RIF Telecon 29 May 2007 -- 29 May 2007

<ChrisW> Scribe: Jeff Pan

<csma> scribe: Jeff Pan

<csma> scribenick: Jeff_Pan

<AxelPolleres> Hi!

<ChrisW> scribenick: JeffP

Admin

ChrisW: we don't have telecon next week

as the weekend will be our FTF

<csma> no

F2F

ChrisW: any changes to the agenda?

FTF starts at 9am

<csma> we do not hear you

<ChrisW> hmmm

(JeffP will arrive about noon)

josb: THe meeting will be at the university, outside the centre
... two ways to reaching it: taxi or bus
... see the wiki

csma: there is a need of using the phone

Liason

ChrisW: any updates?

none

<PaulVincent> PRR update: final submisson supposedly in for June OMG meeting

Core - Builtins

<sandro> Okay, Jeff, I answered the registration survey for you.

Thanks, sandro

<csma> ACTION-291 done

<sandro> (bringing us to 25 people who will attend at least part of the meeting.)

ACTION-291 done

<csma> ACTION-290 complete

ACTION-285 done

ACTION-282 done

<csma> ACTION-281 continued

ACTION-280 complete

ACTION-230 dropped

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

ChrisW: any discussions on this?

Axel_Polleres: is there a function translating URIs to strings and vice versa?

<PaulaP> fn:encode-for-uri is what you are looking for, Axel?

csma: maybe, but at the biginning we should focus the core ones

<AxelPolleres> paula, this is only one direction, the otherone is problematic, I think

<PaulaP> fn:escape-html-uri?

ChrisW: seems that one direction is possible

Axel_Polleres: maybe good to keep it recorded

<josb> yes

yes

<MichaelKifer> we should have toString for every type. why reinvent the wheel

<DavidHirtle> yes, anyURI

<DavidHirtle> http://www.w3.org/TR/2001/REC-xmlschema-2-20010502/#anyURI

<PaulaP> http://www.w3.org/TR/xpath-functions/#anyURI-functions

<PaulaP> there is just one function fn:resolve-uri

<AxelPolleres> what i wanted to state is that we might need something special for rif:iri

anyURI is not a sub-type of string

<PaulVincent> http://www.w3.org/TR/xmlschema11-2/#anyURI

ChrisW: is there a general transforming functions?

<PaulaP> http://www.w3.org/TR/xpath-functions/#casting

<AxelPolleres> An example from RDF: FOAF uses tel: URI's for telephone numbers, vCard uses strings... if you want to convert, you need something...

ChrisW: we need to reuse XML Schema, if possible

<AxelPolleres> Summary: probably we only need to extend the casting of XPath/XPQuery to new primitive datatypes we introduce/use, such as e.g. rif:iri, yes?

<AxelPolleres> what I mean is, we should nail down which basic sets of types AND which conversions we allow.

<PaulaP> the wiki page points already to that document

<sandro> http://www.w3.org/TR/xpath-functions/#func-concat

<PaulaP> perhaps is better to first decide on F&O

<PaulaP> and then link them

ChrisW: add links to their definitions

<PaulaP> ok

<PaulaP> I see

<PaulaP> thanks, Sandro

<scribe> ACTION: Sandro to put links to the F&O wiki page [recorded in http://www.w3.org/2007/05/29-rif-minutes.html#action01]

<rifbot> Created ACTION-292 - Put links to the F&O wiki page [on Sandro Hawke - due 2007-06-05].

ChrisW: we should read F&O before the FTF

in particular on what should/shouldn't in core and why

<AxelPolleres> +1 to harold, allow all in principle, and only declare a subset normative!

<sandro> Sandro: so the use of an xpath function that we exclude will not be syntactically detectable.

csma: we have a mandated list datatype in the charter

<Harold> A F&O sequence can be represented as nested Uniterms in the normal Prolog manner.

<scribe> ACTION: on Axel to prepare a datatype conversion proposal [recorded in http://www.w3.org/2007/05/29-rif-minutes.html#action02]

<rifbot> Sorry, couldn't find user - on

<sandro> ACTION: Axel to prepare a datatype conversion proposal [recorded in http://www.w3.org/2007/05/29-rif-minutes.html#action03]

<rifbot> Created ACTION-293 - Prepare a datatype conversion proposal [on Axel Polleres - due 2007-06-05].

<Harold> [e1, e2, ...., eN] --> Uniterm(e1 Uniterm(e2 Uniterm(... eN)...))

<ABG> I have to go; see you in Innsbruck

<scribe> ACTION: on Harold to prepare a proposal on sequences [recorded in http://www.w3.org/2007/05/29-rif-minutes.html#action04]

<rifbot> Sorry, couldn't find user - on

<sandro> ACTION: Harold to propose a treatment of sequences [recorded in http://www.w3.org/2007/05/29-rif-minutes.html#action05]

<rifbot> Created ACTION-294 - Propose a treatment of sequences [on Harold Boley - due 2007-06-05].

<sandro> Gary: Is there a way I can tell if a function is logical or non-logical (built-in) ?

<sandro> csma: (ie unbound variables allows as argument).

<sandro> chris: ie can only tell at runtime?

sandro: one solution by Harold is to have one builtin ...

ChrisW: Harold, Michael?

MichaelKifer: I propose a special syntax

<Harold> User-defined functions can often be inverted: [a b c d] = append(?L ?M) tries to split [a b c d] into all prefix-postfix pairs ([],[a b c d]), ([a],[b c d]), ..., ([a b c d],[]).

<AxelPolleres> that would mean we talk about different types of atoms then...

MichaelKifer: different dialects have different name spaces

<AxelPolleres> that would touch on the CORE model, wouldn't it?

<sandro> Sandro: I think implementors would like to know when they are seeing a call to a builtin vs using a logic function. Although we might want to hide the difference.

<Harold> Functions can already be defined in RIF WD1 using equations: append([] ?M) = ?M append([?H | ?T] ?M) = ...

ChrisW: we seem to need more discussion on this

csma: we might not need a solution in the 2nd WD

AxelPolleres: we need to have a solution

<sandro> (Axel, your audio quality is quite poor)

<scribe> ACTION: chairs to open an issue [recorded in http://www.w3.org/2007/05/29-rif-minutes.html#action06]

<rifbot> Sorry, couldn't find user - chairs

<sandro> proposed ACTION: DeborahNichols open issue on whether the abstract syntax will distinguish between builtins and logic functions.

<sandro> ACTION: DeborahNichols open issue on whether the abstract syntax will distinguish between builtins and logic functions. [recorded in http://www.w3.org/2007/05/29-rif-minutes.html#action07]

<rifbot> Sorry, couldn't find user - DeborahNichols

<sandro> ACTION: Deborah open issue on whether the abstract syntax will distinguish between builtins and logic functions. [recorded in http://www.w3.org/2007/05/29-rif-minutes.html#action08]

<rifbot> Created ACTION-295 - Open issue on whether the abstract syntax will distinguish between builtins and logic functions. [on Deborah Nichols - due 2007-06-05].

<AxelPolleres> I can propose a strawman syntax, taken from a system I work with which at least opens some requirements

<Harold> For logic functions there are again two possibilities: uninterpreted functions and interpreted (equation-defined) functions. Only the latter ones can be compared with builtins.

<sandro> ACTION: Axel to propose strawman for resolving this new issue (builtins vs logic functions) [recorded in http://www.w3.org/2007/05/29-rif-minutes.html#action09]

<rifbot> Created ACTION-296 - Propose strawman for resolving this new issue (builtins vs logic functions) [on Axel Polleres - due 2007-06-05].

Core - XML Syntax

Christw: Did you propose the XML syntax for RIF Core will not use stripe-skipping

Gary: yes

<sandro> we lost you Chria.

<sandro> we lost you Chris.

<Harold> Gary, you're right, it's quite verbose to write And(arg1(Conjunct1) arg2(Conjunct2) ... argN(ConjunctN)) rather than just And(Conjunct1 Conjunct2 ... ConjunctN).

<PaulVincent> Apologies had to leave.

csma: We should avoid discussions on stripe-skipping in FTF

<Harold> I propose to omit stripes if they are 'artificial' (position-simulating) anyway such as arg1, arg2, ..., argN.

ChrisW: we need to proposal from abstract syntax to W3C XMLS

<Harold> Think back to the IRI vs. URI discussion: human readability.

sandro: human readability should not have 0 weight

<PaulaP> :)

<Harold> rdf:_1, rdf:_2, ...

<sandro> I basically agree.

<sandro> Gary: I agree, unless the position has some special role, as in uniterm.

<sandro> +1 Gary

Gary: if the first argument is a name for function etc. then it should keep a stripe

ChrisW: do we need a propsoal in FTF on this?

thanks, Harold

<Harold> Gary also mentioned the remaining positional arguments could get stripe-skipped.

UCR

<ChrisW> allen - your action #274 status?

<PaulaP> yes

<csma> http://www.w3.org/2005/rules/wg/wiki/Design_Constraints

<PaulaP> I added a couple of requirements today

<PaulaP> yes

Issues 14-21 to be closed?

Deborah: does it mean resolved?

sandro: delay till phase 2

<sandro> the typical term is "Postponed"

AOB

csma: we will start at 9am on June 2

<AxelPolleres> THe brave can meet at 6am to climb a mountain!

<csma> a small mountain?

<PaulaP> +1

<csma> +1

<sandro> rdml (roll down mountain laughing...)

<PaulaP> bye

<AxelPolleres> +1

<GaryHallmark> better than moving the mountain

<Francois> bye.

- DRAFT -

RIF Telecon 29 May 2007

29 May 2007

Attendees

Contents

Admin

F2F

Liason

Core - Builtins

Core - XML Syntax

UCR

AOB

Summary of Action Items

Scribe.perl diagnostic output