<csma> http://lists.w3.org/Archives/Public/public-rif-wg/2009Sep/att-0045/15-Sept-2009-rif-mins.html

<csma> PROPOSED: to accept the minutes from Sept 15 telecon

<csma> RESOLVED: to accept the minutes from Sept 15 telecon

Liason

Sandro confirmed that OWL 2 went to PR with auto-inc ref to old version of XML Schema datatypes

<ChrisW> Owl is at PR

Actions

<ChrisW> close action-925

<trackbot> ACTION-925 Review RIF Guide closed

<ChrisW> close action-924

<trackbot> ACTION-924 Review test cases closed

<ChrisW> close action-923

<trackbot> ACTION-923 Review RIF Guide closed

<ChrisW> close action-921

<trackbot> ACTION-921 Review the OWL2RL by 1 week closed

<ChrisW> close action-917

<trackbot> ACTION-917 Review changes to xml document by 2 weeks closed

<ChrisW> close action-916

<trackbot> ACTION-916 Review changes to xml document by 2 weeks closed

<ChrisW> close action-915

<trackbot> ACTION-915 Update xml data document by 1 week closed

close action-910

<trackbot> ACTION-910 Update Sept 1 minutes closed

close action-908

<trackbot> ACTION-908 Fix/update AssertRetract test case closed

action-907 pending-review

close action-833

<trackbot> ACTION-833 Send email about the XML schemas TF closed

close action-791

<trackbot> ACTION-791 Write test cases for PRD closed

<cke> action 791 is done

<trackbot> Sorry, couldn't find user - 791

Public comments

<csma> http://lists.w3.org/Archives/Public/public-rif-comments/2009Sep/0008.html

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

Christian: re comment #2, func: namespace, Christian and Dave both recalled agreement to create rif namespace for predicates and for functions, and have single namespace for functions (which include XF&O operators)

Chris: didn't want to force people to remember which is in RIF and which in XF&O

Sandro: that is a weak reason, because it only really applies to hand authoring
... RIF is supposed to be automatically generated (so only translator implementers need to sort the namespaces)
... in the cases where the names are the same are our semantics really intended to be identical?

Gary: there is a difference with error handling.

Sandro: should we make this difference clearer? See Jim's comment on 'adapted' v. 'adopted'

<josb> we're modifying; this is another motivation for using a different namespace

Jos: we use them in a different way, we package some XF&O functions as predicates (which don't exist in F&O)
... and for errors we leave the behaviour undefined, but in XF&O there is a defined error behaviour

Chris: and that difference is explained in DTB
... the error behaviour is given in bullet 5 at start of section 4 in DTB

Jos: should put 'adapted' in overview and explain it there

<sandro> change from 'adopted' to 'adapted' in overview and abstract

[Chris makes change in real time]

<sandro> The adaptation is primarily to reframe boolean functions as% logical predicates and that instead of error results, we do not define results, as noted above.

Move on to comment #3 on xs:duration

Chris: there is no mapping to an entity in the domain, no defined value space

http://www.w3.org/TR/xmlschema11-2/#duration

Dave: actually it is defined

Dave: it seems to be defined in terms of a pair of months and seconds, with a partial ordering of pairs

Jos: the problem is that with other partial ordering we can't define equality in the way we need

Sandro: there is an equality defined (as in identity)
... this is mostly editorial, can't we just say it is partially ordered?

<sandro> chris: but being partially ordered is okay.

Chris: partial order is OK, but the problem is identity.

Jos: the original decision was made when we were thinking about XML Schema 1.0 and the 1.1 has improved on that

Sandro: we have to use 1.0 for now anyway and if 1.1 advances then we can switch

Chris: can just remove comment, "not included for historical reasons"

Sandro: and there was the goal of alignment with OWL

<sandro> http://www.w3.org/2005/rules/wiki/DTB#Symbol_Spaces

Chris: edited DTB to remove comment in 2.2.1

Looking at comment #4

Sandro: just a question of which spec to reference - XDM or XSD

<sandro> sandro: maybe lets not do this, since it increases our dependency on XSD 1.1

<sandro> point 4: leave it for now -- we'll fix if XSD 1.1 is mature enough.

Christian: in RIF-XML referred to XS 1.1 for all, didn't take those two types out and point to XDM

Move on to point #5

Sandro: just an error, fix in section 2.3 of DTB, the four unsigned should be shifted over to under nonNegativeInteger

<josb> see: http://www.w3.org/TR/xmlschema-2/#built-in-datatypes

Point #6 - negative guards

Sandro - believes our spec is correct but comment suggests we need better explanation

<Gary> but oddly we didn't name it isLiteralAndNotInteger

<sandro> ACTION: sandro to draft text for DTB on negative guards -- due today [recorded in http://www.w3.org/2009/09/29-rif-minutes.html#action01]

<trackbot> Created ACTION-927 - Draft text for DTB on negative guards -- due today [on Sandro Hawke - due 2009-10-06].

Point #7 trivial, sandro will fix.

Point #8 - type promotion

<josb> +q

<sandro> jos: promotion only comes into play with disjoint value space

Dave: the issue is that XPath does the type promotion in the language, so the F&O only apply to uniform args

Jos: but for say int->decimal is OK because those are subtypes, the problem is for things like int->double which are disjoint
... not well defined in XPath and so we are just inheriting that

<cke> Java does type promotion

<josb> http://www.w3.org/TR/xmlschema-2/#built-in-datatypes

<josb> For simplicity, each operator is defined to operate on operands of the same type and return the same type.

See http://www.w3.org/TR/xpath20/#promotion

Jos: agree that if you read it strictly then add(double, decimal) in RIF is undefined at present

<Gary> in my implementation, decimal + double = decimal

<cke> What is decimal? Is this larger than double?

<Gary> decimal is usually mapped to java.math.BigDecimal

<cke> OK, I see.

Christian: from an implementation point of view would expect promotion as in Gary's implementation

Jos: in xpath decimal + double -> double

<josb> http://www.w3.org/TR/xpath20/#promotion

Gary: maybe just following common sense, because decimal can be arbitrary precision

<cke> Decimal definition in the .net platform: http://www.yoda.arachsys.com/csharp/decimal.html

http://www.w3.org/TR/xpath-functions/#func-signatures

Jos: one argument is to avoid introducing spurious precision
... the intention in DTB was to do promotion

Dave: there is a description of type promotion, specially for xs:decimal to xs:float/double, in F&O so maybe we just need to refer to this more explicitly

<josb> http://www.w3.org/TR/2007/REC-xpath-functions-20070123/#op.numeric

<josb> For simplicity, each operator is defined to operate on operands of the same type and return the same type.

Jos: the mapping is only defined for the case where the values have have the same type, see link above

<scribe> ACTION: jos to update mappings section of DTB to resolve type promotion issue [recorded in http://www.w3.org/2009/09/29-rif-minutes.html#action02]

<trackbot> Created ACTION-928 - Update mappings section of DTB to resolve type promotion issue [on Jos de Bruijn - due 2009-10-06].

Sandro: other non-trivial comment is #11 on numbering list elements from 1 to match substring etc

<Gary> +1 for +1

Sandro: switching the indexing of lists might affect definition of negative indexes, as well as require rework of test cases and formal semantics

<sandro> +1 "it is unfortunate" and leaving it as is.

<Gary> 0 for 0

<josb> http://www.w3.org/2005/rules/wiki/DTB#Numeric_Functions

Jos: [Referring to his action to edit DTB] what are the a1 and a2 in the mapping definition?

Chris: they are the actual values passed in

Sandro: will do other editorial changes in response to Jim comments, aim to do that today, will need quick review of text by WG, so can send response tomorrow

OWL 2 RL publication

Jos: it can be published as is

<csma> PROPOSED: Publish RIF-OWL2/RL [5] as first public working draft.

<adrianpaschke> +1

<ChrisW> +1

<cke> +1

<csma> +1

<sandro> +1 (W3C)

<dave> +1 (HP)

<ChrisW> +1 (IBM)

<Gary> +1 Oracle

<LeoraMorgenstern> +1 (self)

<MichaelKifer> +1

<csma> RESOLVED: Publish RIF-OWL2/RL as first public working draft.

XML Data

Christian: a number of substantial comments from Dave and Gary
... can we publish with editorial notes?

Dave: Ok if editorial notes on the name mapping and semantics issues

Gary: this is a first working draft so OK with editorial comments, there are quite a few of those needed

<csma> PROPOSED: Publish RIF-XMLdata as FPWD, subject to agreement on editorial notes

<ChrisW> +1 IBM

<Dave> +1 HP

<Gary> +1 Oracle

<ChrisW> sandro: +1 (W3C)

<adrianpaschke> +1

<LeoraMorgenstern> +1

<josb> +1 FUB

<MichaelKifer> +1(self)

RESOLUTION: Publish RIF-XMLdata as FPWD, subject to agreement on editorial note

Guide

Chris: would like to change title before first publication

Sando: ok to publish as first working draft

Michael: how about "Overview of RIF"

Leora: "Overview of RIF Documents"?

Michael: it is more than the documents

<adrianpaschke> what about "Introduction to RIF"

<csma> PROPOSED: extend to 10 to

<csma> RESOLVED: Extend to 10 to

<ChrisW> PROPOSED: publish "Guide" as "Overview"

<csma> "Roadmap to RIF documents"

<josb> ACTION: completed to http://www.w3.org/2005/rules/wiki/DTB#Numeric_Functions [recorded in http://www.w3.org/2009/09/29-rif-minutes.html#action03]

<trackbot> Sorry, couldn't find user - completed

Chris: key thing is to have the short name be overview rather than guide

<MichaelKifer> +2

<josb> +1 RIF Overview

Chris: who prefers RIF Overview

<ChrisW> Prefer RIF Overview:

+1

<ChrisW> +1

+1 (Sandro)

<ChrisW> prefers RIF Document overview

<LeoraMorgenstern> +1

<csma> +1

<ChrisW> PROPOSED: publish "Guide" as "RIF Overview"

<ChrisW> +1 (IBM)

<adrianpaschke> +1

<Gary> +1 Oracle

<Dave> +1 HP

<Sandro> +1 W3C

<josb> +1 (FUB)

<MichaelKifer> +1

<ChrisW> RESOLVED: publish "Guide" as "RIF Overview"

<LeoraMorgenstern> +1

Test cases

Chris: reviews were positive

Sandro: OK to publish

<csma> PROPOSED: Publish RIF Test cases as 2nd public working draft.

<LeoraMorgenstern> When I reviewed what was then called The Guide, I reviewed it as a guide to documents, not as a RIF Overview.

<ChrisW> +1

+1

<adrianpaschke> +1

<LeoraMorgenstern> I made the point in my review about the slant of the guide, but since I felt it was basically just a guide to the documents, I didn't feel it was of great importance to change.

<MichaelKifer> +1

<LeoraMorgenstern> +1

<josb> +1 (FUB)

<ChrisW> Sandro: +1 (W3C)

<Gary> +1 Oracle

RESOLUTION: Publish RIF Test cases as 2nd public working draft

<LeoraMorgenstern> However, if this document is really considered an overview to RIF, I would make more pointed suggestions about how the document should be modified.

<ChrisW> leora: Overview is not finished, we can change it

<csma> RESOLVED: Publish RIF Test cases as 2nd public working draft

<csma> Next meeting 13 October

<MichaelKifer> There was nothing in the document that suggested it is just a list of documents. The old title was "A Guide to RIF" not to RIF documents.

<LeoraMorgenstern> Chris, I realize that. I just wanted to record what I said, since Dave didn't.

<LeoraMorgenstern> Okay, Michael, I misunderstood.

<LeoraMorgenstern> I am still in favor of publication of this as a first

<LeoraMorgenstern> working draft; however, my comments would have been different.