See also: IRC log
<ChrisW> Scribe: Francois Bry
Christian: Review actions. No
... Approve minutes of last week telecon. Objections?
... No objections. Minutes approved.
<ChrisW> RESOLUTION: Accept minutes of 3/13 telecon
Christian: of March 13 telecon. Minutes of F2F 5 approved? Objections? None.
<ChrisW> RESOLUTION: Approved F2F5 minutes
Christian: Minutes of F2F 5
... Amendments to the agenda? None.
... First topic F2F 6. Looking at the answers to the questionnaire. Only one way where all can attend, that is June 2nd-3rd (ie week end) in Innsbruck.
... comments? None.
<ChrisW> ACTION: jdebruij to confirm availability of F2F meeting facilities in Innsbruck [recorded in http://www.w3.org/2007/03/20-rif-minutes.html#action03]
<rifbot> Created ACTION-267 - Confirm availability of F2F meeting facilities in Innsbruck [on Jos de Bruijn - due 2007-03-27].
Christian: Who offers to host F2F
7 late Aug to early Oct? Preferences Sept in North
... No deadline for offers yet but F2F 7 should be finalize before F2F 6 (June 2-3).
Chris: Plenary will be fist week of November. Does not fit well with RIF F2F.
Christian: Next topic
... Reports on liaisons? None.
... Related to liaison. The MISMO WG have an example of proof-of-concept interchange of rules.
... Does someone want to translate their rules in RIF?
Harold: I might be interested in doing it.
Christian: I'll ask for getting this example and putting it on the RIF Wiki.
<ChrisW> ACTION: Christian to check with MISMO that their proof of concept example can be published to RIF [recorded in http://www.w3.org/2007/03/20-rif-minutes.html#action04]
<rifbot> Created ACTION-268 - Check with MISMO that their proof of concept example can be published to RIF [on Christian de Sainte Marie - due 2007-03-27].
Christian: MISMO: Mortgage Industry Standard ...
Chris: A new incubator activity ask RIF to look at what they are doing on incertainty reasoning.
Christian: JeffPan and Giorgos should be interested.
<agiurca> both of them are there
Christian: More on liaison? No.
Let us move to next topic technical design..
... actions have been worked out or are continued.
... The frozen version for RIF core has been published on Friday.
... We decided we cannot stop it any longer. Can be reviewed up till the end of the week. We'll vote on it next Tuesday.
... Two reactions, one from Dave. Comments?
Dave: Small editorial things. Two
issues: syntax of sorts and examples all in the non-sorted
... I suggest to explain that in a caveat.
Christian: This has already been decided.
Michael: there was a proposal to
start with sorts but leaving examples without sorts for not
... Regarding syntax of sorts. This has not been discussed at the F2F.
<Harold> Dave's caveat wording for his i1) is fine with me.
Christian: Dave's is proposal to
add a caveat about the unsorted examples.
... Objections to doing this?
Michael: Maybe should we say something in the sorted part that future version will give examples with sorts.
Dave: My proposal is to add something before Example 3.
Christian: Or after introduction of multisorted logic.
Chris: Close to an example is preferable.
<csma> The examples of BNF and XML rule syntax given here use the unsorted version of the condition syntax and fail to illustrate the use of URIs for constants. This will be addressed in a future working draft.
Christian: Objections to adding Dave's caveat before Example 3?
<ChrisW> RESOLVED: Add " "The examples of BNF and XML rule syntax given here use the unsorted
<ChrisW> version of the condition syntax and fail to illustrate the use of URIs
<ChrisW> for constants. This will be addressed in a future working draft."
<ChrisW> to the CORE WD
Christian: No objections? None. Issue resolved, caveat will be added.
Chris: Harold and Michael, do you take the action?
Harold: I take it.
Michael: We could easily add sorts in Example 3.
Christian: This could convey a wrong message on how sorts are to be.
Michael: But we already decided on this.
<ChrisW> ACTION: Harold to add DaveR's caveat text to CORE [recorded in http://www.w3.org/2007/03/20-rif-minutes.html#action05]
<rifbot> Created ACTION-269 - Add DaveR\'s caveat text to CORE [on Harold Boley - due 2007-03-27].
Chris: Let us keep the example as
it is for now.
... let us have this caveat added.
Christian: Regarding Dave's second comment (syntax for primitive sort). What do we do?
<agiurca> the usual way is to keep concepts which are already known
Francois: Dave's syntax proposal is standard and rather nice. Why not take it?
Dave: It seems to be a small syntactic thing. Benefice would be to have it look like in other SW languages.
Michael: You mean N3 syntax?
Hassan: I agree with Francois.
This might conflict with namespaces.
... The notation in the WD is contrive and ugly.
Dave: in N3 there are quotes. Integer unquoted is like in our draft, a shortcut.
Michael: If there is consensus to change this syntax, I'll change it.
Hassan: I'll prefer the simpler and most naturakl and most known. value column sort or something like that.
<sandro> Very considerate of them, Harold. :-)
<josb> because it is standard!
<agiurca> Why we don't use the textual syntax as for example "lexicalValue"^^"sort"
<agiurca> +1 josb
Chris: Only two choices: keeping how it is now or choose the notation of another formalism.
<sandro> The single-hat was used for something else in N3.
Christian: Second choice is "lexicalValue"^^"sort" for WD 1.
<Hassan> No but ...
<sandro> (The single-hat was used for the same thing "." and "->" are commonly used for in other languages.)
Hassan: There are no reasons for _. It does not make it easier to parse.
<Harold> I don't know why "sort" be in quotes?
Christian: Decision is between
_sort"value" or "lexicalValue"^^"sort".
... No objections against "lexicalValue"^^"sort".
Hassan: One character should be enough ^ not ^^.
<Hassan> Great! Ubu roi...
Chris: Like in RDF or like in the current draft. No other choice.
Christian: It is _sort"value" or "lexicalValue"^^"sort". Objections against "lexicalValue"^^"sort".
<josb> +1 for ^^
<mdean> +1 for ^^
Christian: No objections. "lexicalValue"^^"sort" chosen. Who implements the modification in WD Core?
Michael: How is the syntax exactly? "lexicalValue"^^"sort" or "lexicalValue"^^sort ?
<Harold> I take the action.
Michael: What should be the syntax for variables?
<Harold> Variables should be "lexicalValue"^^?Varname I think.
<agiurca> In XML syntax variables may be something like xs:NCName
Christian: Action on Harold for changing syntax of sort annotations.
<mdean> N3 does not have sorted variables
<Harold> I meant the latter.
Dave: I propose the N3 syntax for
values not for variables.
... My sugestion was to align on something already published.
Christian: Harold's proposal ?varname^%^ sort.
<Harold> So we have ?Varname^^sort for sorted variables and "lexicalValue"^^sort for sorted literals.
Michael: this is reasonable to me.
Chris: replace _sortWHATEVER by WHATEVER^^sort.
<ChrisW> ACTION: Harold to change _sort syntax to ^^sort syntax [recorded in http://www.w3.org/2007/03/20-rif-minutes.html#action06]
<rifbot> Created ACTION-270 - Change _sort syntax to ^^sort syntax [on Harold Boley - due 2007-03-27].
Christian: action on Harold implementing this change.
Christian: An orther sort of comments by Francois.
<sandro> scribenick: sandro
Francois: It was systematic
issues in the text, most of which have been resolved.
... Not happy with "I", the interpretation function.
... the subscript, "e" constant, "e" function, etc.
... these are stylistic details.
... in the Semantics section
MichaelKifer: There is a function, I<sub>truth</sub> .... Francois wants to call it I<sub>formula</sub>. This is truth evaluation, in every text book
Francois: I don't really agree,
but it's okay.
... It makes the text harder to understand, but it's not impossible to understand.
CSMA: Stylistic questions like this, we'll have to deal with before the final draft, but let's keep it like this for now.
<scribe> scribenick: Francois
Christian: We'll vote on next Tuesday. Mention objections or issues before. I expect the document to be accepted.
Sandro: will the WD1 be published soon after the vote?
Christian: If we vote to publish the text, there should only be minor changes.
Sandro: Formating and the like will be needed. A few days work.
Christian: Then we'll have 2
months for producing a much better WD2.
... Let us more to the next topic is issue 30, the RIF URI.
Chris: There has been discussions
on this at the last F2F. rif:uri has not been defined. Hence
the issue recorded.
... We were pretty close to an understanding at F2F, I think.
Christian: Can someone summarize
the positions on rif:uri? Jos?
... josb are you there?
josb: The discussion was to
clarify what is meant by URIs. It seems that a URI in RIF is
the same as the XML schema data type anyUri. This is not the
case. We do not want URI be interpretaed as themselves but
instead as individual in some particular domain.
... we have to characterize the syntactical class of uri.
Michael: I changed the document.
URI can now denote resources on the web. It is not clear
whether we need anyUri at all.
... the mention is in section multisorted syntax at the end.
Christian: josb what decision do you suggest?
josb: I do n ot understand why we need a sort for URIS.
Michael: Do you mean it should be called "resource" instead of "uri"?
josb: we should have constants interpreted in the usual abstract domain. A sort "resource" might be closer to what we actually want.
<ChrisW> Harold - I found three instances of "sortal" in the frozen draft
Michael: syntactic control is need.
<ChrisW> "sortal" should be changed to "sort"
<Harold> In http://www.w3.org/2005/rules/wg/wiki/Core/Positive_Conditions, "Syntax for Primitive Sorts" it says: rif:uri. Constants of this sort have the form _rif:uri"XYZ" where XYZ is a URI as specified in RFC 3986.
Francois: There are two ways. Uris interpreted as "resources" or interpreted freely by any individual in the domain.
Hassan: If you assign a sort to Uris, since uris are used for defining sorts, ehat is the sort of a sort?
Hassan: Uris should denote only
resources. They are syntactic object. Uri have a particular
syntactic form. Then have to be given a denotation.
... If it is a syntactic object, it can only have a meta-sort.
Michael: Sorts are not constant in the domain.
Chris: What would belong to the sort Uri?
<Harold> Jos, at F2F5 we discussed the issue that several syntactic URIs can denote the same individual in the domain. RFC 3986 introduces a normalization algorithm, but I have not yet seen a formal equality theory for URIs (URI forwarding is hard to deal with).
Christian: Can you give an example where we need a constant of sort URI?
Michael: predicate name refer to an object on the web and we use a uri to retrieve it.
<josb> These are signatures, not sorts!
Michael: URIs should be used for identifying names of predicates.
Christian: Part of the probelm seems to be we do not know what uri are to be used for.
josb: interesting thing is that
URI have a semantical domain associted with this sort.
... uri for identifying names of predicate has nothing to do with sorts.
Michael: two issues: syntax of uri and domain for their interpretation.
josb: What is the point in having this sort?
Michael: for using uris as names of predicated.
josb: no need of a URI sort for this.
Michael: we need to assign
signature to them. this does not seem feasible.
... we need to group them.
... we need a mechanism to assign signature to constants.
josb: why is sort needed for that.
Michael: what other mechanism is suitable?
josb: think of any other kind of mechanism.
Christian: can you propose such a mechanism?
Christian: action on josb on that.
<ChrisW> ACTION: jdebrui to propose a mechanism that doesn't have a problem with signatures for uris [recorded in http://www.w3.org/2007/03/20-rif-minutes.html#action07]
<rifbot> Sorry, couldn't find user - jdebrui
Hassan: I am confused. Take variables instead of URI. If a syntactic category of variables are capitalized.
<ChrisW> ACTION: jdebruij to propose a mechanism that doesn't have a problem with signatures for uris [recorded in http://www.w3.org/2007/03/20-rif-minutes.html#action08]
<rifbot> Created ACTION-271 - Propose a mechanism that doesn\'t have a problem with signatures for uris [on Jos de Bruijn - due 2007-03-27].
Hassan: Ambiguity is built
... Using URIS for syntactical categories and fdor constants is ambiguous.
Michael: this is not done.
Hassan: p(f, x) if all of p f x can be uri, what is the variable?
<ChrisW> ACTION: hassan to propose example of ambiguity [recorded in http://www.w3.org/2007/03/20-rif-minutes.html#action09]
<rifbot> Created ACTION-272 - Propose example of ambiguity [on Hassan Ait-Kaci - due 2007-03-27].
<ChrisW> DaveR: If the question is about using URIs for e.g. predicate names as well as other kinds of constants, then why?
Michael: THis is what the current linear syntax has. I need to think of thne abstract syntax.
Christian: We are not able to
resolve the issue right now. I would like the paragraph about
RIF mentioning that there is an open issue on the topic.
... I do not think it is clear what it means using uris as references to resources.
... the reference to "rdfs resources" should be more specific.
<ChrisW> ACTION: michael to add reference to URI issue [recorded in http://www.w3.org/2007/03/20-rif-minutes.html#action10]
<rifbot> Sorry, amibiguous username (more than one match) - michael
<rifbot> Try using a different identifier, such as family name or username (eg. merdmann, mkifer, msintek, uscholdm)
<ChrisW> ACTION: mkifer to add reference to URI issue [recorded in http://www.w3.org/2007/03/20-rif-minutes.html#action11]
<rifbot> Created ACTION-273 - Add reference to URI issue [on Michael Kifer - due 2007-03-27].
Michael: I can mention the issue in the doc. I am not sure what can be done for the second request.
Christian: "in a way similar to a wohle doc." is not clear.
Michael: we should only say we'll resolve the issue later.
<Harold> I will do the sortal->sort change (without action item).
Chris: in 18.104.22.168 typos. I can fix it in the wiki if this is ok.
Christian: Second issue not being formally raised by Deborah yet: distinct names for predicate, functions and constants.
Harold: it is a typical
misundserstanding between "constant" and individual
... the can be "function constant", "predicate constants" as well.
Michael: the issue was whether the sets of individual constants, function symbols and predicate symbols should be pairwise disjoint.
Christian: the wording in the
issue uses the word "constant" in a maybe misleading
... If the above mentioned sets are pairwise disjoint in RIF CORE, how to to remove it in a RIF dialect.
Michael: no problem, it is a matter of assigning more signature to constants.
Christian: AOB? Any other URGENT business? None. Adjourned.