Admin

<ChrisW> Scribe: Mike_Dean

<ChrisW> http://lists.w3.org/Archives/Public/public-rif-wg/2008Jun/att-0042/10-June-2008-rif-minutes.html

<ChrisW> PROPOSED: Accept last week's minutes

<Harold> +1

+1

<ChrisW> RESOLVED: Accept last week's minutes

Harold: no agenda item about equality

Liaison

Adrian: HCLS interested in liaison with RIF
... on use cases

Adrian will be liaison

F2F11

confusion due to Chris's travel constraints - need to meet 1 day on weekend - Eastern US

early September or late August

need all editors to attend

Axel can't attend proposed dates

still working on dates

Action Review

<ChrisW> ACTION-524: closed

<trackbot> ACTION-524 Start the wiki page for the active members of the WG notes added

<trackbot> If you meant to close ACTION-524, please use 'close ACTION-524'

<ChrisW> ACTION-524: completed

<trackbot> ACTION-524 Start the wiki page for the active members of the WG notes added

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

text for acknowledgments section

scribe: contributors, chairs, regular WG attendees

<AdrianP> action-523 is done

Harold: some updates from Michael to BLD

<Harold> Hi Michael?

ChrisW: change editor's notes to at risk notes

<ChrisW> ACTION: Chris make sure DTB discusses URI base directive [recorded in http://www.w3.org/2008/06/17-rif-minutes.html#action02]

<trackbot> Created ACTION-526 - Make sure DTB discusses URI base directive [on Christopher Welty - due 2008-06-24].

Michael: BLD doesn't use any duration subtypes
... should reference DTB

ChrisW: list datatypes in BLD (fixed) with pointer to DTB

Michael: OK - no duration

Casting

discussion on email list

scribe: Jos and Axel not here

ChrisW: seemed to reach conclusion

Michael: agree with Jos

Publication plan

ChrisW: revisit when everyone here

Harold: BLD pretty good - should finish edits in 2 weeks, extending a week
... need to coordinate with Michael
... metadata throughout - expand both syntaxes and examples
... makes grammars harder to read - perhaps publish with and without

ChrisW: XML syntax is more important than presentation
... BNF

Michael: in mathematical representation as well - first action

Harold: have tested XSD
... would like someone else to try

<Harold> http://www.w3.org/2005/rules/wiki/BLD#Appendix:_XML_Schema_for_RIF-BLD

<Harold> Especially:

<ChrisW> ACTION: Gary to test the XSD examples http://www.w3.org/2005/rules/wiki/BLD#Appendix:_XML_Schema_for_RIF-BLD [recorded in http://www.w3.org/2008/06/17-rif-minutes.html#action03]

<trackbot> Created ACTION-527 - Test the XSD examples http://www.w3.org/2005/rules/wiki/BLD#Appendix:_XML_Schema_for_RIF-BLD [on Gary Hallmark - due 2008-06-24].

<Harold> http://www.jdrew.org/rif/bldLC/

<Harold> (xsds plus examples)

<ChrisW> ACTION: Hallmark to test the XSD examples http://www.w3.org/2005/rules/wiki/BLD#Appendix:_XML_Schema_for_RIF-BLD [recorded in http://www.w3.org/2008/06/17-rif-minutes.html#action04]

<trackbot> Created ACTION-528 - Test the XSD examples http://www.w3.org/2005/rules/wiki/BLD#Appendix:_XML_Schema_for_RIF-BLD [on Gary Hallmark - due 2008-06-24].

<Harold> Gary, I used W3C's XSV to test them.

<ChrisW> ACTION: mkifer to add data type list to BLD [recorded in http://www.w3.org/2008/06/17-rif-minutes.html#action06]

<trackbot> Created ACTION-529 - Add data type list to BLD [on Michael Kifer - due 2008-06-24].

<Harold> XSV: http://www.w3.org/2001/03/webdata/xsv

Harold: need to use consistent namespaces
... need about 10 days for updates

ChrisW: freeze BLD on July 27

<ChrisW> BLD: Freeze on 27th

ChrisW: need someone to review final BLD draft diffs

<LeoraMorgenstern> I can review

<ChrisW> ACTION: leora to review final BLD LC draft [recorded in http://www.w3.org/2008/06/17-rif-minutes.html#action07]

<trackbot> Created ACTION-530 - Review final BLD LC draft [on Leora Morgenstern - due 2008-06-24].

<AdrianP> added a Java script which allows hiding the BLD and PRD examples http://www.w3.org/2005/rules/wiki/UCR#Use_Cases; the document is now much more compact and the use cases are readable

Adrian: several updates to UCR - JavaScript to hide code examples - now more readable

<AdrianP> added a very compact (abridged) presentation syntax + some conventions http://www.w3.org/2005/rules/wiki/UCR#Use_Cases, close to ISO Prolog standard, but using e.g. <- for both PRD and BLD (to describe them in an independet and interchangeable way)

<scribe> ... new examples of abridged presentation syntax

<AdrianP> reworked the BLD and PRD examples of the use cases as far as possible

<AdrianP> updated requirements according to our resolutions

<AdrianP> removed discussion of implications of requirements from the use cases

ChrisW: very nice

<AdrianP> open issue: examples overflows the right border of the containing box

<AdrianP> open issue: coordinate with PRD presentation syntax

Adrian: not much more to do, but depends on BLD and PRD finalization

<Harold> Why is '<-' better than ':-'?

Adrian: independent for BLD and PRD

<AdrianP> the idea is

<AdrianP> [Actions] <- [Conditions]

<AdrianP> [Conclusion] <- [Conditions]

<Harold> Well, PRD actions are not the exact same things as BLD conclusions.

<markproctor> when LHS then RHS

Christian: production rule community uses LHS (condition) and RHS (action) - reverse will be a problem

<GaryHallmark> clearly there is no standard, so we can do what we want....

<AdrianP> [Condition] -> [Actions]

<AdrianP> the problem is the -> which is also used in frames

Christian: divergence between LP and PR communities

<GaryHallmark> we are here in part to bridge the 2 communities

Christian: don't need common presentation syntax that will only confuse either community

<GaryHallmark> is everyone stupid?

<GaryHallmark> if we explain what we mean, they will understand

Christian: other community won't use presentation syntax

<Harold> In the XML syntax we could have something equivalent to -- BLD: if [Conditions] then [Conclusion] vs. PRD: if [Conditions] do [Action]

<MichaelKifer> +q

<AdrianP> derivation rules are often written as conditions -> conclusion

Christian: would object to non-standard form of negation

<markproctor> for me this comes from the idea of interchange between production rule and prolog rules. that really only interests fringe use cases, and better solved else where.

Christian: PRD should be basic PR dialect

<AdrianP> not sure about inflationary semantics for negated procuction rules as unifying semantics. One of the problems with inflationary semantics is that if rule designers don't write the rules carefully, the program winds up asserting certain facts ``too soon,'' that is, using a negated literal (not assumption) that eventually turns out not to be true (i.e., it is of the form NOT p, where p eventually becomes true)

<markproctor> the way that 'not' and truth maintenance work in production rules has real value to real users.

Christian: all deployed systems use same form of negation

<markproctor> jess, drools, ilog, opsj every PR system works the say with 'not'.

<Harold> Christian, also in LP dialects you would normally NEED negation (as failure).

<Harold> Michael is on the q ;-)

ChrisW: base that can be extended in different implementations

<Harold> Yet, we are doing a BASIC logic dialect.

<markproctor> by experimenting with 'not' you refering to those silly non inference sequential engines? ;)

<AdrianP> just to clarify, I see negation as very important for both logic dialect and production rule dialect

Gary: agree on consensus within PR community

<Harold> q

Gary: no problems with negation
... for eval, need examples to show utility
... continue to have problems with running example not using presentation syntax (too much like iLog) - artificially difficult to compare to BLD
... presentation syntax 95% similar to BLD
... remove formulas off forall

Christian: do not need 1 presentation syntax

Gary: ad hoc syntax

<AdrianP> but it is good for the user to have one presentation syntax in introduction which is used in all RIF documents and show a mapping to the concrete XML serialization

Christian: examples should be in presentation syntax if we have one - should translate them

Gary: please put everything in one condition
... use BLD syntax where it's common

<AdrianP> the objective is to enable interchange

<markproctor> we talking about this?

<AdrianP> also between logical rule engines and production rule engines

<markproctor> http://www.w3.org/2005/rules/wiki/PRD#Syntactic_compatibility_between_RIF-PRD_and_RIF-BLD

<AdrianP> whenever possible!!!

<AdrianP> business rules, policies, contract etc. they are often not specific to production rules, logical rules, decision tables etc.

ChrisW: do we need to fix examples before next WD?

<AdrianP> and we want to describe them in RIF and interchange them

Gary: would take action to update examples in next couple days

<markproctor> or this? http://www.w3.org/2005/rules/wiki/PRD#Running_example

Christian: OK

<ChrisWelty> ACTION: gary to update PRD examples for next wd [recorded in http://www.w3.org/2008/06/17-rif-minutes.html#action08]

<trackbot> Created ACTION-531 - Update PRD examples for next wd [on Gary Hallmark - due 2008-06-24].

<ChrisWelty> ACTION: christian to open issues on PRD as per editors notes [recorded in http://www.w3.org/2008/06/17-rif-minutes.html#action09]

<trackbot> Created ACTION-532 - Open issues on PRD as per editors notes [on Christian de Sainte Marie - due 2008-06-24].

Christian: multiple issues marked in PRD - need to open
... OK with <- if we include negation

ChrisW: <- impacts UCR?

<Harold> I now think the backward arrow is fine in UCR since this document needs to be general enough for all kinds of rules.

Adrian: suggests BLD use <- also

<Harold> (BLD, PRD, ...)

ChrisW: add as alias for :-

Harold: OK for UCR, keep :- for BLD

Michael: bad idea to add <- - confusing with frames - too many arrows

<AdrianP> exactly the examples in UCR should not be specific to BLD or PRD

ChrisW: frames only use ->

Michael: first order languages in FLD need <-

<Harold> I agree with Michael that in BLD '<-' is not great because it will encourage people to also use '->' for PRD rules (the "natural" direction).

Michael: doesn't subsume :-

<Harold> But '->' is needed for slots.

Michael: not the same as material implication

<AdrianP> We could use a different sign

<GaryHallmark> what about (conditions) => (actions)

Adrian: <- motivated by PRD

Michael: OK with different sign

<Harold> What about LaTeX-like \Longleftarrow (looks somewhat like '<==')?

<GaryHallmark> or what about (actions) "if" (conditions)

<markproctor> I second that. I find the common common presentation syntax very hard to read

<markproctor> especially with the forall

<AdrianP> yes, good idea we could use if ... then ...

<markproctor> heh, I'd rather we just use "lisp" for a common presentation.

<markproctor> pretty much all the PR companies would map easily to that and agree.

<Harold> Gary, yes, perhaps '==>' and '<==' with a long shafts?

ChrisW: schedule for next draft of PRD?

Christian: before end of this week

will use :- in PRD and UCR

<AdrianP> ok, will do. but also "if then" would work or <==

Gary: PR vendors happy to see any standards in this area

ChrisW: freeze PRD document by Monday

Adrian: freeze UCR on Monday

ChrisW: freeze BLD following Monday

Michael: xsd:int vs. xsd:long?

ChrisW: didn't decide - broke for lunch

Michael: add at risk note for Last Call

<ChrisWelty> RESOLVED: reuse NumericLiteral from SPARQL http://www.w3.org/TR/rdf-sparql-query/#rNumericLiteral giving us INTEGER, DECIMAL, and DOUBLE to the Presentation Syntax

ChrisW: resolution to reuse numeric literal from SPARQL
... integer, decimal, or double
... not yet reflected in BLD

Michael to add to BLD

ChrisW: Axel leaves on vacation soon

Equality

Harold: equality an issue for user-defined functions
... can't distinguish left and right

ChrisW: discussed at F2F11 - not enough support

Harold: changed mind

<Harold> http://lists.w3.org/Archives/Public/public-rif-wg/2008Jun/0048.html

<Harold> changing it: Jos, Sandro, (dave reynolds)

ChrisW: propose for next week

<Harold> Look at this: http://lists.w3.org/Archives/Public/public-rif-wg/2008Jun/0046.html

<AdrianP> it is quite useful for assignment to have a left and right side

<Harold> You cannot even say Equal(x y) = Equal(y x) without splitting side into left and right.

<Harold> This is the job of the semantics.

<GaryHallmark> left and right aren't really roles, though

AOB

no other business

ChrisW: adjourned 5 minutes early

<ChrisWelty> adjourned