are idle (when they are quite significant in other contexts) can be confusing."

<csma> owl:real is the only reference to the owl namespace in RIF

chris: Hierarchy is: real - rational - decimal, no further subclasses

<csma> PROPOSED: drop owl:real from RIF primitive data types.

<DaveReynolds> +1

<AxelPolleres> +1

<AdrianP> +1

<sandro> +1

<MichaelKifer> +1

<ChrisW> 0 see no harm in keeping it

<Harold> +1

<GaryHallmark> +1

<csma> RESOLVED: drop owl:real from RIF primitive data types.

Now on the lists

<ChrisW> http://www.w3.org/2005/rules/wiki/Lists

chris: we agree on the semantics of lists, we agree that the semantics can be described, ...

<csma> PROPOSED: have lists in Core, as described in http://www.w3.org/2005/rules/wiki/Lists, with the builtins listed there.

<sandro> (Some remaining List issues: (1) mapping to/from RDF lists and sequences, (2) indexing )

<sandro> +1

csma: some details need to be refined.

<sandro> Other remaining List issue: can the list constructor syntactically include external or variable terms?

<sandro> (ie what does ground list really mean)

jos: do we keep only grounded lists in Core?, we chose option 1

<csma> PROPOSED: have lists, as described in http://www.w3.org/2005/rules/wiki/Lists, with the builtins listed there (with only immutable ground lists in Core, per previous resolution). Closing ISSUE-

<sandro> issue-95?

<trackbot> ISSUE-95 -- Does RIF need a primitive data type (and associated builtins) for lists? -- OPEN

<trackbot> http://www.w3.org/2005/rules/wg/track/issues/95

<csma> PROPOSED: have lists in BLD, as described in http://www.w3.org/2005/rules/wiki/Lists, with the builtins listed there (with only immutable ground lists in Core, per previous resolution). Closing ISSUE-95

<GaryHallmark> +1

<ChrisW> +1

<sandro> noting that there are still several design issues about lists not yet resolved.

<AdrianP> 0

<Harold> +1

<sandro> +1

<DaveReynolds> 0

<MichaelKifer> +1

<AxelPolleres> +1

<csma> RESOLVED: have lists in BLD, as described in http://www.w3.org/2005/rules/wiki/Lists, with the builtins listed there (with only immutable ground lists in Core, per previous resolution). Closing ISSUE-95.

csma: now need to solve design, editorial issues

<AdrianP> decision on list built-ins was a quick majority vote - no thorough technical reasons for the current selection

<csma> ACTION: michael to add the syntax and semantics of lists to BLD [recorded in http://www.w3.org/2009/04/17-rif-minutes.html#action01]

<trackbot> Sorry, amibiguous username (more than one match) - michael

<trackbot> Try using a different identifier, such as family name or username (eg. msintek, mkifer, merdmann)

we also need to change the schemas in Core, BLD and PRD

<csma> ACTION: mkifer to add the syntax and semantics of lists to BLD [recorded in http://www.w3.org/2009/04/17-rif-minutes.html#action02]

<trackbot> Created ACTION-759 - Add the syntax and semantics of lists to BLD [on Michael Kifer - due 2009-04-24].

<csma> ACTION: Sandro to add the list builtins in DTB [recorded in http://www.w3.org/2009/04/17-rif-minutes.html#action03]

<trackbot> Created ACTION-760 - Add the list builtins in DTB [on Sandro Hawke - due 2009-04-24].

<csma> ACTION: Adrian to add the restriction on list in Core [recorded in http://www.w3.org/2009/04/17-rif-minutes.html#action04]

<trackbot> Created ACTION-761 - Add the restriction on list in Core [on Adrian Paschke - due 2009-04-24].

<csma> ACTION: Harold amends the XML schema for Core. [recorded in http://www.w3.org/2009/04/17-rif-minutes.html#action05]

<trackbot> Created ACTION-762 - Amends the XML schema for Core. [on Harold Boley - due 2009-04-24].

<csma> ACTION: Gary add lists to PRD. [recorded in http://www.w3.org/2009/04/17-rif-minutes.html#action06]

<trackbot> Created ACTION-763 - Add lists to PRD. [on Gary Hallmark - due 2009-04-24].

list of issues now

<ChrisW> ACTION: chris to close issue-95 [recorded in http://www.w3.org/2009/04/17-rif-minutes.html#action07]

<trackbot> Created ACTION-764 - Close issue-95 [on Christopher Welty - due 2009-04-24].

Issues are 98, 77, etc.

Issue 98: update all specs to reference xml scema datatype 1.1

discussion about how wiki can support references to XML, XML data types

chris: we still have to update the documents to use datatypes 1.1

<csma> ACTION: sandro to make the references be to XML schema 1.1 in all RIF documents [recorded in http://www.w3.org/2009/04/17-rif-minutes.html#action08]

<trackbot> Created ACTION-765 - Make the references be to XML schema 1.1 in all RIF documents [on Sandro Hawke - due 2009-04-24].

<csma> PROPOSED: Close issue-98, based on action 765.

<ChrisW> +1

<DaveReynolds> +1

<AdrianP> +1

<Harold> +0.99

axel: xpath function still references datatypes 1.0. Is this a problem?

<AxelPolleres> 0 unclear what the implication with XPath F&O is

sandro: will finf out later

<sandro> sandro: We could ask XS or XP folks ...

<ChrisW> ACTION: Chris to close issue-98 [recorded in http://www.w3.org/2009/04/17-rif-minutes.html#action09]

<trackbot> Created ACTION-766 - Close issue-98 [on Christopher Welty - due 2009-04-24].

<sandro> _+1

<sandro> +1

<csma> RESOLVED: Close issue-98, based on action 765.

next: breakout sessions

csma: editors should meet, they can plan the work, then we' ll see

issue 37

combination of RIF and XML data, and/or XML schemas

<csma> http://www.w3.org/2005/rules/wiki/RIF%2BXML_data-schema

there are 2 strawman proposal in the wiki

csma: one for combining RIF with XML schemas, other: same thing, roughly

First: map frames to XML instance, second: more schema related, linked to the style
... possible to use the XML data without schema, for instance

Second: you need a schema, it's necessary. The ref to data is through the schema

csma: the second option should address the only OO model, while mine is more general
... but they are not incompatibles, we can keep them, or mix the two

sandro: can you show some examples?

csma: look at the example if shiporder fragment

now look at the rule

csma: the doc should be updated with namespaces qualifications
... the RIf rules should be the same regardless the source of data, which can be XML, OWL, etc.
... we just have to change the import section, rules will be portable
... it remains a few minor issues: attribute, namespace, etc.

the other approach is to rely on the xml schemas

<AdrianP> could be generalized to a general approach to allow constructive queries on external data sources, e.g. SPARQL, SQL, XPATH, XQUERY

Gary's proposal starts with a schema

Expressions using some kind of XPath lead to values

it's very straight

<sandro> sandro: you could implement this as a pre-processor, turning an XML document, with or without a schema, into a RIF document consisting of just frame assertions.

But we need typing information. We need the XML schema at the end, otherwise PRD engines cannot work

<sandro> sandro: when you import from some URI, you get a frame like (that_url) [ rif:root-element -> the_root_element ]

<sandro> sandro: this is like SWC, orthogonal to RIF, but defines a way to use RIF with XML data.

mk: Can you put rif:IRI everywhere? Issue with @attribute?

<sandro> sandro: they are rif:iri's that happen to kind of look like xpath expressions.

csma: the expression are not xpath based, even they appear to resemble to xpath

<sandro> poll: -1 don't publish anything on this in the next few months; 0 don't care; +1 publish a WD on this soon

<GaryHallmark> +1

<sandro> +0.75

<csma> +1

<AxelPolleres> 0

<Harold> +1

<MichaelKifer> +1

<AdrianP> +1

+1 (but I would like to refine the technical solution)

<josb> +1

<DaveReynolds> +0.2

<ChrisW> -1

<GaryHallmark> I think we can do this with no new syntax except an import statement

<josb> +1 to no new syntax

<csma> +1 to no new syntax

<sandro> break until 11:10

break till 11:10

<GaryHallmark> e.g. import (myDoc) generates a _myDoc[rif:root->x] frame that points to the root element from myDoc

<sandro> sandro and gary chatting --- (1) PSVI / type information; without it you need to cast in your rules -- import needs a flag about whether you'll get all strings or typed values;

<sandro> --- (2) ordering information -- do you get multiple values and lose the ordering, or do you get rif:Lists for everything? that's another flag on import.

<sandro> --- (3) what about striping? gary says we can probably just assuming the element names are all property names -- it's like an OO x.y.z. expression; sandro says yeah, that might work, maybe.

<sandro> --- (4) id/idref

FLD

<csma> scribe: John Hall

<csma> scribenick: johnhall

Michael: organized FLD for more dialects
... added aggregates and modules
... now need to add lists and complete the XML work

ChrisW: same as BLD

Michael: yes
... XML schema syntax, now harder

Chris: to account for generalized quntifiers

Michael: feedback from last draft - from mail archives

josb: last draft July 2008

ChrisW: lists and XML syntax - all that needs to be done
... imports - specialization of modules

Michael: didn't do
... How to proceed with other dialects?

csma: another agenda item

ChrisW: for FLD to pass to rec we need implementations
... individuals should work on dialects and submit as member contributions

Michael: cumbersome

Sandro: costs a lot to keep a group running
... regardless of whether WG continues with FLD, we need people to implement and report back

Harold: we need 2 implementations of FLD. We have one - what could be the other?

ChrisW: in order for FLD to reach rec, other implementations don't need to standards

last call plan

ChrisW: wiki page called 'Last Call Plans'

<csma> http://www.w3.org/2005/rules/wiki/LastCallPlan

ChrisW: most important - 6 documents for last call
... are we going to release BLD with a new last call?
... adding lists seems significant

josb: predicates with multiple arities

ChrisW: big impact on implementations

Sandro: need to give people a chance to object

ChrisW: second last call for BLD?

csma: restrict scope of comments to just the concerns here

Sandro: WG is not gong to make changes except in response to external input

<sandro> sandro: While I think the bulk of the WG wants NAUs gone, I don't think it's appropriate for us be making changes like that now.

ChrisW: plan for future is to have enough meetings to deal with external input

Harold: risk of losing commitment from WG members' employers

Sandro: all want to get to rec

ChrisW: restrict BLD comment to just the changes since last 'last call'?

<Harold> http://www.w3.org/2005/10/Process-20051014/tr.html#last-call

csma: difficult to manage

Michael: list of changes to BLD

Harold: look at 'last-call' process. What are we allowed to do next?

Michael: we say we are acting in good faith and must provide help to implementers

ChrisW: significant BLD changes are only arities and lists

Sandro: easier to manage of everything foes to last call

josb: How can you have a BLD implementation without DTB?

Harold: risk of what might come in

csma: if there is a show-stopper, we need to know

ChrisW: looking in detail at the list, less concerned about going to last call

csma: would prefer to go with all as a whole

Sandro: we can include a paragraph about a second last call, and why it might be needed

<sandro> sandro: We'll explain in the Second Last Call paragraph that the biggest changes were the inclusion of lists and allowing multiple arity symbols, but that the main reason for LC2 is just to keep all of RIF in sync.

Michael: if we don't do LC2

Sandro: options are LC2, candidate recommendation, member contribution
... downplay the different specs, present RIF as one thing

<sandro> kifer: LC2 --- give people a chance to comment on rif-as-a-whole.

Michael: say that we want to give people a chance to comment on RIF as a whole

<sandro> PROPOSED: Our next publication of BLD will be as a second Last Call

<sandro> +1

<ChrisW> +1

<MichaelKifer> 0

<cke> +1

<sandro> (it's simpler)

<Harold> 0

<csma> +1

<josb> +1

<AxelPolleres> +1

<GaryHallmark> +1

<StellaMitchell> +1

<AdrianP> +1

<sandro> RESOLVED: Our next publication of BLD will be as a second Last Call

<sandro> note that the same is true for SWC, but it's not controversial

<sandro> PROPOSED: Our next publication of SWC will be as a second Last Call

<sandro> +1

<ChrisW> +1

<josb> +1

<AdrianP> +1

<Harold> +1

<MichaelKifer> +1

<csma> +1

<sandro> RESOLVED: Our next publication of SWC will be as a second Last Call

ChrisW: need 2 reviewers for each document
... everyone does 1, can do more

<sandro> ACTION: harold review PRD [recorded in http://www.w3.org/2009/04/17-rif-minutes.html#action10]

<trackbot> Created ACTION-767 - Review PRD [on Harold Boley - due 2009-04-24].

csma: editors do not review their own documents

<sandro> ACTION: cke review Core [recorded in http://www.w3.org/2009/04/17-rif-minutes.html#action11]

<trackbot> Created ACTION-768 - Review Core [on Changhai Ke - due 2009-04-24].

<sandro> ACTION: cke review PRD [recorded in http://www.w3.org/2009/04/17-rif-minutes.html#action12]

<trackbot> Created ACTION-769 - Review PRD [on Changhai Ke - due 2009-04-24].

<sandro> ACTION: jos review Core [recorded in http://www.w3.org/2009/04/17-rif-minutes.html#action13]

<trackbot> Sorry, amibiguous username (more than one match) - jos

<trackbot> Try using a different identifier, such as family name or username (eg. jdebruij2, jderoo)

<sandro> ACTION: josb review Core [recorded in http://www.w3.org/2009/04/17-rif-minutes.html#action14]

<trackbot> Created ACTION-770 - Review Core [on Jos de Bruijn - due 2009-04-24].

<sandro> ACTION: josb review DTB [recorded in http://www.w3.org/2009/04/17-rif-minutes.html#action15]

<trackbot> Created ACTION-771 - Review DTB [on Jos de Bruijn - due 2009-04-24].

<sandro> ACTION: axel review swc [recorded in http://www.w3.org/2009/04/17-rif-minutes.html#action16]

<trackbot> Created ACTION-772 - Review swc [on Axel Polleres - due 2009-04-24].

<sandro> ACTION: axel review BLD [recorded in http://www.w3.org/2009/04/17-rif-minutes.html#action17]

<trackbot> Created ACTION-773 - Review BLD [on Axel Polleres - due 2009-04-24].

<sandro> ACTION: chris review FLD [recorded in http://www.w3.org/2009/04/17-rif-minutes.html#action18]

<trackbot> Created ACTION-774 - Review FLD [on Christopher Welty - due 2009-04-24].

<sandro> ACTION: adrian review dtb [recorded in http://www.w3.org/2009/04/17-rif-minutes.html#action19]

<trackbot> Created ACTION-775 - Review dtb [on Adrian Paschke - due 2009-04-24].

<sandro> ACTION: csma review bld [recorded in http://www.w3.org/2009/04/17-rif-minutes.html#action20]

<trackbot> Created ACTION-776 - Review bld [on Christian de Sainte Marie - due 2009-04-24].

<sandro> ACTION: josb review prd (lower priority than his other reviews) [recorded in http://www.w3.org/2009/04/17-rif-minutes.html#action21]

<trackbot> Created ACTION-777 - Review prd (lower priority than his other reviews) [on Jos de Bruijn - due 2009-04-24].

<sandro> ACTION: stella review FLD [recorded in http://www.w3.org/2009/04/17-rif-minutes.html#action22]

<trackbot> Created ACTION-778 - Review FLD [on Stella Mitchell - due 2009-04-24].

<sandro> ACTION: gary review SWC [recorded in http://www.w3.org/2009/04/17-rif-minutes.html#action23]

<trackbot> Created ACTION-779 - Review SWC [on Gary Hallmark - due 2009-04-24].

ChrisW: actions on core (from last call page)

josb: Core - need to deal with some problems on formulas
... section 2.3 of Core spec

ChrisW: get rid of second bullet?

csma: resolution - external can appear everywhere a term can

ChrisW: nothing at risk in Core
... actions on BLD

csma: resolution to modularize schema

Harold: is already modularized
... if we have a lot of includes, from experience is harder to maintain

Gary: there is substantial duplication

Harold: Core is not stable

csma: there is a resolutin that Core schema will be included in BLD and PRD

cke: we need to explain the architecture

Gary: would me willing to refactor the schema

<ChrisW> ACTION: gary to refactor BLD schema to import Core (help from cke) [recorded in http://www.w3.org/2009/04/17-rif-minutes.html#action24]

<trackbot> Created ACTION-780 - Refactor BLD schema to import Core (help from cke) [on Gary Hallmark - due 2009-04-24].

csma: why does FLD have a schema - it's not a dialect

Michael: it's a framework

ChrisW: copy and edit from BLD?
... if so, relationships need to me maintained over imports
... actions on SWC

josb: three more actions added

ChrisW: actions on DTB

Axel: need to refine definitions of mapping for all informal built-ins

ChrisW: actions on PRD

csma: four items added

<sandro> $ echo tag$RANDOM

<sandro> tag30193

<Harold> NmNot could be renamed to Inot (for Inflationary not).

Adrian: tag for 'pnot'

<AdrianP> semantics of PRD probably needs to be revisited - see PRD test cases on retract, assert combinations

ChrisW: actions on Core

Harold: ready for review by April 30

ChrisW: actions on BLD

Michael: ready for review April 24, except XML part

Harold: XML modularization by April 30

ChrisW: actions on SWC

josb: May 4

<ChrisW> ACTION: josb to write some list test cases [recorded in http://www.w3.org/2009/04/17-rif-minutes.html#action25]

<trackbot> Created ACTION-781 - Write some list test cases [on Jos de Bruijn - due 2009-04-24].

ChrisW: actions on DTB

Axel: April 30

ChrisW: actions on PRD

csma: April 30

ChrisW: actions on FLD
... all documents ready for review by May 4
... last call vote on May 12

<AxelPolleres> scribe: Axel Polleres

<AxelPolleres> scribenick: AxelPolleres

future of the WG

csma: we talk about the time between and of may and end of november
... if we have all docs at lc end of may, we are done basically.

chrisw: 3 work items remaining.
... 1) comments 2) tests 3) xml data

<sandro> chrisw: three things to do after LC: responding to public comments, test, and xml-data.

chrisw: reduced time committments, no more weekly telecons.

csma: probably we sshould have monthly, and more as necessary.
... monthly sync point seems reasonable.

chrisw: possibly adding task forces as we do now with specific topics.

sandro: we miss one item: implementations.

chrisw: deadline for comments 4 weeks from lc... probably end of june.

sandro: we probably publish may 19
... comment deadline june 16
... depends a lot how many comments we get and how hard they are to implement.

chrisw: what is the cr period?

sandro: runs as long as you need to get implementations.
... to some extent we can define ourselves what "implementations" mean.
... general rules whould be tewo interoperating implementations per dialect.

jos: means 4 implementations, 2 prd, 2 bld which are interoperable.

sandro: should we target CR mid july?

<sandro> LC pub on 19 May, comment deadline 16 June, ...

csma: that is optimistic.

<sandro> Optimisitcally, go to CR 15 July

csma: announcement must indicate a minimum duration.

sandro: do we need producers or jsut consumers?

cke: we should do both.

josb: what do we need for SWC? somewhat swc defines a bunch of new dialects.

mk: swc with OWL DL needs a SWRL implementation

josb: there are implementations of dl safe swrl.
... someone could make an implementation on top of that.
... I assume the jena people will do something for Core+RDF?

axel: depends whether I find someone to write a parser/compiler to dlvhex.

harold: flora-2?

mk: I would add a new reader, I guess.

sandro: you (mk) would expcet do have a BLD implementation?

mk: also a matter of finding a student.

csma: finding people to implement is time consuming.

sandro: so, do we need a 1month CR or a 6month CR?
... if we want PR in nov we need CR in oct.
... oct 1 or oct 15 for CR deadline?
... to accomodate for management time in between.

<sandro> October 1 for end of CR?

<sandro> kifer: I wont be ready by then....

mk: willl probably not be ready by then.

<sandro> kifer: I dont have any student to do this right now....

sanrdo: in terms of all the work with the DTB built-ins.
... we don't need all BLD require all of DTB.

<Harold> We may mostly need translators between RIF and the languages of existing engines, rather than brand-new engines.

axel: what of DTB is in BLD and not in Core?

<sandro> Implementations need: 2 BLD, 2 DTB, 2 SWC-with-Core, 2-DTB-with-BLD.

<sandro> kifer: incompete notion of equality....?

sandro: I meant not all of the built-ins, not "core" in the sense of Core.

mk: the reason why implementation is at risk because a complete implementation is hard.

<sandro> (I don't know....)

mk: i have no idea how far ontoprise will get.

chrisw: we should assign actions to gather implementations.
... hopefully no later than the end of the year.

mk: producer is much easier than consumer.

<ChrisW> ACTION: christian to talk to ontoprise about RIF implementation & timeframe [recorded in http://www.w3.org/2009/04/17-rif-minutes.html#action26]

<trackbot> Created ACTION-782 - Talk to ontoprise about RIF implementation & timeframe [on Christian de Sainte Marie - due 2009-04-24].

chrisw: who to contact ontoprise?

<AdrianP> implementation plans

<AdrianP> http://www.w3.org/2005/rules/wg/wiki/Implementation_plans

<ChrisW> ACTION: Chris to talk to Dave about implementation [recorded in http://www.w3.org/2009/04/17-rif-minutes.html#action27]

<trackbot> Created ACTION-783 - Talk to Dave about implementation [on Christopher Welty - due 2009-04-24].

<Harold> Re Equality we could weaken the requirement on implementors to allow for certain logical incomplete implementations such as those that have only oriented equations.

<ChrisW> ACTION: christian to talk to Mark Proctor about RIF implementation [recorded in http://www.w3.org/2009/04/17-rif-minutes.html#action28]

<trackbot> Created ACTION-784 - Talk to Mark Proctor about RIF implementation [on Christian de Sainte Marie - due 2009-04-24].

chrisw: any other implementations we haven't talked about?

sandro: planning some BLD implementation, if not all builtins are required.

<ChrisW> sandro plans an implementation - Oct 1 seems reasonable

<Harold> http://www.w3.org/2005/rules/wg/wiki/Implementation_plans

harold: I contacted igor.

<AdrianP> http://www.w3.org/2005/rules/wg/wiki/List_of_Rule_Systems

chrisw: everyboduy fine with the plan?

csma: w3c tech plenary is nov 1st
... in santa clara, CA.
... ac meeting there, around the time we go to CR.
... I will probably go, a lot of AC Reps there.

axel: might try to go as well.

csma: need for a latest f2f before that?

sandro: don't think so.

<sandro> PROPOSED: Ask to extend the Working Group by 6 months, through 30 November 2009. We will not plan for any more F2F meetings, and expecting fewer telecons.

<sandro> +1

<AdrianP> +1

<ChrisW> +1

<csma> +1

<josb> +1

<Harold> +1 (or 9 months :-)

<sandro> RESOLVED: Ask to extend the Working Group by 6 months, through 30 November 2009. We will not plan for any more F2F meetings, and expecting fewer telecons.

test cases

csma: revisit approved ones, volunteers for new ones.

axel: do we need testcases for all built-ins?
... would one TC be ok that calls ALL built-ins?

<sandro> re taxi: <Ralph> they could call a taxi for 3:30 and wait on Vassar outside the Gates entrance or they could walk up to the Marriott and get a taxi from the stand

<scribe> ACTION: axel to provide an ALL-builtins testcase. [recorded in http://www.w3.org/2009/04/17-rif-minutes.html#action29]

<trackbot> Created ACTION-785 - Provide an ALL-builtins testcase. [on Axel Polleres - due 2009-04-24].

sandro: we need testsuite done mid july.

chrisw: where to start in our review?

<sandro> Every Core test MUST also work in all PRD and BLD engines.

discussion whether TCs can be labelled just core

http://www.w3.org/2005/rules/wiki/Argument_names_in_uniterms_3

sandro: XML is wrong.

<josb> http://www.w3.org/2005/rules/wiki/BLD#Alphabet_of_RIF-BLD

jos: according to spec named args are unicodes strings.

sandro: the testcase should be positive

I asked already what about a->b->c and strange stuff like that.

<sandro> the negative test case in the XML would be ANY kind of markup in name. The name is just text.

csma: that means that this TC is not approved?

or: "asdasd->asdasd"->"asdasd"

<sandro> POSITIVE SYNTAX: <Name>"http://example.com/example#color"^^xs:string</Name>

<sandro> NEGATIVE SYNTAX: <Name><Const type="&xs;string">green</Const>

<sandro> NEGATIVE SYNTAX: <Name><Const type="&xs;string">green</Const></Name>

chrisw: is it possible for the exact same literal to be a const and an arg name?

csma/jos: Yes!

chrisw: why is there no ps for this?

<sandro> chris: So you can't write the bad-xml in the PS. Okay.

<sandro> POSITIVE SYNTAX: <Name><Const type="&xs;string">green</Const></Name>

chrisw: the thing that shouldn't parse can't be said in PS.

<sandro> everyone: okay.

<sandro> everyone: yeah, of course.

chrisw: shall we drop the test?
... what about http://www.w3.org/2005/rules/wiki/Argument_names_in_uniterms_1
... basically the same... so we drop all these?

<sandro> chris: the thing in the argname position is NEVER a rif Const.

mk: I disagree.

discussion on what is a symbol/constant in the language

PROPOSED: drop http://www.w3.org/2005/rules/wiki/Argument_names_in_uniterms_1-4

<josb> +1

<ChrisW> 0

<MichaelKifer> 0

<sandro> +1

<AdrianP> 0

<Harold> +1

+1->":-)"

<sandro> these test cases speak to an important confusion in NAUs, but they don't actually clarify it right now.

RESOLUTION: drop http://www.w3.org/2005/rules/wiki/Argument_names_in_uniterms_1-4

<ChrisW> ACTION: Stella to makr argnamesinuniterms as rejected [recorded in http://www.w3.org/2009/04/17-rif-minutes.html#action30]

<trackbot> Created ACTION-786 - Makr argnamesinuniterms as rejected [on Stella Mitchell - due 2009-04-24].

<sandro> drop == mark as "Rejected"

<ChrisW> http://www.w3.org/2005/rules/wiki/Assert

<Harold> A solution would be to be (much) more restrictive wrt what unicode strings are allowed to represent ArgNames.

subtopic: http://www.w3.org/2005/rules/wiki/Assert

<sandro> csma: the name does not need to be repeated.

axel: is assert a constant?

csma: no, a language primitive in PRD.

<sandro> sandro: YOu can't have a PRD test case using just assert, since then it's core.

<sandro> csma: right.

<Harold> ArgNames, as 'indexes' into terms, could, e.g., be restricted to just *alphanumeric* ASCII unicode strings (disallowing the characters for double-quote and hat among other characters).

<sandro> instead have: forall ?x if ?x[ex:status -> "gold"] then ?x[ex:discount -> 10]

<sandro> but this is a Core test case.

<sandro> could be done as a PRD test case if we add modify, or retract+assert.

chrisw: adrian, can you fix that one?

<ChrisW> ACTION: Adrian to change Assert test case as in meeting minutes [recorded in http://www.w3.org/2009/04/17-rif-minutes.html#action31]

<trackbot> Created ACTION-787 - Change Assert test case as in meeting minutes [on Adrian Paschke - due 2009-04-24].

subtopic: http://www.w3.org/2005/rules/wiki/AssertRetract

<sandro> +1 harold -- ArgNames being just [_a-zA-z][_a-zA-Z0-9]*

we had that discussion on argnames before...

<Harold> Right, but it's some progress to say we should restrict their alphabet.

it would be (I just gave up on it last time, nobody seemed to be willing to restrict it)

<ChrisW> ACTION: adrian to modify AssertRetract in the same way as Assert [recorded in http://www.w3.org/2009/04/17-rif-minutes.html#action32]

<trackbot> Created ACTION-788 - Modify AssertRetract in the same way as Assert [on Adrian Paschke - due 2009-04-24].

subtopic: http://www.w3.org/2005/rules/wiki/Retract
... http://www.w3.org/2005/rules/wiki/Modify

both the same issue?.

gary: .../Retract should be a negative entailment test

jos: empty set is entailed by everything.

<ChrisW> ACTION: adrian to update Retract test case as with assert, fixing object retraction [recorded in http://www.w3.org/2009/04/17-rif-minutes.html#action33]

<trackbot> Created ACTION-789 - Update Retract test case as with assert, fixing object retraction [on Adrian Paschke - due 2009-04-24].

subtopic: http://www.w3.org/2005/rules/wiki/Class_Membership

adrian: I simplified that one.

<sandro> take our parent and child as classes and make male be a class.

axel: I suggest to replace classes child and parent with class person

chrisw: remove the classes at all.
... but make Male a class.

Adrian: nested molecules a la f-logic would be nice but not allowed in RIF.

<ChrisW> ACTION: adrian to fix ClassMembership and rename [recorded in http://www.w3.org/2009/04/17-rif-minutes.html#action34]

<trackbot> Created ACTION-790 - Fix ClassMembership and rename [on Adrian Paschke - due 2009-04-24].

<sandro> sandro: it would be NICE to allow membership predicates as terms....

<sandro> chris: yeah, whatever. :-)

subtopic: http://www.w3.org/2005/rules/wiki/Classification-inheritance

chrisw: that is a bad thing to do, should we mark that?

mk: this is basically default values.
... not really a default, it would override all.

<sandro> gary: I, P, and V should be quantified

<sandro> chris: We need to say this is a REALLY BAD PRACTICE.

chrisw: can be approved modulo fixing XML.

<sandro> PROPOSED: approve test case http://www.w3.org/2005/rules/wiki/Classification-inheritance

<Harold> +1

<ChrisW> +1

<sandro> +1

<MichaelKifer> +1

<josb> +1

<sandro> RESOLVED: approve test case http://www.w3.org/2005/rules/wiki/Classification-inheritance

subtopic: http://www.w3.org/2005/rules/wiki/Conflict_resolution

<ChrisW> ACTION: cke to write test cases for PRD [recorded in http://www.w3.org/2009/04/17-rif-minutes.html#action35]

<trackbot> Created ACTION-791 - Write test cases for PRD [on Changhai Ke - due 2009-04-24].

<sandro> PROPOSED: Approve http://www.w3.org/2005/rules/wiki/Conflict_resolution

<sandro> +1

<josb> +1

<ChrisW> +1

<Harold> +1

<csma> +1

<AdrianP> +1

<sandro> RESOLVED: Approve http://www.w3.org/2005/rules/wiki/Conflict_resolution

<cke> +1

<MichaelKifer> +1

<ChrisW> http://www.w3.org/2005/rules/wg/meeting/2009-04-17

<ChrisW> scribe: MichaelKifer

<sandro> http://www.w3.org/2005/rules/wg/meeting/2009-04-17

core non-safeness test case

sandro: negative tests must be explicit as to which dialect extensions it applies to

<sandro> A negative syntax test for PRD is be definition a negative syntax test for Core.

jos: changed the dialect of the example to prd from core

<ChrisW> ACTION: Stella to add note in test case document that negative tests 'go down' and positive tests "go up" [recorded in http://www.w3.org/2009/04/17-rif-minutes.html#action36]

<trackbot> Created ACTION-792 - Add note in test case document that negative tests 'go down' and positive tests "go up" [on Stella Mitchell - due 2009-04-24].

<ChrisW> PROPOSED: accept http://www.w3.org/2005/rules/wiki/Core_NonSafeness

<ChrisW> +1

<cke> +1

<josb> +1

<sandro> as a PRD case.

<sandro> +1

<ChrisW> RESOLVED: accept http://www.w3.org/2005/rules/wiki/Core_NonSafeness

<Harold> +1

<sandro> no need to rename.

<ChrisW> PROPOSED: accept http://www.w3.org/2005/rules/wiki/Core_NonSafeness2

<AdrianP> +1

<ChrisW> PROPOSED: accept http://www.w3.org/2005/rules/wiki/Core_NonSafeness_2

<ChrisW> +1

<ChrisW> RESOLVED: accept http://www.w3.org/2005/rules/wiki/Core_NonSafeness_2

<ChrisW> PROPOSED: accept http://www.w3.org/2005/rules/wiki/Core_Safeness

<Harold> +1

<ChrisW> +1

<ChrisW> RESOLVED: accept http://www.w3.org/2005/rules/wiki/Core_Safeness

<AdrianP> +1

<AxelPolleres> +1

<GaryHallmark> +1

jos: too lazy to type +1

<GaryHallmark> +1

<Harold> +1

<ChrisW> PROPOSED: accept http://www.w3.org/2005/rules/wiki/Core_Safeness_2

<ChrisW> RESOLVED: accept http://www.w3.org/2005/rules/wiki/Core_Safeness_2

<AdrianP> +1

<GaryHallmark> +1

<ChrisW> PROPOSED: Accept http://www.w3.org/2005/rules/wiki/Core_Safeness_3

<Harold> +1

<ChrisW> +1

<AdrianP> +1

<ChrisW> RESOLVED: Accept http://www.w3.org/2005/rules/wiki/Core_Safeness_3

<ChrisW> PROPOSED: accept http://www.w3.org/2005/rules/wiki/EntailEverything

<ChrisW> +1

<sandro> +1

<AxelPolleres> +1

<AdrianP> +1

<ChrisW> RESOLVED: accept http://www.w3.org/2005/rules/wiki/EntailEverything

<Harold> +1

<sandro> Any BLD or PRD test cases, it's a NEGATIVE SYNTAX test for Core. Otherwise, it would be marked as a Core test.

<ChrisW> PROPOSED: accept http://www.w3.org/2005/rules/wiki/Factorial_Functional

<ChrisW> +1

<GaryHallmark> +1

<StellaMitchell> +1

<sandro> +1 wow it'll be hard to implement.

<ChrisW> RESOLVED: accept http://www.w3.org/2005/rules/wiki/Factorial_Functional

<AdrianP> +1

<AxelPolleres> +1

<sandro> gary: do a core version of this.

<ChrisW> ACTION: GARY to write a core version of factorial [recorded in http://www.w3.org/2009/04/17-rif-minutes.html#action37]

<trackbot> Created ACTION-793 - Write a core version of factorial [on Gary Hallmark - due 2009-04-24].

<ChrisW> PROPOSED: accept http://www.w3.org/2005/rules/wiki/Factorial_Relational

<ChrisW> +1

<sandro> Gary: I would expect the test-taker to manually guide how many iterations they run to get the required entailment, so it wont "run forever"

<StellaMitchell> +1

<ChrisW> RESOLVED: accept http://www.w3.org/2005/rules/wiki/Factorial_Relational

<GaryHallmark> +1

<AxelPolleres> +1

<ChrisW> PROPOSED: accept http://www.w3.org/2005/rules/wiki/IRI_from_IRI

<GaryHallmark> +1

<AdrianP> +1

<AxelPolleres> +1

<ChrisW> RESOLVED: accept http://www.w3.org/2005/rules/wiki/IRI_from_IRI

<sandro> wow, the one is totally twisted. I like it.....

<sandro> +1

sandro: the previous case was nasty

<Harold> +1

<AxelPolleres> xmlns:foaf="http://xmlns.com/foaf/0.1/"

<AxelPolleres> +1

sandro: this test is worth living for

<ChrisW> PROPOSED: accept http://www.w3.org/2005/rules/wiki/IRI_from_RDF_Literal

<Harold> +1

<AxelPolleres> +1

<ChrisW> RESOLVED: accept http://www.w3.org/2005/rules/wiki/IRI_from_RDF_Literal

<Harold> Before LC, we could bring in a few extra meaningful test cases, e.g. by Jos De Roo for the Euler Proof Mechanism (http://www.agfa.com/w3c/euler/): http://www.agfa.com/w3c/euler/authen.axiom.n3

<AdrianP> +1

<sandro> sandro: So this means that in practice, to implement your test cases, you'll probably have to rename all you rif-locals

<sandro> jos: correct.

<ChrisW> PROPOSED: accept http://www.w3.org/2005/rules/wiki/Local_Constant

<ChrisW> +1

<Harold> +1

<StellaMitchell> +1

<ChrisW> RESOLVED: accept http://www.w3.org/2005/rules/wiki/Local_Constant

<AdrianP> +1

<GaryHallmark> +1

<sandro> +1 although i don't really like it.

note: conclusions are not part of the document that entails them (to keep in mind for the tests that use rif:local).

<GaryHallmark> +1

<ChrisW> PROPOSED: accept http://www.w3.org/2005/rules/wiki/Local_Predicate

<Harold> +1

<AxelPolleres> +1

<AdrianP> +1

<ChrisW> RESOLVED: accept http://www.w3.org/2005/rules/wiki/Local_Predicate

<ChrisW> PROPOSED: adjourn

<GaryHallmark> +1

<AxelPolleres> -1 let's continue! :-)

<ChrisW> PROPOSED: gang up on Axel

<AxelPolleres> -1 on that one , then I rather accept adjourning

<josb> more nasty string-iri stuff: http://www.w3.org/2005/rules/wiki/Multiple_IRIs_from_String

- DRAFT -

RIF F2F13

17 Apr 2009

Attendees

Contents

issue 37

FLD

last call plan

future of the WG

test cases

Summary of Action Items

Scribe.perl diagnostic output