See also: IRC log
<ChrisW> Scribe: Hassan
<ChrisW> PROPOSED: accept minutes of Aug 12 telecon
<ChrisW> RESOLVED: accept minutes of Aug 12 telecon
Sandro: OWL WG meeting a few weeks ago: go to last call in Oct 08. Draft due next month.
<sandro> see http://lists.w3.org/Archives/Public/public-rdf-text/
Axel: reporting on rdf:text
<ChrisW> ACTION: chris to figure out how to link to rdf:text comments [recorded in http://www.w3.org/2008/08/19-rif-minutes.html#action01]
<trackbot> Created ACTION-559 - Figure out how to link to rdf:text comments [on Christopher Welty - due 2008-08-26].
<AxelPolleres> sandro will add a link to http://www.w3.org/2007/OWL/wiki/InternationalizedStringSpec with the mailinglist address to solicit feedback?
Axel: still open issues remaining; on his (Axel's) side nothing new to tell
<AxelPolleres> I completed ACTION-552, see http://lists.w3.org/Archives/Public/public-rif-wg/2008Aug/0074.html
<AdrianP> currently the semantics does not account for pattern formula associated with an enclosing Forall
No suggestion for hotels
ChrisW: try hotels in Brooklyn or even Queens using public trans. to get lower rates
<ChrisW> ACTION: chris to put DTB review on agenda for next week [recorded in http://www.w3.org/2008/08/19-rif-minutes.html#action02]
<trackbot> Created ACTION-560 - Put DTB review on agenda for next week [on Christopher Welty - due 2008-08-26].
ChrisW: What should CORE be?
BLD intersected PRD = CORE?
DaveReynolds and GaryHallmark agree on this def
ChrisW: Assume this is so (CORE *is* the intersection) - how do we define this language?
GaryHallmark: decidability or tractability issues for CORE are not relevant for RIF
ChrisW: What does such an intersection look like?
GaryHallmark: we could extract the syntax from the common grammar rules
<AdrianP> core production rule syntax, i.e. without negation; only assert in the head and without the special pattern formula in the forall
GaryHallmark: intended semantics overlap is a fuzzy concept; make things that are relevant or not more explicit
AdrianP: Suggests a specific language... will work it out (for the condition language)
<ChrisW> ACTION: apaschke to document the BLD/PRD syntax intersection [recorded in http://www.w3.org/2008/08/19-rif-minutes.html#action05]
<trackbot> Created ACTION-561 - Document the BLD/PRD syntax intersection [on Adrian Paschke - due 2008-08-26].
ChrisW: Handling of external functions; skolem functions as well; subclassing also
DaveReynolds: prefers that CORE stay minimal and not have membership and subclass
ChrisW: Who would disagree as a maximal intersection of PRB and BLD?
Jos: how useful a language would that be?
ChrisW: a maximal intersection of PRB and BLD would make interchange easier
Jos: yes, but there may be other uses for CORE
ChrisW: so it should be easy to "implement"
<AdrianP> other uses e.g. the integration of CORE + ontologies. Decidability is often a requirement for many application scenarios
<GaryHallmark> PRD has no equality in the head, so neither does Core
ChrisW: Jos would an Easy-To-Implement core bigger or less?
Jos: it would be less
GaryHallmark: it would be good to itemize "things" that make this hard ... disjunction?
<AdrianP> disjunction in the body can be split into two rules, so there is no need to have it in Core
DaveReynolds describes features that he'd wish for CORE (a la Jena)
ChrisW: another aspect is to relate it to existing languages
<ChrisW> ACTION: Dave to open CORE issues on tracker [recorded in http://www.w3.org/2008/08/19-rif-minutes.html#action06]
<trackbot> Created ACTION-562 - Open CORE issues on tracker [on Dave Reynolds - due 2008-08-26].
ChrisW: More on Core?
No more on core
<AdrianP> the discussion points: http://lists.w3.org/Archives/Public/public-rif-wg/2008Aug/0056.html
ChrisW: shooting for one-month after the F2F to have the final draft ready
<StellaMitchell> about how many test cases by then?
ChrisW: review of the approach?
Discussing format of the document
StellaMitchell: Test Case Number of Results as an example
AdrianP: explains what a query is in the context on the test case
Jos: Why do you need such a concept as a "query"?
<AxelPolleres> I made some comments on test cases and "query" answers in my mail.
Jos: proposes a simpler scheme instead based on logical entailment
ChrisW: we have not defined what a query is
<AxelPolleres> +1 to jos because enumerating "answers" is not possible, especially if there are infinite answers.
Jos: exactly: so better not use it
AdrianP: we need it because entailment is not enough in certain test cases
Jos: argues against the need
<AxelPolleres> ... also, we'd need a defined output format, etc.
ChrisW: I think Adrian is referring more to a unit test
ChrisW: does not like the arrow of the frame notation
(Test case Positive Entailment Test)
ChrisW: understands the notation Premiss -> Conclusion
<AxelPolleres> no arbitrary condition...
ChrisW: hesitant to define a whole new language for just test cases
<AxelPolleres> ... I was suggesting such more general built-ins some time ago.
<DaveReynolds> positive and negative entailment tests seem like the main ones to me
<GaryHallmark> can a BLD expert comment on the "answer" to http://www.w3.org/2005/rules/wiki/PositiveEntailment_Entail_everything
Axel: agrees - need a new language for the queries, the output, etc ...
ChrisW: try to focus on entailment
StellaMitchell: question on the exact nature of the test cases
<josb> Gary: Indeed, A(t), for any ground term t, is entailed
ChrisW: test cases should use things that are expressible in BLD
<josb> If you say everything is in A, you can derive that everything is in A
ChrisW: let's keep it simple
<AxelPolleres> I would have more comments when it comes to test cases for core... but I don't know whether this is in scope of the discussion now and whether I haven't said all in the mails yet on this topic.
ChrisW: Jos and Michael - please think up some test cases for BLD
Jos: also w/ RDF and OWL - will do
ChrisW: anyone else please as well
<AdrianP> a test case for each DTB built-in
StellaMitchell: do we need things like shown in "Built-Ins numeric add" test case?
ChrisW: this looks more like what I was imagining ...
StellaMitchell: do we need to verify that all the builtins are implemented?
<ChrisW> ACTION: chris to add "test case for every builtin?" to issues list [recorded in http://www.w3.org/2008/08/19-rif-minutes.html#action07]
<trackbot> Created ACTION-563 - Add \"test case for every builtin?\" to issues list [on Christopher Welty - due 2008-08-26].
<AdrianP> usually test cases a simple, at least from the point of view of test-driven development in agile programming
<ChrisW> areck jos
Jos: wonders if test cases for consistency checking are needed?
<AxelPolleres> jos, could you specify an example for where builtins cause inconsistency?!?
<AdrianP> beside entailment tests we currently have syntax tests
ChrisW and Axel do not believe so
StellaMitchell: it was question to have some consistency checks where they are needed
<AdrianP> I listed some categories of potential test cases here http://www.w3.org/2005/rules/wiki/Test#Categories_of_RIF_Test_Cases
<josb> Axel: currently not sure whether there is anything besides equality that can cause inconsistency, but I suspect there is
Sandro: agrees that as much consistency check should be done (necessary conditions, not sufficient)
<AdrianP> inconsistency e.g. in PRD due to negation
Sandro: discusses the format of the tests (for nested tables) and how to simplify it
AdrianP: asks a question about
editing the test cases
... using a standard syntax or XML
StellaMitchell: yes we need an official format
<GaryHallmark> stella: nice to be able to enter tests in presentation syntax and then auto convert to XML
<StellaMitchell> did Hassan write a program to convert PS to XML?
Sandro: discusses how to organize
the wiki to automate all this
... using RDF data and XHTML tables
<ChrisW> ACTION: hassan to finish PS to XML conversion based on LC BLD grammar [recorded in http://www.w3.org/2008/08/19-rif-minutes.html#action08]
<trackbot> Created ACTION-564 - Finish PS to XML conversion based on LC BLD grammar [on Hassan Ait-Kaci - due 2008-08-26].
Sandro: discussing using Pres. Syntax and then generate XML form
ChrisW: how do we ensure that all
versions of the same syntax are kept consistent with one
... the idea is to keep all the wiki versions of the same thing be kept consistent (in PS or XML)
Sandro: there are ways to do that ...
ChrisW: test metadata and the submission process?
AdrianP: describes the current submission process ...
ChrisW: what about an editorial process?
ChrisW: how - email? wiki?
<LeoraMorgenstern> I think email submission is fine.
Sandro: not sure of ways to do that ...
<LeoraMorgenstern> If it's emailed to the WG, we will have a record of it, but still not allow access to the wiki.
Sandro: ... from the wiki
<LeoraMorgenstern> In any case, we have to vet and modify the examples, so there will be work involved in any case with exernal submissions.
Sandro: let us use email for now
<LeoraMorgenstern> I agree Sandro; I don't think we'll get many email submissions
ChrisW: we'll review this again next week
<ChrisW> ACTION: sandro to get CVS access for TCG [recorded in http://www.w3.org/2008/08/19-rif-minutes.html#action09]
<trackbot> Created ACTION-565 - Get CVS access for TCG [on Sandro Hawke - due 2008-08-26].
ChrisW: giving a pat on the back of Adrian et al...
ChrisW: external review from Peter Patel-Schneider
<ChrisW> who is on the phone?
Jos: perhaps Michael should respond
<ChrisW> ACTION: Chris to ask MK to look at PFPS4 [recorded in http://www.w3.org/2008/08/19-rif-minutes.html#action10]
<trackbot> Created ACTION-566 - Ask MK to look at PFPS4 [on Christopher Welty - due 2008-08-26].
<sandro> Is this page public? http://www.w3.org/Systems/Accounts/w3t/
ChrisW: saying something wise ... ;-)
<ChrisW> ACTION: chris to ask axel to look at 2cnd to last comment on PFPS4 [recorded in http://www.w3.org/2008/08/19-rif-minutes.html#action11]
<trackbot> Created ACTION-567 - Ask axel to look at 2cnd to last comment on PFPS4 [on Christopher Welty - due 2008-08-26].
<AdrianP> Sandro, no it is not public
+1 to adjourn