RIF telecon 22 January 2008 -- 22 Jan 2008

Admin

csma: agenda ammendments?
... none

<ChrisW> PROPOSED: accept minutes of telecon January 15

csma: objections to accept the minutes of last week

<ChrisW> RESOLVED: accept minutes of telecon January 15

Liaisons

csma: ChrisW, did you take a look at the response to Peter's comments?
... you said you wanted to take a look at them

ChrisW: they are ok
... responses will be sent tomorrow

<ChrisW> http://www.w3.org/2005/rules/wg/wiki/Response_to_PPS1

<ChrisW> http://www.w3.org/2005/rules/wg/wiki/Response_to_PPS2

csma: do you want an action to do that?

ChrisW: responses are to be found at the given links

csma: are there any liaison reports?
... Jos sent regrets

Mike_Dean: high-level discussions, no proposal to use MathML in the OWL 1.1. WG

csma: when is the next telecon
... is it before ours or not?
... probably not

csma: any other report?
... none

Mike_Dean: OWL 1.1. telecon after RIF telecon

F2F9

csma: F2F9 in one month from now

<Harold> Doug, do you plan to join us at the f2f9 in France?

csma: you should start to consider booking a flight and hotel
... useful to know how many people come

sandro: should I make a registration form?

csma: yes, this would be useful
... Harold told me that I should remind everybody how you get the ILOG rates at the hotels

<Harold> http://www.w3.org/2005/rules/wg/wiki/F2F9

csma: the page about F2F9

<csma> http://www.ilog.com/corporate/training/infoEurope.cfm

csma: there are a few selected hotels
... there is a list of hotels where rates were negotiated
... you must make reservations through ILOG

<josb> Hello, it turns out I returned a bit earlier than expected so I could join the conference.

csma: you may call ILOG directly

<ChrisW> hi jos

<csma> Desk at ILOG: +33 1 49 08 35 00

sandro: the phone number should be on the wiki page

<sandro> (should be is not)

<sandro> (should but is not)

<csma> Prani: +33 1 49 08 36 88

<ChrisW> that number connects you to a production rule system

<ChrisW> be careful with negation!

csma: call Prani if you need assistance

sandro: say that you'll attend a W3C meeting organized by csma

Harold: I still don't understand the workflow

csma: I will describe the workflow on the wiki page

<AdrianP> early confirmations of the booking are important to apply for funding

sandro: or the first person who makes a reservation

csma: please update the wiki if you have more information

Harold: working hours at ILOG necessary

csma: I will update the wiki page with all information needed
... anything else?

<Hassan> The hotels mentioned on the Wiki by CSMA standard aseptic beehive american style hotels situated in not so pretty area (e.g., the Hol.Inn express has a view on the Paris "Peripherique" freeway and a garbage dump!). If interested in good chep and pictureque places, ask me.

Actions review

csma: Action 152
... is Paul Vincent here?
... continued

<sandro> Hassan, are there any places you like that are a reasonable walk from ILOG?

csma: Action 274, ChrisW, did you get reply from Allan?

csma: Paula could perhaps take over Action 274

<Hassan> yes - and on the line of the 57 bus (8mins to ILOG). I myself stay here:http://www.cofrase.com/hotel/verlaine/

<Harold> Hassan, Can you put your *one* most recommended not too far away (< 20 mins), not too expensive (< 100 EU), pictureque place on the f2f9 wiki?

sandro: my action (292) is also continued

<Hassan> Harold: OK

csma: Harold, would you take over action 292?
... to add links for functions and operators in the BLD

end of next week?

is it ok?

yes

<Harold> Hassan, also it should not be noisy from the street.

csma: action 274's new deadline is Feb 1, 2008
... close action 362
... close 364, which is done
... sandro, 373?

sandro: continued

csma: action 375?

sandro: continued

csma: when can it be done?

sandro: before F2F9

csma: Feb 20, 2008
... action 378 continued
... 384 is pending discussion

<scribe> done

csma: action 396 done

it becomes pending discussion

csma: action on Dave on collation issues for built-ins

csma. action is done and the result becomes pending discussion

csma: better leave the action open as pending discussion

Dave: no issue needed

Issue 45 (lists)

csma: issue 45

<csma> PROPOSED: to close Issue 45 by including lists in BLD, where the pair

<csma> syntax will be used in the language definition and the sequence syntax

<csma> will be used for presentation and XML.

csma: discussion last week on lists
... proposal given above

<ChrisW> PROPOSED: to close Issue 45 by including lists in BLD, where the pair syntax will be used in the language definition and the sequence syntax will be used for presentation and XML.

<Harold> http://www.w3.org/2005/rules/wg/wiki/Core/List_Constructor

josb: the point is that it should always be possible to express the tail of the list

csma: the sequence symtax proposed by Harold allows for that
... should the resolution be more specific?

josb: I don't really understand the syntax on the page

Harold: recursive definition

josb: can you put a variable instead of list term?

Harold: sequence terms are syntactic sugar for pair terms

csma: you would not agree to the proposed resolution?

Harold: what do you mean by lang definition?

csma: the semantics

Harold: we need the pairs for the presentation syntax

csma: ChrisW, do you want to add something?

jobs: the lang definition and the semantics should not use different syntaxes
... why do we need the sequences?

Harold: very useful for syntactic sugar
... the pairs are needed for the semantics

<Harold> Sequences are regarded as syntactic sugar for Pairs.

ChrisW: probably you want to keep the syntax uniform

csma: why don't we define everything in terms if sequences?

josb: we can also use direct definitions

Harold: pairs are just binary constructors
... the semantics of sequences not that straightforward
... we could discourage people to use pairs

csma: another proposal would be to allow sequences and pairs
... but forbid to mix them

Harold: possibly doable also in XSD

csma: so better use them only on the semantics

Harold: translation table updated on the wiki

csma: allowing both - it seems you are in minority Harold, since most people supported josb's statement

josb: I can help with the semantics of sequences

Harold: vertical bar quite different on the meta syntax
... I would like to have an intermediant step

josb: it is just a proposed resolution

<AdrianP> semantic definitions for lists for both sequences and pairs already exist in the LP community

csma: try to rephrase the proposed resolution

<csma> PROPOSED: to close issue 45 by including lists in BLD, with the sequence syntax for presentation syntax and XML as described on the wiki...

csma: what do we do with pairs?
... do we remove them? do we allow them only in the presentation syntax?

Igor: why don't we leave the editors to do their job?

csma: the job of the editors is to represent the whole WG

Igor: I think we are getting into too low-level details

csma: so, to close issue and have lists
... but the issue is what kind of lists

josb: I object of the phrasing to the second resolution

csma: this means it is unfinished

ChrisW: better to simplify more or less equivalent issues here
... seqeunces are simpler to represent them and their semantics is established work
... let Jos take an action here

Harold: we only need the pairs, because we develop an interchange format

<csma> PROPOSED: close issue 45 by including lists in BLD as sequences (for both presentation and XML syntax)

<josb> +1

ChrisW: my personal opinion on this is given above

csma: any objections to the last proposed resolution?

<csma> PROPOSED: close issue 45 by including lists in BLD as sequences (for both presentation and XML syntaxes) as they are defined in the wiki page

csma: currently sequences are defined through pairs

josb: there is no definition of the semantics
... ChrisW did make a constructive proposal

csma: new proposal

<csma> PROPOSED: close issue 45 by including lists in BLD as sequences for obth presentation and XML syntaxes and with one of the standard semantics.

<Harold> PROPOSED: close issue 45 by including lists in BLD as sequences for obth presentation and XML syntaxes and with a binary-reduction semantics.

csma: ChrisW, do you have a proposal for the semantics of sequences?

ChrisW: we don't need to resolve this now

Dave: Harold would like to use the binary reduction semantics

<Harold> DaveR, exactly.

Dave: difference between presentation syntax that people might use and the internat syntax used in the document

<Harold> However, we could 'hide' the pair syntax in the semantics.

ChrisW: I don't think the constraint on the semantics needs to be in the resolution

<AdrianP> +1 for Chris

csma: proposed resolution would be only

<csma> PROPOSED: close issue 45 by including lists in BLD as sequences for both presentation and XML syntaxes.

josb: fine

Harold: fine

csma: further discussion on the latest proposal?

<DougL> I decline to abstain at this time.

<sandro> heh

csma: does somebody wants to abstain?
... RESOLVED

<csma> RESOLVED: close issue 45 by including lists in BLD as sequences for both presentation and XML syntaxes.

csma: for the 2nd WD on BLD, the next issue is on named arguments uniterms

Issue 44 (named arguments Uniterm)

csma: the current picture is that we have exactly the same number of supporting persons, objecting, and abstaining persons

josb: not clear enough what they are

csma: confused between the relation between named arguments and frames
... not anymore
... any other confusion?
... Gary, do you still object?
... Axel is not here
... Axel objected too

Gary: many possibilities put a burden on translators

<josb> +1 to Gary

Gary: better to have just one way of doing things

csma: this is the reason why Sandro asked about the implementation
... I hope we can have a resolution next week

ChrisW: not speaking as chair, support for simplifying things

<josb> +1 to simplifying

csma: I also support that

ChrisW: while we have an objector on each side, IMO people lean more to simplifying things

csma: I would also push in the direction of simplifying the interchange
... any other opinion?

<Harold> Also MichaelK was not here last week and is not here now: so polling next week would be better.

sandro: engineering decision this time

<josb> lexicographic ordering

<josb> (of arguments)

sandro: basically, you need an extension to become a standard

Dave: you need to translate your data model to RIF

csma: the only benefit is in writing arguments in any order
... you can also put a constraint on the order

sandro: you're developing separate modules and need to coordinate between them

<Harold> Christian, let's not do too much out of band: we have Signatures

<Harold> (http://www.w3.org/2005/rules/wg/wiki/FLD/Syntax)

sandro: if you have rules against RDF
... and the subject is a BNode
... it seems not to map to a frame

Dave: the point is about fixed number of arguments

sandro: now it is clear to me

Dave: I agree that there is a pressure to go for simplicity
... we can remove named arguments uniterms and then vote on future meetings

csma: maybe there is no real objection

<csma> PROPOSED: to close issue 44 by removing named argument Uniterms from BLD.

csma: the simplest proposal
... would there be objections here?

<AdrianP> named argument uniterms have a benefit if you want to use the rules on top different fact bases

Harold: I would abstain, Michael would object

Adrian: if you have named arguments and are not specific about the order, it is easier to build the rules on top of different fact bases
... such as relational databases

csma: the question is what are the consequences of this? is it really a drawback?

<Harold> Fact base 1: p(a->1 b->2) and Fact base 2: p(b->2 a->1) can be easily integrated.

<DougL> -1 (I object also -- but not rabidly -- decades of being bitten by revisions of fixed-order schemes motivate me to prefer named args)

ChrisW: we estimate based on our experience

<Harold> Fact base 1: p(1 2) and Fact base 2: p(2 1) can NOT be easily integrated.

Adrian: e.g. CLIPS

<AdrianP> CLIPS

ChrisW: not clear if this is really relevant for BLD

csma: does that mean that you would object to the resolution, Adrian?

Adrian: no, I won't object

<Harold> But Adrian's use case shows more discussion here would help.

<ChrisW> it is established that you can interchange rules in any case here

csma: more discussion?

Harold: more discussion needed here

<ChrisW> we always try to announce resolutions in the agenda of a telecon

Issue 36 (mapping between prez and XML syntaxes)

csma: we might go fast here
... the issue was raised on Sept 27, after F2F7
... after the discussion on the triangle of syntaxes
... but then we had a poll and it showed that the preferences were for what we do currently

<csma> http://www.w3.org/2007/09/28-rif-irc-ed.html#item08

csma: there is a second resolution later
... structural model diagrams like UML
... and a direct mapping between the presentation and the XML syntax

Harold: we go directly from the presentation syntax to the semantics

ChrisW: issue not closed, but forgot to close it

csma: we agree that we can close it
... proposed resolution next week
... would there be any objections to that?
... no objections

Meta data

csma: the item meta data in the list of TODOs for the 2nd WD BLD
... any part of RIF that couldn't be ignored is not meta data
... the issue is of ignorable statements
... list of such statements
... poll on open/closed list issues
... other issues here?

josb: yes, the issue of the non-ignorable meta data
... e.g. references to external data sets
... they say something about the rule sets

<AdrianP> import statement are not meta data

csma: we have to list this kind of stuff too
...

meta data may be ignorable for a dialect and not ignorable by another dialect

csma: round tripping point of view should be also considered

<josb> right

csma: comments you can forget, but meta data not

Dave: comments are just another piece of meta data
... this for some use cases
... there are not two classes of meta data

csma: this was actually my point too
... what we do not have is a strawman proposal on how to include them in BLD

josb: I volunteer to write one
... within one week

JeffP: any discussion about RIF test cases

csma: probably on the agenda next week
... scribe for next week
... any volunteer to scribe?

<JeffP> I will try

ChrisW: Jeff as scribe next week, if not Adrian?

<AdrianP> I need to check too

<Hassan> +1 on adjourning

- DRAFT -

RIF telecon 22 January 2008

22 Jan 2008

Attendees

Contents