W3C

- DRAFT -

RIF Telecon 27-Jan-2009

27 Jan 2009

Agenda

See also: IRC log

Attendees

Present
AdrianP, Stella_Mitchell, ChrisW, LeoraMorgenstern, Harold, Sandro, Hassan_Ait-Kaci, GaryHallmark, Michael_Kifer, csma
Regrets
DaveReynolds, PaulVincent, JosDeBruijn
Chair
Chris Welty
Scribe
Harold

Contents


 

 

<ChrisW> Hi Adrian, can you scribe today?

<ChrisW> Scribe: Harold

<ChrisW> scribenick: Harold

<AxelPolleres> dialing

<ChrisW> http://lists.w3.org/Archives/Public/public-rif-wg/2009Jan/att-0007/06-rif-minutes.html

<AxelPolleres> (I can scribe, if needed. feeling guilty for not having done quite some time)

F2F13

<ChrisW> http://www.w3.org/2002/09/wbs/38457/f2f13_dates/results

Chris: Will be last f2f, if things go according to plan.

<AdrianP> would be great if we can fix dates as early as possible

Chris: Most people who responded so far are ok if F2F13 will be in the US.

Axel: Ideally closer to the AC meeting (Apr 24?).

<AxelPolleres> no, march 24th, that's why i said, I guess that's unrealistic.

Axel: Perhaps near Madrid, collocated with WWW2009?
... could arrange a meeting room in Madrid.

<AxelPolleres> ok

<AxelPolleres> will check back

Liason

Chris: Split OWL comment into two.
... Try to move fwd with OWL WG collaboration: Axel, Jos would be key.

Axel: Yes.

Public COmments

Chris: Seems done.

Action Review

<AxelPolleres> Sandro, when is the next OWL f2f? maybe some overlap or adjacent meeting would make sense?

Gary: Still working on Core as specialization of PRD.

<Hassan> Parser continued - will work on it this Feb.

Issue-50

<AdrianP> there was a discussion on the last F2F about using meta data to indicate multi-valued

<AdrianP> frames

Harold: Terminologically, let's separate the more specific "metadata" and more general "annotation".
... I am ok with no semantic import of annotations.

<AxelPolleres> adrian, would such metadata be "per frame" or "per ruleset"?

Adrian: Maybe metadata for multi-valued frames.
... proposed by Michael Kifer?

Chris: Cardinality with metadata?

Sandro: Was just in response to PRD discussion.

<AdrianP> yes, agree too

Michael: fine.

<ChrisW> PROPOSED: Core, BLD, PRD will not have semantic metadata; removing metadata from a ruleset will not change the entailments - closing ISSUE-50

<AdrianP> +1

+1

<Gary> +1

<Michael_Kifer> +1

<Hassan> 0

<ChrisW> Sandro: +1

<AxelPolleres> +1

<ChrisW> +1

<ChrisW> RESOLVED: Core, BLD, PRD will not have semantic metadata; removing metadata from a ruleset will not change the entailments - closing ISSUE-50

<ChrisW> ACTION: Chris to close issue-50 [recorded in http://www.w3.org/2009/01/27-rif-minutes.html#action01]

<trackbot> Created ACTION-699 - Close issue-50 [on Christopher Welty - due 2009-02-03].

ISSUE-83

<ChrisW> Ambiguity in PS

Hassan: As far as I can remember, all remaining ambiguities are mild.
... First, canonical PS has no serious ambiguity.
... Then, Abridged PS, which is even less binding, may have more ambiguity (e.g. dash as minus vs. as part of identifiers)

<AdrianP> yes, abridge syntax as compact representation, e.g., 2 - 1 (numbers) vs. "2" - "1" (text)

<ChrisW> PROPOSED: Close ISSUE-83 - the existing presentation syntax ambiguities are resolvable

Chris: Let's deal with ambiguities in Abridged PS as they arise.

<AdrianP> +1

<Hassan> +1

+1

<AxelPolleres> +1

<ChrisW> +1

<Michael_Kifer> +1

<ChrisW> RESOLVED: Close ISSUE-83 - the existing presentation syntax ambiguities are resolvable

DTB

Chris: Dave had suggested a meta-predicate for equality.

Axel: For normal equality, we have all we need.
... For negated, we may need what Dave suggested.

<ChrisW> pred:isLiteralEqual(?l1 ?l2)

<ChrisW> pred:isLiteralNotEqual(?l1 ?l2)

Chris: Is pred:isLiteralEqual false if I1 or I2 is not a literal?

Axel: Yes.

Chris: What does it solve?

Harold: Would isLiteralEqual's mapping to false of both "wrong arguments" and "non equal arguments" be acceptable to RIF users (and to other communities)?

Axel: Perhaps parameterize.

Chris: maybe type as first arg?

Axel: Def of built-ins could depend on underlying datatypes.

<AxelPolleres> literalNotEqual

<AxelPolleres> isLiteral, isNotLiteral ?

Hassan: Either parameterized type args or already instantiated in built-in identifier.
... (from Jos)

Axel: fair enough.

Chris: There are 'equal' preds for all datatypes?

Axel: Yes, even though not named completely uniformly as such.

Chris: The issue is for numeric-equal if args are of different numeric types.

Sandro: XPath tells you how to do the type promotion.

Chris: Yes, e.g. equal for integer and double.

Axel: Idea was to back up each datatype with an equal and a notequal predicate.
... We could have done differently.
... Will draft something.
... For email discussion.

<AxelPolleres> ... will draft something on both the generic guards and the generic equals/notequals predicate.

<ChrisW> ACTION: axel to report on possibility of closing literal-not-equal issue-80 [recorded in http://www.w3.org/2009/01/27-rif-minutes.html#action02]

<trackbot> Created ACTION-700 - Report on possibility of closing literal-not-equal issue-80 [on Axel Polleres - due 2009-02-03].

Postpone to next week's telecon ISSUE-81 [10] (Support for additional OWL-RL data types)

ISSUE-81 [10] (Support for additional OWL-RL data types)

Test cases

http://www.w3.org/2005/rules/wiki/Category:Test_Case

<ChrisW> Test case: http://www.w3.org/2005/rules/wiki/Argument_names_in_uniterms_1

Name ::= UNICODESTRING

UNITERM ::= Const '(' (TERM* | (Name '->' TERM)*) ')'

<Michael_Kifer> maybe call negative syntax examples "syntax errors"?

<ChrisW> PROPOSED: Accept test case http://www.w3.org/2005/rules/wiki/Argument_names_in_uniterms_1

<csma> +1 to look them all first

<AxelPolleres> What about this one: _p(http://example.com/example->color->"green")

<AxelPolleres> ?!?

<csma> ak csma

<AxelPolleres> or this one:

<AxelPolleres> _p(http://ex"ampl"e.com/exam"ple#color->"green")

Stella: Arg names are defined 'negatively': every UNICODESTRING that's not a constant is ok.

Chris: Hassan, does this meet your expectation?

Hassan: Would have been easier to quote them, but it's ok.

Michael: Wasn't this a different case?

Hassan: Right.

Michael: Maybe find a universal solution, maybe single quotes.

Hassan: yes.

Sandro: A valid name could have a space inside, so need single quotes.
... Why not double quotes?

<AxelPolleres> again the two ugly examples, which seem awkward to parse: _p(http://example.com/example->color->"green") _p(http://ex"ampl"e.com/exam"ple#color->"green")

Hassan: Keep the grammar as is. Fine.

Michael: Looks a bit strange.

<Hassan> No - not keep the grammar as is - modify it to allow for STRING there

<AxelPolleres> ncnames only?

Chris: If 'strangeness' bothers you then the whole named-arg thing should bother you.

<AxelPolleres> how bout that?

Hassan: Grammar should allow string as arg name.

<AxelPolleres> we could say: ncnames OR (singlequote uc-strings singlequote)

Chris: Examples?

Axel: Any issues with single quotes?

Chris: Would be (even) harder to read.

Sandro: Bad idea to use up single quotes for corner case.

Michael: agree with Sandro.

<StellaMitchell> can you just limit it to alphanumeric characters, no spaces, etc

<StellaMitchell> something like that

<Gary> of course, we could remove NAUs...

<AxelPolleres> so, you say: Name ::= UNICODESTRING

Hassan: Suggest to postpone until we finish use cases.
... So can look for solutions for all issues.

<Hassan> I agree

Chris: Before we decide test case we need to decide syntax.
... No one against strings?

Sandro/Axel: Well, we.

<AxelPolleres> :-)

<AxelPolleres> http://www.w3.org/TR/xmlschema-2/#NCName

Chris: Would those in favor of named args be fine with specializing Name ::= UNICODESTRING into Name ::= CIDENTIFIER.
... Or just use quotes.

Harold: Yes.

(specialized names)

<AxelPolleres> oh my, full URIs only, no abbreviations?

Chris: Which ones are URIs and which ones only look like URIs?

<ChrisW> Prefer NCNAMES

<LeoraMorgenstern> which is the choice that doesn't allow abbreviations?

<LeoraMorgenstern> Something that forces you to do full URIs

<StellaMitchell> +1 ncnames

<AxelPolleres> +1 NCNames

+1 (to specialized names, which we have anyway, albeit defined 'negatively')

<Gary> _p is an ncname

<LeoraMorgenstern> I'm still confused about what an nc name is.

<Hassan> 0 - not sure at this point

<LeoraMorgenstern> I looked at Axel's link, but need more time.

<AxelPolleres> -1 to negatively defined (see weird corner cases above)

<Michael_Kifer> I prefer quotation

<Hassan> I agree with Axel's proposal (NCNAME or else quoted)

<StellaMitchell> Gary raised a problem with ncname?

Gary, this is interesting but they can occur elsewhere in an expression, too.

_p(_p->1) might be a problem.

<ChrisW> straw poll: is this worth redoing last call?

<StellaMitchell> so, Harold, it would be ncname that is not a valid constant?

<AxelPolleres> Sandro, why would (ncname or quotetdstring) not work for you, except taste?

<AdrianP> -1

-1

<LeoraMorgenstern> -1

<Hassan> -1

<Michael_Kifer> -1

<LeoraMorgenstern> Given that it's already 12:21, let's accept the 2 test cases we have looked at,

<ChrisW> _p(2->"green")

<LeoraMorgenstern> rather than looking at all test cases first.

<sandro> am I finally connected?

<Michael_Kifer> +1 accept case 2

<LeoraMorgenstern> +1 accept case 1 and accept case 2

<Hassan> +1 accept both

<ChrisW> http://www.w3.org/2005/rules/wiki/Argument_names_in_uniterms_3'

Stella, some parsers might not notice that both _p's in _p(_p->1) have the same name.

<LeoraMorgenstern> Chris, the link doesn't work for me

<sandro> link worked for me.

<LeoraMorgenstern> http://www.w3.org/2005/rules/wiki/Argument_names_in_uniterms_3

+1

<ChrisW> _p("http://example.com/example#color"->"green")

<LeoraMorgenstern> Right, I have the case; just pointing out that clicking on that link isn't working for me.

<AxelPolleres> 3 and 4 are problematic with respect to what we discussed now.

<sandro> ewwwwwww

<AxelPolleres> _p(http://example.com/example->color->"green") _p(http://ex"ampl"e.com/exam"ple#color->"green") are similar to 4.

<Hassan> Stella, Harold, Gary: solution is to tokenize _foo as as Local no matter what and reject _p as named-argument

Sandro: Just editorial change, since only changing non-normative PS, not changing normative XML.

<ChrisW> PROPOSED: Accept http://www.w3.org/2005/rules/wiki/Argument_names_in_uniterms_1

<AxelPolleres> +1

<LeoraMorgenstern> +1

<ChrisW> +1

+1

<Hassan> +1

<sandro> sandro: My preference is for a solution where argument names can be (ncname|string) -- users get to pick whether to have quotes.

<sandro> +1

<Michael_Kifer> +1

<Gary> +1

<StellaMitchell> yes

<ChrisW> PROPOSED: Accept http://www.w3.org/2005/rules/wiki/Argument_names_in_uniterms_2

<sandro> +1

<LeoraMorgenstern> +1

+1

<Gary> +1

<ChrisW> +1

<AdrianP> +1

<Hassan> +1

<AxelPolleres> +1

<csma> RESOLVED: Accept http://www.w3.org/2005/rules/wiki/Argument_names_in_uniterms_1

<csma> RESOLVED: Accept http://www.w3.org/2005/rules/wiki/Argument_names_in_uniterms_@

<ChrisW> RESOLVED: Accept http://www.w3.org/2005/rules/wiki/Argument_names_in_uniterms_2

<ChrisW> AOB?

<csma> yes

<csma> yes to accept the action

<ChrisW> ACTION: csma to put this on agenda (syntax discussion on NAU names) [recorded in http://www.w3.org/2009/01/27-rif-minutes.html#action03]

<trackbot> Created ACTION-701 - Put this on agenda (syntax discussion on NAU names) [on Christian de Sainte Marie - due 2009-02-03].

<sandro> note that ncnames are allowed to contain minus signs.

Chris: put on agenda for next week Sandro's "My preference is for a solution where argument names can be (ncname|string)."

<Hassan> +1 to adjourn

Summary of Action Items

[NEW] ACTION: axel to report on possibility of closing literal-not-equal issue-80 [recorded in http://www.w3.org/2009/01/27-rif-minutes.html#action02]
[NEW] ACTION: Chris to close issue-50 [recorded in http://www.w3.org/2009/01/27-rif-minutes.html#action01]
[NEW] ACTION: csma to put this on agenda (syntax discussion on NAU names) [recorded in http://www.w3.org/2009/01/27-rif-minutes.html#action03]
 
[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.133 (CVS log)
$Date: 2009/01/27 17:30:53 $

Scribe.perl diagnostic output

[Delete this section before finalizing the minutes.]
This is scribe.perl Revision: 1.133  of Date: 2008/01/18 18:48:51  
Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Guessing input format: RRSAgent_Text_Format (score 1.00)

Succeeded: s/ different type/ different numeric types/
Succeeded: s/Chris: Why not double quotes?/...  Why not double quotes?/
Succeeded: s/stings/strings/
Succeeded: s/No on/No one/
Found Scribe: Harold
Found ScribeNick: Harold
Default Present: AdrianP, Stella_Mitchell, ChrisW, LeoraMorgenstern, Harold, Sandro, Hassan_Ait-Kaci, GaryHallmark, Michael_Kifer, csma
Present: AdrianP Stella_Mitchell ChrisW LeoraMorgenstern Harold Sandro Hassan_Ait-Kaci GaryHallmark Michael_Kifer csma
Regrets: DaveReynolds PaulVincent JosDeBruijn
Agenda: http://lists.w3.org/Archives/Public/public-rif-wg/2009Jan/0085.html
Got date from IRC log name: 27 Jan 2009
Guessing minutes URL: http://www.w3.org/2009/01/27-rif-minutes.html
People with action items: axel chris csma

[End of scribe.perl diagnostic output]