See also: IRC log
<ChrisWelty> Meeting: RIF Telecon May 6, 2008
<csma> list agenda
<ChrisWelty> Scribe: mdean
csma: 2 comments from Dan Connolly on comment list
Sandro: discuss process in meeting - likely to get more comments soon
Chris: probably assign someone for response, rather than just on Wiki page
csma: IRI comment pretty close to
email list discussion
... second comment is question for Michael
Chris: add to end of agenda
15 people have answered, 12 are coming
closest hotel is full
bed and breakfast has space
Axel only on IRC today due to conflicting meeting
please fill out form
<Harold> About Sandro's ACTION: http://www.w3.org/2005/rules/wiki/FLD#XML_Serialization_Framework
<csma> all actions continued except 471, which is closed
<csma> ACTION-459: closed
<AdrianP> we might discuss it together with the UCR review
<csma> ACTION-437: closed
<csma> ACTION-253: closed
<AdrianP> nothing specific. there is currently a XBRL conference
no news from OWL
3 editor's notes in SWC
Axel: a proposal on table
Jos: takes care of 52 and 53
Jos to redial due to noise
Jos: notion of context comes from BLD document - this specifies specific contexts
Sandro: email thread from last
night - open issues
... prefer "language" over "context"
... or "format"
<sandro> or "formalism"
Jos: agree - took "context" from BLD - second argument to import statement
<Harold> Import ::= 'Import' '(' IRI CONTEXT? ')'
Jos: definition of import directives
<Harold> "The context specifies what kind of entity is being imported and under what semantics (for instance, the various RDF entailment regimes)."
<sandro> Sandro: So I'm proposing that this parameter, "language", be understood to be a default language identifier to use in case the language is not sufficiently self-identifying.
<sandro> Jos: I'm okay with that
Chris: consistency with current OWL WG?
Sandro: too early to say
<sandro> Chris: I am too -- how does it relate to what OWL-WG uses?
Michael: OWL uses profile
<sandro> mkifer: OWL uses "Profile"
<sandro> Sandro,Jos,Harold: Okay
consensus on "profile"
resolution not needed
Sandro: could call it default profile, to indicate that it doesn't override
Michael: other cases: semantics, data types, etc.
<Harold> Re Import(t c): "The constant t indicates the address of another rule set to be imported and c is called the context of import."
Chris: document takes precedence
if it specifies a profile
... but could imagine various precedence strategies
mkifer: should be an error
Sandro: ok with error
... now OK with profile given discussion
<sandro> ACTION: Harold to change "context" to "profile" in BLD and propose an XML syntax [recorded in http://www.w3.org/2008/05/06-rif-minutes.html#action01]
<trackbot-ng> Created ACTION-472 - Change \"context\" to \"profile\" in BLD and propose an XML syntax [on Harold Boley - due 2008-05-13].
<sandro> (and in FLD)
Sandro: next issue - URIs for
... probably makes sense to create new URIs linked to spec
<sandro> Sandro: let's make up new IRIs, like http://www.w3.org/2008/rif-import-profile/Foo
<sandro> Sandro: let's make up new IRIs, like http://www.w3.org/2008/rif-import-profile#Foo
<sandro> Chris: just for the ones mentioned in this document
<sandro> Sandro: right.
mkifer: is profile a
... why do we need to specify this?
Sandro: about 7 enumerated in Jos' document
Chris: will these be dereferenceable?
<sandro> Chris: this is just for Jos' document -- we don't need to talk about these in general.
Sandro: in principle yes
... great if someone is motivated to put statements there
<sandro> Chris: Just special IRIs that some implementations know. When you see this IRI, use this form of interpretation.
<sandro> Sandro: Yes.
<sandro> ACTION: jdebruij2 to pick IRIs for profiles [recorded in http://www.w3.org/2008/05/06-rif-minutes.html#action03]
<trackbot-ng> Created ACTION-473 - Pick IRIs for profiles [on Jos de Bruijn - due 2008-05-13].
Jos: what about RDF graphs with different profiles - pick highest?
Sandro: what about transitive imports?
Jos: pick highest, but haven't thought about transitivity - not feasible to use multiple profiles
Sandro: could make it an error or
... would like open issue on importing things with different profiles
... OK with last call draft as is
<sandro> ACTION: jdebruij2 to open a non-CP issue on importing with mixed profiles [recorded in http://www.w3.org/2008/05/06-rif-minutes.html#action04]
<trackbot-ng> Created ACTION-474 - Open a non-CP issue on importing with mixed profiles [on Jos de Bruijn - due 2008-05-13].
Jos: generic profiles: RDF vs
... could be combined - no formal distinction
... don't require specification of a profile
... imports currently defaults to RIF ruleset
Sandro: can distinguish based on
RIF MIME type
... also need to work on that
<sandro> ACTION: Sandro to look into mime type registration for RIF [recorded in http://www.w3.org/2008/05/06-rif-minutes.html#action05]
<trackbot-ng> Created ACTION-475 - Look into mime type registration for RIF [on Sandro Hawke - due 2008-05-13].
<sandro> Sandro: How about make Profile optional, and if omitted use mime-type.
Jos: this is a BLD issue
<sandro> Chris: So -- no generics, profile is optional, and up to spec how to handle this if document isn't self-identified.
<sandro> Jos: Michael, harold?
mdean: file: IRI doesn't have a MIME type
Sandro: can live with second
argument of "RDF" but this seems silly
... file: IRI library may have file extension mappings to MIME types
<sandro> Sandro: I can live with second-parameter-required for non-RIF imports.
<sandro> Sandro: I think that works for XML, OO style and not.
resolution not required since things stay the same
wait until next week to close Issue 52
<sandro> postponing closing issue-52 until we can check over the documents with these edits.
josb: issue 53 is resolved
Sandro: would like review by Bijan, who's thinking about annotations in OWL
Chris: seems to account for where OWL is going as well as providing backward compatibility
<sandro> PROPOSED: close ISSUE-53 as in current rdf-owl text
<csma> RESOLVED: close ISSUE-53 as in current rdf-owl text
<sandro> PROPOSED: close ISSUE-53 as in current http://www.w3.org/2005/rules/wiki/SWC
<sandro> RESOLVED: close ISSUE-53 as in current http://www.w3.org/2005/rules/wiki/SWC
Jos: contact OWL DL people as
soon as possible
... syntactic restriction on use of variables in RIF rules
<Harold> I guess we are talking about a subdialect of BLD, which fulfills Safeness Restrictions?
<sandro> ACTION: jdebruij2 to propose solution to ISSUE-54 that he's happy with and OWL-RIF TF is happy with [recorded in http://www.w3.org/2008/05/06-rif-minutes.html#action06]
<trackbot-ng> Created ACTION-476 - Propose solution to ISSUE-54 that he's happy with and OWL-RIF TF is happy with [on Jos de Bruijn - due 2008-05-13].
Sandro: leaning toward option 4
Chris: first option corresponds to "out-of-band agreement"
<csma> What I have in mind when I say "out-of-band agreement, I mean that the requirement must be mentioned in the RIF doc
Chris: shouldn't require ignoring of datatypes not in spec
<sandro> kifer: how about something between 1 and 2 ---- two systems may agree on some extra datatypes, if they see an unrecognized datatype, then they issue an error.
mkifer: combination of 1 and 2?
Sandro: might work
<sandro> jos: I prefer option 2 -- you reject if you cannot deal with the datatypes that are there.
<sandro> kifer: I think that's the same as I'm saying. You assume systems can support official types plus some extra types, and you reject only if it's in neither.
mkifer: agree - reject datatypes that you don't support
csma: option 1 is not quite out-of-band agreement
<sandro> csma: allow a doc to make ref to non-std datatypes; an impl should be able to use the dt's they know.
<sandro> csma: they should be required to throw an error/warning if they don't recognize the datatype, and use it if they do. only issue is how to make sure no ambiguity. maybe RIF docs have to list their non-std datatypes.
<sandro> csma: all the datatypes in RIF document must be either std or unambiguously named; and any implement that knows the DT can use it, and if you don't you must warn/error.
<sandro> jos: yes
<sandro> mkifer: yes
<josb> OWL says: "If an input document uses datatypes that are not supported by the datatype map of an OWL consistency checker then it MAY report a warning. "
Harold: option 4 subsumes 1 and 2
<sandro> Harold: I would hope we could go for option 4, it subsumes 1 and 2 -- the xform could be identify, or refine error, etc. It's nice an general. It's at the heart of what RIF is about.
Harold: heart of RIF as interoperability mechanism
Chris: how much work needs to be done to include this?
<sandro> Harold: it could be left open, in the sense of our externals.
<Zakim> sandro, you wanted to talk about use case where a big publisher wants to include a new data type
Sandro: compelling use case for 4
is big web publishers - can't introduce new datatype until
consumers all implement - chicken and egg precludes
... but don't see a solution by the end of May
... prefer to leave the door open for programmable fallback
mkifer: which document(s) will specify this?
<Harold> Sandro, is XTAN at all related to GRDDL?
<sandro> Sandro: I wish XTAN were done ahead of time, but it's not, so the question is how to leave the door open for them.
Sandro: XTAN in separate
REC-track document used by RIF but separate
... need to finish enough for people to build against BLD - shouldn't require major changes to implementations
<Harold> I think XTAN can also be used for transformations that return a normal form.
<josb> well, people do use xsd:date in OWL, I think
<Harold> (XTAN could thus allow users to work with non-normal-form information.)
mkifer: need to say something about compliance, beyond data types
<josb> uses xsd:float: http://protege.cim3.net/file/pub/ontologies/camera/camera.owl
mkifer: strawman compliance statements that can later be changed
<sandro> Chris: let's leave compliance clause until after we have implementaiton experience
<Harold> What about at least an Editor's note about compliance?
Chris: no way to know what compliance means for BLD
<sandro> kifer: let's provide some straw man, at least.
<csma> you mean, later than LC?
Sandro: can procedurally wait until CR
mkifer: discussion a year ago
<Harold> What will happen regarding answering public comments, gathering errata, etc. between RIF phase 1 and phase 2?
mkifer: should take up seriously or not, not just for datatypes
<csma> +1 to michael
Chris: much clearer idea of what BLD is than we had a year ago, but experience still limited
<sandro> ACTION: kifer to draft some text on compliance [recorded in http://www.w3.org/2008/05/06-rif-minutes.html#action07]
<trackbot-ng> Created ACTION-477 - Draft some text on compliance [on Michael Kifer - due 2008-05-13].
<Harold> What will happen with the OWL-RIF Task Force after May 31?
<csma> rssagent, make log public