See also: IRC log
<DaveReynolds> Christian: +1 617 761 6200
<AxelPolleres> Zakim ??P63 is me
<AxelPolleres> I can scribe, if hassan just did...
<ChrisW> scribenick: Hassan
<ChrisW> Scribe: HassanAitKaci
<AxelPolleres> ok, also good ;-)
Action review ...
Action 242 ctd
Minutes from last week to approve ... Adrian sent the latest version ...
<csma> the version in the agenda is the last one
Minutes accepted ...
<ChrisW> RESOLVED: accept sept 18 minutes
<csma> PRR was voted for being voted for adoption yesterday
metavoting ... :-)
CSMA: the PRR voted to send the prop to the Architecture board - that will decide this Friday if it is to be adopted or not (he hopes it will be!)
F2F meeting questions?
<csma> can you pls stop issuing new frozen drafts every second minutes?
<csma> October 3, I think
Sandro: making hotel reservation - will send forms . expire october 6th (maybe even 3rd!)
<csma> I confirm October 3
CSMA: deadline is Oct 3 !
<ChrisW> ACTION: sandro to create/update f2f8 wiki page [recorded in http://www.w3.org/2007/09/25-rif-minutes.html#action01]
<rifbot> Created ACTION-344 - Create/update f2f8 wiki page [on Sandro Hawke - due 2007-10-02].
Jos: Sandro pls put up info re: schedules on the wiki ...
Review comments on the BLD drafts - MK making corrections. Need to document what the open issues are and make sure they are on the agenda for the F2F
ChrisW: thinks we have to contention ... ? are there any ... ?
Related to semantics
Jos: yes - I raised the issue
related extensibility of data types
... Not clear what the semantics for (1) symbols and (2) extensibility of data type support
... is . need to discuss at the F2F
Sandro: issue on whether data types should be extensible within the dialect ...
<csma> What does "extensiility of data types in a dialect" mean?
<AxelPolleres> Shall we put an action on collectign the ALL comments for a session in the F2F to be able to effectively go through them?
Jos: Extensibilty of data type, yes, but also restrict some data types ...
<csma> Is tha related to application-specific data types?
Jos: can use many types but can
implement only a part
... should open an issue on this point
ChrisW will schedule an issue on this point
Sandro: argues that data types make up a fixed set in BLD and that's it !
<Harold> From Jos' first email today:
<Harold> 17- section 2.1.2: it is unclear to me why the list of datatypes is
<Harold> fixed. By fixing this list, every implementation needs to support all
<Harold> mentioned data types, and no other data types can be used in
<Harold> meaning-preserving fashion. I propose to make a list of datatypes which
<Harold> need to be supported by every RIF implementation (e.g. xsd:string,
<Harold> xsd:integer), and a list of additional data types which are recommended
<Harold> for use with RIF (e.g. xsd:gYearMonth)
<csma> Would that forbid specific application to import data types?
ChrisW: action on Jos to open two issues...
ChrisW: See URL of the syntax issues ...
Any missing syntax issues?
<ChrisW> ACTION: jdebruij to open 2 issues on datatype extensibility and symbol spaces [recorded in http://www.w3.org/2007/09/25-rif-minutes.html#action02]
<rifbot> Created ACTION-345 - Open 2 issues on datatype extensibility and symbol spaces [on Jos de Bruijn - due 2007-10-02].
DaveReynolds: there are issues on
the XML serialization ...
... also abstract syntax issues being confused ...
<csma> This is on the agenda
ChrisW: Also how many syntaxes do we have ?
DaveReynolds: that - yes - but how do we manage all these syntaxes ...
<csma> Second discussion
Any other point?
Sandro: should we spend some time
on these ? Just to have a feel ?
... What AS do we use ?
ChrisW: any other points on syntax ?
Sandro: there was a mail from MK that we should NOT talk about naming conventions?
MK: yes - I confirm - perhaps on the 2nd day
Sandro: I think it's important: we need to settle the syntax involves making up a convention for symbols. This is crucial!
<josb> +1 with Sandro: some of my comments were related to naming conventions
MK: sure - but that is not the most immediately needed. Also, the naming is not a big issue as opposed to othetr more pressing. If we have time, then ok ...
Sandro: I thinks this is critical path !
MK: there are more important issues ... than naming conventions.
ChrisW: Is there a contention
there on this issue?
... Is there a contention on the naming convention?
<GaryHallmark> I like Sandro's conventions
Sandro: yes - at least one point ...
MK: can do so by email etc...
<Harold> Sandro, Michael said it's less important than other, technical issues, because it's a matter of a 1-to-1 isomorphism.
ChrisW: I agree with MK
Sandro: yes - but we need a consensus and F2F is good for that
MK: There are more crucial issues ... (again!)
ChrisW: take this off line
... any other point on syntax?
ChrisW: any other point is the breakup of the document that has been suggested several times
<AxelPolleres> ok, we can discuss my points on f2f, no prob
ChrisW: Can we discuss this at
... Anyone thinks we should NOT?
Harold: it is just an editorial matter ... we can postpone the few details left by email
Sandro: When do we split the doc?
ChrisW: after the F2F
Sandro: So we should postpone
ChrisW: any other point on the F2F?
MK and ChrisW chatting about breakfast ... :-)
ChrisW: Sandro more on issues on syntax ?
Sandro: ... ?
ChrisW: Sandro asked feedback on using EBNF syntax ...
Sandro: the mapping of
presentation syntax to AS is not clear
... it is more of a style sort of thing ...
... that is related to Harold's proposal
... the new AS has no hand waving and make sense
... I prefer working with objects ... others may prefer othewise (BNF, ...)
Harold: justifies his notation as it supports order-independence ... makes a clear cut separation between the AS and the semantics
DaveReylnods: I like Sandro's
proposal and his notation make sense to me
... We need a linear notation so this one is good
... the current prop with complicated mappings is to complex - need simpler schemes
Harold: there was no formal mapping before (was hidden in implementation); only now we have made it explict
MK: the EBNF is better because it
is easy to map to XML
... also other notations makes examples bulky ...
... the mappings should be straighforward from the EBNF syntax
Jos: there is not real justification for the EBNF syntax (other than that it is a short hand)
<Harold> Jos, The presentation syntax has been very convenient throughout
<Harold> (principles, examples, semantics).
<Harold> It can be seen as a stripe-skipped, infix-operator version of
<Harold> the abstract EBNF syntax, which is a striped, prefix EBNF syntax.
MK: sure - but a shorthand make it easy to manipulate things and make other notations derivable from it. So I agree
Sandro: this is only partially ok
with me ...
... I still do not understand the objections expressed by Harold
... intuitively, what I need as a programmer should be there ...
... not all things are mappable ...
<Harold> Sandro, we would need to *interface* TWO semantics if the Abstract Syntax already requires a formal *semantics* (for unorderedness etc.)
Sandro: as an implementor I need to know that things are ordered or not
<Harold> Sandro, With the right default, ordered="no", logics such as BLD can be easily represented in XML and can be conveniently modified to ordered dialects using an explicit ordered="yes" where required. For example, a BLD <Ruleset> is equivalent to <Ruleset ordered="no">. For a Prolog-like dialect <Ruleset ordered="yes"> can be used.
ChrisW: Sandro: how is this related to ASN/EBNF notation
Harold: explains how he handles order of things ...
Sandro: argues that it is not
... are args of uniterms ordered? Does the syntax reflect that?
MK: = does distinguish between
... can only capture *some* properties - not all ... why choose some and not others ?
Sandro: finds these details (ACI of and or or) important
<Harold> Sandro, There can also be advanced data structure support for semantic axioms other than commutativity, e.g. for associativity.
ChrisW: Sandro are you designing ans API?
Sandro: it is a matter regarding how this should be serialized ...
<csma> But isn't the diffrence, in the case of RIF, that there is a semantics?
<csma> In data formats, do you have a smeantics in the same sesne,
<sandro> (No, dialects have semantics, RIF per se does not.)
ChrisW: OK - let us talk about this at the F2F
<sandro> (oh, okay.)
<Harold> Example: plus(2 5 plus(3 1 4) 6) can be represented in efficient data structures -- with associativity built in -- as plus(2 5 3 1 4 6).
DaveReynolds: MK said that it is a linear notation, but what Sandro is saying is correct
if the notation is non linear
<Harold> DaveR, so far we only have one XML attribute: type
Gary cannot attend F2F
<sandro> DaveReynolds: RDF/XML takes the non-linearity a lot farther, with xml elements also not ordered.
Will dial in ...
ChrisW: review actions ...
action 343 - complete
action 337 - harold?
action 337 - complete
action 336 - complete
<csma> continued (plan to do that today)
action 331 - continued
action 333 - completed
action 333 - awaiting discussion
action 305 - completed
action 305 - awaiting discussion
ChrisW: what the status of the architecture documents?
Sandro: between rough text and
jumbled discussions (???)
... explains a convention of editing it ...
ChrisW: is anyone interested in being an editor?
<csma> Maybe Sandro just starts and other people will jump in?
Loud roar of volunteers ! :-)
<AxelPolleres> I was hesitatin, since I am not sure where it goes at this moment... otherwise I would volunteer!
<csma> Axel just volunteered
Axel: I was just waiting to see
where it is all going ...
... I'll decide by the F2F
ChrisW: will try to convince you
... if we need to make the BLD doc simpler we need to trim things out ... Splitting into two part might make more sense, but need more editors
Axel: I will discuss this with Sandro ...
ChrisW: DavidReynolds pls feel free to participate as well
<DaveReynolds> Dave is balancing extreme work overload with a desire to help :-(
None voiced ...
ChrisW: Let us review the XML
syntax document ...
... Should we attack any now?
Sandro: I have not read the update
Axel: discusses his comments re: RDF ground facts being kept as is or not
<Harold> Axel asks: s p o. (in RDF/XML) =?= s p o. (in RIF/RDF)
Harold: Section 3 about orders was discussed
<sandro> (Axel, I wonder if you can add some pro/con points?)
Harold: Section 4 structure sharing
<AxelPolleres> (Sandr, good point!)
Harold: pointers should be avoided
<sandro> 4. Are there local pointers in the XML serialization?
<Harold> Reasons for Local Pointers:
<Harold> * Allows structure sharing (see below)
<Harold> * Some dialects may have loop structures in their data
<Harold> Reasons against:
<Harold> * Makes serialization and deserialization more complicated
<Harold> * Loop structures can also be represented using symbolic labels
<Harold> * Pointers should be avoided on the level of rule interchange (specifying the "What"), as they can be introduced on the implementation level (realizing the "How")
<Harold> * Reuse of rules is harder when they need to be disentangled from general graph structures, and easier when rules can be picked at the roots of their tree structures
<Harold> * As far as we know, all practical rule languages use syntax that avoids local pointers. It is unclear as to why should an XML syntax radically depart from the established practice of defining syntaxes for rule languages.
<AxelPolleres> I added Section 1.2, basically.
<csma> It would be useful to have examples (fragment of schema and instances) for the alternatives
Harold: using symbols for structure reuse is safer
<Harold> Reasons to mandate structure sharing:
<Harold> * Saves higher-level work on input processing (eg no symbol tables needed for rule variables)
<Harold> * RIF documents will be smaller
<Harold> Reasons to forbid it:
<Harold> * Structure sharing is somewhat complicated and implementation-level (see above)
<Harold> The Equal predicate of RIF BLD can be used as a declarative way to obtain small RIF documents by naming structures once and reusing their names as often as necessary
<csma> I do not understand that
<csma> can you give an example?
<Harold> x = largetree
<Harold> p(a x x)
Sandro: this functionality
belongs in the syntax
... I am picking up a cultural difference between objects-people and tree-people
Harold: clarified issues of references, local constants
<csma> Re issue 7 and 8: isn't that related to the question of data schemas?
<csma> More specifically isn't the serialisation of globally named variables (issue 8) dependant on how we handle externally specified data schemas?
CSMA: issues 7 and 8 depend on how we define our data schema
<Harold> Christian, if you regard http://example.com/books#LeRif as a data schema, then yes.
CSMA: metadata might also be used to disambiguate - how we serialize symbols etc ... Esp. when the data schema is externally specified
ChrisW: I do not understand ...
DavidReynolds: issue is symbol matching between schemas
<csma> but not only
Sandro: expresses concerns about serialization procedure being BLD-specific ...
<Harold> If we are 'BLD and PRD specific' at the same time, then all other dialects could be happily built on that 'Core'.
<sandro> <op><Const type="rif:iri">http://example.com/products#book</Const></op>
ChrisW: ok - we'll dig into all this deeper at the F2F
<csma> +1 to examples
<csma> Do not hesitate to prepare slides with examples
ChrisW: if you argue some point,
pls give concrete examples to back up qualitative claims
<csma> See you in Hawthorne
<AxelPolleres> See you!
<csma> +1 to adjourn
<csma> apparently, I was cut off
<csma> Shall I recall?
<ChrisW> the only questin is regarding dinner
<ChrisW> do we need to meet for dinner tomorrow?
<sandro> Attendees: josb, Dave_Reynolds, StellaMitchell, Hassan_Ait-Kaci, Harold, ChrisW, csma, AxelPolleres, Sandro, Gary_Hallmark, IgorMozetic, Michael_Kifer
This is scribe.perl Revision: 1.128 of Date: 2007/02/23 21:38:13 Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/ Guessing input format: RRSAgent_Text_Format (score 1.00) Succeeded: s/6th/6th (maybe even 3rd!)/ Succeeded: s/Rview/Review/ Succeeded: s/type/types/ Succeeded: s/me/we/ Succeeded: s/thise/these/ Succeeded: s/are/were/ Succeeded: s/ there is no formal mapping right now/ there was no formal mapping before (was hidden in implementation); only now we have made it explict/ Succeeded: s/ Sandro: With/ Sandro, With/ Found ScribeNick: Hassan Found Scribe: HassanAitKaci WARNING: No "Present: ... " found! Possibly Present: Attendees Axel AxelPolleres CSMA ChrisW Christian DaveReylnods DaveReynolds Dave_Reynolds DavidReynolds Example GaryHallmark Gary_Hallmark Guest Harold Hassan_Ait-Kaci IBM IgorMozetic Jos MK MichaelKifer Michael_Kifer NRCC P42 P6 P63 P75 P9 Sandro StellaMitchell aaaa inserted josb rifbot scribenick was xsd You can indicate people for the Present list like this: <dbooth> Present: dbooth jonathan mary <dbooth> Present+ amy Agenda: http://lists.w3.org/Archives/Public/public-rif-wg/2007Sep/0143.html Got date from IRC log name: 25 Sep 2007 Guessing minutes URL: http://www.w3.org/2007/09/25-rif-minutes.html People with action items: jdebruij sandro WARNING: Input appears to use implicit continuation lines. You may need the "-implicitContinuations" option.[End of scribe.perl diagnostic output]