See also: IRC log
<IgorMozetic> is zakim working properly? (74394# not valid)
ACTION-374: done (extension requested)
Sandro: committee meets tomorrow, might have answer then
<sandro> Odd, IgorMozetic.... Sometimes the tones get too distorted by the phone system to be recognized......
Harold: haven't seen notification
of WD publication on RDF list
... should we have larger scale notification of the new WDs?
Sandro: only concern is limited attention, people will only look at docs so many times. If we are going to have a refactored draft in the new couple of months would be reasonable to wait to till then publicize more aggressively.
Jos: would be interested in getting as much f/b as possible on the compatibility doc
cmsa: suggested tabling this question until discuss the BLD plan
No objections. Minutes accepted.
Paul: all input appreciated on PRR review, finalization due around Feb
csma: on first page of doc is email address for comments
csma: notes that Bijan is liaison
from OWL-WG to RIF, do we want a reverse liaison?
... currently Sandro is our liaison
<josb> yes; I can help
josb volunteers to help Sandro with tracking the OWL-WG liaison issues
Sandro notes that the chair hasn't filled out the survey!
Igor: ideally would like to decide F2F9 dates by Dec 4th
Sandro: encourages everyone,
especially chairs and editors, to fill survey in next day. Then
can set up location survey for next week or so.
... need next survey to be up for a full week, that would mean that at next telecon will have to both decide on dates and confirm hosts can support dates
csma: deadline for date survey end of this week
Sandro: each host work out which dates for them and propose a couple of dates over the weekend ready work for next telecon (27th)
<scribe> ACTION: Igor to post proposed F2F9 dates by next telecon, i.e. sometime Monday 26th [recorded in http://www.w3.org/2007/11/20-rif-minutes.html#action02]
<rifbot> Created ACTION-379 - Post proposed F2F9 dates by next telecon, i.e. sometime Monday 26th [on Igor Mozetic - due 2007-11-27].
<scribe> ACTION: christian to post proposed F2F9 dates by next telecon, i.e. sometime Monday 26th [recorded in http://www.w3.org/2007/11/20-rif-minutes.html#action04]
<rifbot> Created ACTION-380 - Post proposed F2F9 dates by next telecon, i.e. sometime Monday 26th [on Christian de Sainte Marie - due 2007-11-27].
<scribe> ACTION: Axel to post proposed F2F9 dates by next telecon, i.e. sometime Monday 26th [recorded in http://www.w3.org/2007/11/20-rif-minutes.html#action05]
<rifbot> Created ACTION-381 - Post proposed F2F9 dates by next telecon, i.e. sometime Monday 26th [on Axel Polleres - due 2007-11-27].
ACTION-365: pending discussion, drafted reply
ACTION-364, ACTION-362: pending discussion
<csma> PROPOSED: Close issue 2, based on Jos's analysis in 
PROPOSED: close ISSUE-2 based on Jos' analysis
RESOLUTION: close ISSUE-2 based on Jos' analysis
Jos: originally wanted rif:text
to be able to treat all constants uniformly in the syntax
... if we violate this for one type of constant then we might want to re-evaluate how we represent other constants (e.g. IRIs, local constants)
... drawback of sticky with rif:text is that we are creating a new datatype which is not really our business
csma: why not only allow addition of language tags and stop there?
Jos: that then creates two kinds of constants, a strange lack of uniformity
Harold: can't we have both a symbol space and a language tag?
csma: but the lang tag might apply to other types?
Jos: no! (scribe agrees)
Sandro: by definition XML literals don't allow language tagging
Harold: e.g. lang tag even for IRIs
<sandro> (chat.com could be english or french, right?)
DaveR: no, lang just applies to text for user presentation not to other data types
<IgorMozetic> I vote for uniformity
csma: seems it needs more discussion, so postpone to later telecon
<Hassan> +1 with MK
MK: benefit of uniformity is just one mechanism for defining constants, as opposed to mess of special cases
<IgorMozetic> +1 for MK
csma: XML 1.1 is a little more permissive but is still WD
Sandro: clarifies this about XML Schema Datatypes not XML!
<sandro> XSD1.1, not XML1.1
csma: so the proposal is to refer to XSD 1.0 and change that if XSD 1.1 reaches Rec before BLD
Jos: no real drawback to
referring to XSD 1.0 but should explicitly state we will switch
to XSD 1.1 when Rec, to satisfy people who want to use XML
... that remark should be explicit in doc
Sandro: makes sense
<josb> PROPOSED: we refer to XSD 1.0 while it is not a Rec, but add an explicit remark that we will switch to 1.1 if it reaches Rec before RIF
<sandro> PROPOSED: We'll refer to XSD 1.0 instead of XSD 1.1 in our document for now, including a clear note that it our intention to change to XSD 1.1 when it becomes available, so that people can use XML 1.1.
<sandro> RESOLVED: We'll refer to XSD 1.0 instead of XSD 1.1 in our document for now, including a clear note that it our intention to change to XSD 1.1 when it becomes available, so that people can use XML 1.1.
csma: is equality issue critical path or not?
<sandro> +1 equality is CriticalPath
csma: so ISSUE-47 is critical
... so 13 critical path issues
... Sandro needs to add issue on invisible extensions, is that CP?
Sandro: there is extensibility work that is critical path though that issue doesn't capture all of that
csma: if people think there are other issues that are CP to last call should raise them as soon as possible
<sandro> ACTION: sandro to report any critical-path issues related to extensibility [recorded in http://www.w3.org/2007/11/20-rif-minutes.html#action06]
<rifbot> Created ACTION-382 - Report any critical-path issues related to extensibility [on Sandro Hawke - due 2007-11-27].
csma: roundtripping - CP or
... aren't ISSUE-46, ISSUE-39 and ISSUE-33 interdependent yet 46 is not marked CP
Jos: modules is just one way to do this
csma: ISSUE-34, should that be closed?
Jos: this was originally raised to record Sandro's concern on whether this conflicts with extensility
Sandro: right, can't close that yet
csma: said at F2F8 that 2nd BLD WD will resolve issues 45 (lists), 41(#), 43(##), 44(named args) and 40(builtins)
<Harold> If a set of builtins uses unrelated URIs to point to each builtin's specification, there is nothing ressembling a module. If, however, a set of builtins uses a shared base URI and different local URIs to point to each builtin's specification, the base URI can be regarded as a module.
<josb> (by the way, regarding an earlier discussion in the telecon, Chris advertised our working drafts during his keynote at ISWC)
csma: can we do work on 40 and get quick decisions on the others in 1.5 months
MK: someone has to work out the actual document for builtins
csma: want to keep OWL compat doc in sync with BLD, can we have 2nd WD on 1.5m timescale?
Jos: yes, but would like to discuss the current page in an up coming telecon
<scribe> ACTION: christian to schedule discussion on OWL compatibility document at future telecon [recorded in http://www.w3.org/2007/11/20-rif-minutes.html#action07]
<rifbot> Created ACTION-383 - Schedule discussion on OWL compatibility document at future telecon [on Christian de Sainte Marie - due 2007-11-27].
csma: are there issues on the BLD XML syntax, is that just ISSUE-36 or is it related to extensibility?
Harold: current XSD is monolithic, doesn't use any extensibility of XML Schema, could try to modularize it more
MK: builtins should be in a separate document since they are shared by different dialects
<Harold> The current builtins can also be used by PRD.
<Harold> Yes, Paula, kind of Core builtins.
<Harold> So the Core could start with the builtins.
csma: asks Paula to include the additional builtins DaveR mentioned in email
Paula: already in, but will check
csma: what's the issue with binding patterns?
MK: no issue, just prefers not to
have binding patterns in semantics, complicates things
... e.g. a + b = 3, if a and b are unbound have infinite relation so impossible to compute
<AxelPolleres> allowing ONLY constants would be too limiting.
MK: binding pattern more
complicated, e.g. for plus have to say that two need to bound
but one can be unbound
... have to specify this but don't have to put them in the semantics
Axel: binding pattern is similar
issue to safety
... binding patterns give patterns such that builtin is safe (finite extension) so similar to rule safety
... agrees with MK it is not concern for the semantics
<Harold> To prevent 'plus' used as in plus(3 ?a ?b), only allowing deterministic calls like plus(?x 1 2), we can specify the single 'plus' [binding pattern or] mode plus(out in in).
Axel: rule engines vary with the binding patterns they support, some might further limit patterns (e.g. only run plus forwards)
<AxelPolleres> you mean swrl?
MK: what do they do in SWRL?
MikeDean: they don't specify binding patterns, just as relations
<AxelPolleres> +1 to mike dean: exactly, if not specified, then it causes difficulties for implementers...
Hassan: tend to disagree that binding patterns is part of semantics, whether rules pass value by unification is important
MK: part of the semantics, point is they are not part of the model theoretic semantics
<AxelPolleres> I also agree that it need't be part of the semantics.
<AxelPolleres> but for a dialect, not only the list of built-ins, but also the supported binding patterns should be specified.
<AxelPolleres> this is more an issue of arch/extensibility than semantics.
Hassan: can have models for infinite relations perfectly happening
csma: can't we just have simplest binding patterns?
Axel: part of extensibility,
dialect should be allowed to specify binding patterns
... in generic defn of builtins allow higher order predicates (builtins with take as inputs other predicates), make sense for things like aggregation
Axel: is this something we want to account for in BLD?
Hassan: higher order or meta? Seems more like meta - e.g. prolog's univ.
Axel: no, takes extension of another predicate
csma: perhaps beyond scope of BLD
<josb> it is beyond bld!
<scribe> ACTION: Axel to post use case for higher order builtins [recorded in http://www.w3.org/2007/11/20-rif-minutes.html#action08]
<rifbot> Created ACTION-384 - Post use case for higher order builtins [on Axel Polleres - due 2007-11-27].
<AxelPolleres> one more point!!!
<scribe> ACTION: Patranjan to start document for the list of builtins [recorded in http://www.w3.org/2007/11/20-rif-minutes.html#action10]
<rifbot> Created ACTION-385 - Start document for the list of builtins [on Paula-Lavinia Patranjan - due 2007-11-27].
<PaulaP> regrets for the next telecon
<LeoraMorgenstern> I'll do it.