SV_MEETING_TITLE -- 20 Feb 2008

<m_schnei> there was no id shown in IRC after I called in - what to do?

<MichaelSmith> ScribeNick: zhe

accept previous minutes

<Rinke> look good

<pfps_> minutes are OK

PROPOSED: accept previous previous minutes

<IanH> Proposed: accept minutes http://www.w3.org/2007/OWL/wiki/Teleconference.2008.02.13/Minutes

<pfps_> +1

<Rinke> +1

RESOLUTION: accept minutes of http://www.w3.org/2007/OWL/wiki/Teleconference.2008.02.13/Minutes

strange phone call ...

<Carsten> Will the whole telecon fit on that answering machine?

<sandro> I'm trying to have Zakim call me, since I don't seem to be able to enter DTMF on this system. :-(

<sandro> But it didn't ring through here, either. :-(

pending actions review

action 87 completed.

action 88 completed (punning and owl lite)

action 89 completed--Investigate QA group advised recently to have "as few parts as possible"

Due and overdue actions

action 43 unlikely to happen soon

IanH: update due date?
... action 43 Develop scripts to extract test cases from wiki

no action on sandro to update

<pfps_> action 43 - move due date to three weeks from now

action 86 Send proposal for issue-91 ontology property

<m_schnei> we are talking about 91 today anyway

IanH: jjc to update due date
... general dicussions

Fragments and Conformance (Continued)

IanH: can we get a working agreement
... Jim is not here.
... everyone see the email from Alan?

<uli> yes

<pfps_> +1

<Carsten> yes

<MichaelSmith> yes

<bcuencagrau> yes

<m_schnei> yes

yes

<pfps_> http://lists.w3.org/Archives/Public/public-owl-wg/2008Feb/0120.html

ivan will post emails

<ivan> Alan's last email on the topic

IanH: common understanding of fragmetns
... too many fragments is not a good idea
... Alan's original email: OWLPrime, EL++, DL Lite
... for rec track

<Carsten> I like the suggestion

Achille: just sent an email. IBM's position is have limited # of fragments.
... not proposing a new one.

<ivan> Achille's answer email

Achille: need something that suits customer's needs
... EL like seems to be nice
... like OWLPrime rule based approach
... not sure if it will scale though
... like OWLPrime because it is simple,

also care about Alan's 3 point for real scalability

scribe: DL Lite + transitivity
... translate to pure sql queries
... close to Alan's position

jjc: consider tradeoffs
... A Box scaliblity is a good measure
... rationale is limit # of fragments. should have clear commercial interests

<Zakim> pfps_, you wanted to ask about IBM and rec-track status

<bijan> Substitute "community" interest for "commercial"....commerical viability is one measure of community interest

pfps: to HP, whether REC-ness is vital part for way forward?

<pfps_> ... and to IBM

<Achille> yes I think it does

<Achille> matter

IanH: does it really matter for a fragment to be in rec or note

<MichaelSmith> I think it does

jjc: not sure.

Achille: does matter. encourage tool and people to adopt

<bijan> users often ask for Rec things; govt agencies sometimes are required to use rec things (as opposed to non-rec things); ISO/ANSI is even stronger

<bijan> OWL Lite's being rec track *did* help it...people who wouldn't otherwise have used OWL Lite used it because it was a rec.

jjc: a rec should focus on interoperability

<Zakim> jjc, you wanted to answer a bit better

<bijan> I wonder if jjc really meand "Commercial" rather than "production quality" implementation

<alanr> +1 to go forward, but adjust if necessary

IanH: commerical support on DL fragments is tricky
... do we count implementation of DL as implementation for fragments?

<alanr> My threshold is available, correct implementation with correct complexity. Ideal is open source.

<alanr> owlprime scales, but is incomplete

owlprime scales really well. we have tested 1 billion triples

<Carsten> depends on how you define the semantics, I guess

alanr: owlprime complements DL lite

<Achille> Zhe, could you please also send a pointer to paper or a document about your scalability results

alanr: important to have one ties closely to completeness

==> Achille, sure.

<Achille> thanks!

bmotik: requirement on at least 2 interoperable implementations
... not sure if owlprime has 2 (or if they interoperate)
... even entire OWL FULL does not have such implementations

uli: can we postpone rec track decision.

<Achille> +1 for Uli's proposal

uli: rather, make clear what exactly is scalability

<Carsten> We seem to have agreement at the moment; should we really defer and wait until agreement vanishes? :-)

uli: and how incomplete

<MarkusK> +1 to explain "scalable+incomplete" better

uli: be happy to study owlprime for example

IanH: is the decision to postpone all fragments?
... or just the one (or ones) that is not totally clear

<bcuencagrau> There is possibly another prior discussion, namely which fragments will at least be included in a note

<Carsten> doubtlessly, each rec-track fragment will require a lot of work from us!

<bcuencagrau> those are likely candidates to go rec-track

<ivan> +1 to jeremy

need scribe help on jjc's comments

Achille: second Uli's suggestion
... no rush here. Different levels of conformance, completeness, not totally clear

<Zakim> alanr, you wanted to respond to Uli

alanr: comment on Uli's suggestion. concern about timeline
... owlprime fits into two spots: 1) rule based owl. need to specifiy clearly; 2) scalable
... hard to imagine to have a second fragment like this.

<Uli> alanr, I see the motivation and highlevel description, I simply want to see some more precise descriptions

alanr: timeline is a factor to be considered.

Carsten: if we rec track something, we should do a good job
... try to have maximal DL fragment and FULL fragment

<alanr> uli: me too. Worried that time spent deciding takes time from figuring out things.

Carsten: for DL Lite, authors don't know current shape of OWL 1.1

<jjc> +1

<alanr> +1

Carsten: should invest in DL Lite if we want to put it on rec track

<MarkusK> +1 to Carsten: DL Lite before Rec needs to be checked for extensibility with OWL1.1 features

Carsten: OWLPrime, the benefit seems to be scalability

<Uli> +1 with Carsten's aiming at "optimal compromises" re. performance & expressiveness

Carsten: hard to maximize it

IanH: I don't think that WG can do a lot of research
... fragments have to be based on what we know, instead of more research

<alanr> but more power to you, Carsten, if you want to try

bmotik: 1) a strong case for rule bases OWL reasoning.

<alanr> Note that we decided to *remove* dlp from the fragments document eariler.

seems to be DLP anyway

scribe: Achille's requirement should be accormodated by DLP
... believe we should go to maximal subset
... if vendors think it is hard to implement, we should not woory much.

<MichaelSmith> alanr, IIRC, *we* didn't decide to remove dlp - the removal was unilateral and reverted.

ivan: we should worry

<alanr> +1 to ivan. At least for part of the spec

<jjc> +1 to bijan

<Carsten> I agree with Bijan

<alanr> yup

<ivan> +1 to bijan

<Achille> +1 for Bijan

<alanr> grist for owl 1.2

<Carsten> But we should make *some* attempt

<IanH> this is what we are doing now isnt it?

<IanH> with OWL 1.1 I mean

bijan: production quality implementation.

<jjc> +1 to bijan

bijan: more importantly, interoperable implementations.

<alanr> resource bound attempt

<Zakim> jjc, you wanted to suggest Carsten adds new issue to issue list

jjc: agree with bijan totally

<Carsten> DLP??

<bmotik> That's trivial jeremy.

<alanr> I think he mispoke. Dl-lite

jjc: maybe not research, but some quick check
... on DL Lite fragment

<bijan> I don't understand it

<bmotik> Sorry jeremy -- I thought you meant DLP.

IanH: I did not mean not to do any investigation

<Carsten> we basically all seem to agree

IanH: do not want lengthy research

jjc: sanity chek DL Lite in the context of OWL 1.1

IanH: anybody?

<bmotik> What does sanity check mean?

<Carsten> is afraid that this might be true

Uli: don't think a simple sanity on DL Lite is sufficient

<bmotik> +1 to Uli

IanH: it could be that we cannot add anything.

<MarkusK> +1 to Uli, DLP is much easier to extend

<bijan> And we can always ask the DL Lite community to do a sanity check

<jjc> jjc is happy if this idea of sanity check is withdrawn

<bmotik> In fact, DLP already contains the maximal set of OWL 1.1 features that obeys the "no existentials and no disjunctions" principle

IanH: do we think it makes sense now to make decisions on fragments?

<bmotik> +1

<jjc> +1

<ivan> +1

<Carsten> +1

<Achille> +1

<bijan> +1

<bcuencagrau> +1

<MarkusK> +1

<Uli> +1

<pfps_> +1

<MichaelSmith> +1

<m_schnei> +1

<Rinke> +

<Elisa> +1

IanH: make some provisional decision on which fragments

<sandro> Ian: Do we think it makes sense, now, to decide about fragments now? Before asking which fragments.... Is it sensible to make some decision about rec track fragments now?

<alanr> +1

<sandro> Ian: Okay, so which fragments, in Alan's order, should be provisionally rec-track?

<Uli> yes

<sandro> Ian: (btw we need a name other than "OWL Prime", I think!)

<m_schnei> RDFS3.0

<pfps_> -1

<alanr> we should do a second run to give fragments marketable names

<ivan> -1

<bijan> -2

IanH: first one OWLPrime. maybe we want to rename?

<alanr> forget about names right now

<m_schnei> RDFS Deluxe ;-)

<sandro> Ian: (Since OWL Prime is the name of Oracle's product)

<pfps_> +1

<bijan> rOWL

<jjc> +1

<alanr> OWL Pro

<sandro> grOWL?

IanH: don't worry about names for now

<Achille> 0

<Carsten> 0

<sandro> +1

<ivan> +1

<MarkusK> 0

<m_schnei> +1

<alanr> +1

<Uli> 0

<bcuencagrau> 0

<Rinke> 0

<pfps_> 0

IanH: do we think it is a good candiate for rec track

<bmotik> +1

<MichaelSmith> +1

<bijan> 0

<sandro> Ian: Is "OWL Prime" a sensible candidate for a rec track fragment.

<jjc> +1

<Elisa> +1

<pfps_> 0 - needs definition

<alanr> sensible there is more work to do for OWL Prime

IanH: OWLPrime folks need to do more convince work

<sandro> (Um - NO. A 0 doesn't need convincing.)

bijan: need a proposal before make a decision

<jjc> (There were a lot of 0s, that does suggest there is a need for convincing)

<sandro> ian: Me too -- I want to see a worked-out proposal in the owl-prime space.

<alanr> this is straw poll - non binding. So say "ok, as long as ....." is perfect

IanH: want to see a proposal for rule based fragment like OWLPrime before go forward
... EL++

<Achille> +1

<bmotik> +1

<MarkusK> +1

<MichaelSmith> +1 on EL++

<Uli> +1

<pfps_> +1

<bijan> +1

<bcuencagrau> +1

<Carsten> +1 (surprise)

<alanr> +1

<m_schnei> +0

<Rinke> +1

<ivan> +1

<jjc> +0

<Elisa> +1

<Achille> +1

IanH: DL Lite kind of fragment? Scalable Abox space?

<sandro> Ian: DL-Lite (Scalable A-box)

<Achille> +1

<MichaelSmith> +1 on DL-Lite or very similar

<alanr> scalable *complete*

<ivan> 0

<alanr> +1

<jjc> +0

<Carsten> +1

<m_schnei> +0

<MarkusK> +0.5

<bcuencagrau> +1

<bmotik> 0 to DL-lite as such

<Uli> +1

<Elisa> +1

<Rinke> +0

<bijan> +1

<pfps_> +1/2 - the target is not quite defined

<alanr> bit better than owlprim

<Uli> or even HornShiq?

<bmotik> Should we call it OWL Prime/DLP?

<alanr> I think the characterization is *rule based* however it turns out

IanH: what about DLP?

<jjc> -1

<bcuencagrau> DLP, Horn-SHIQ and OWL Prime should ''merge''

<bmotik> Not much.

<alanr> 0 unless it turns out equivalent to OWL Prime

<MarkusK> +1 to bcuencagrau if this should work out ...

ivan: how much additional to merge DLP and OWLPrime

<Carsten> And how comprehensible will our documents be?

<Achille> 0 because I am not sure that DLP or HornSHIQ could be defined in a simple way (i.e. non dl-expert can easily understand the restrictions).

<m_schnei> I would rather spend time for the relationship between OWL-Prime and pD*

<Rinke> wondering about the relation between LP and DLP, and eg RIF... or is that willy?

IanH: yes it involves quite some work.
... it was an important link to make

<Carsten> It can have a corresponding fragment, but do *we* have to make that explicit on rec?

<bijan> It would be *nice*

IanH: who think it is important to have a correspondance between rule based fragment

<Uli> +1 to bijan

<alanr> imo, not important. nice to have.

and DL fragment

<jjc> +0

<pfps_> +x, where x is more than 0 but less than 1

<Zakim> bmotik, you wanted to answer to Ivan

<sandro> ian: Is it important to have a DL fragment which corresponds roughly to whatever we have on the rules side (NOT rec-track) ?

bmotik: to answer ivan, with minor work, we ccan come up with a clear definition when these two fragments coincide
... this is something we really should look into
... don't believe it is a lot of work.
... pD* is not that far away from DLP anyway

bijan: is the correspondence a requirement?

<MichaelSmith> +1 to bijan, this is a goal, not a rqmt

<alanr> for fragments: objective

<bijan> +1

IanH: it is an objective

<jjc> +0

<Rinke> +1

<bcuencagrau> +1

<MarkusK> +1

<MichaelSmith> +1

<alanr> 0

<m_schnei> -0

<Achille> 0

<Elisa> +1

<Uli> +0

<bmotik> +1 to establishing a link between OWL Prime and DLP

<Carsten> -0 (depends very much on where it goes and in which form)

<ivan> +1 (if put it that way)

+1 (sorry, typo)

<bijan> Why should we restrict the future?

<sandro> Ian: Are these mentioned fragments ALL the ones we should consider?

IanH: do we cover all fragments (for rec track candidates)?

<jjc> +1

<sandro> Was OWL Full considered?

<alanr> that's it for now. Barring new information.

<alanr> OWL Full isn't a fragment

<m_schnei> not a DL fragment

<Rinke> and DL?

<sandro> Ian: No, OWL Full wasn't considered but it's not a fragment, so the question doesn't apply.

<pfps_> -0, there could be some reason to add something, but I don't see it now

<jjc> +1

<alanr> Dl not considered a fragment either

<m_schnei> 1.0-DL is a fragment, yes! :)

<Uli> do we need to *rule* others out now?

<Achille> +1 with maybe some minor modifications

<ivan> +1

<sandro> +0.5

<MichaelSmith> +1 to *focusing our energy* on these fragments

<alanr> +1

<m_schnei> +1

<bcuencagrau> +1

<Uli> +1 to Mike!

<pfps_> +1 to michael

<bijan> I think what we have is a starting point, but feel no need to rule stuff out

<MarkusK> +1 to michael

<Rinke> +1 to michael

<Carsten> +1 michael

<bijan> +1 to mike

<Elisa> +1 to Michael

<bmotik> +1 to michael

<ivan> +1 to michael

IanH: time is up. follow up with emails

<alanr> to ask - target date for vote?

IanH: in particular, we need specification for rule based fragment

<sandro> Names for ALL the fragments, please. :-)

IanH: suggestion for a new name :)

<alanr> peter

<alanr> there is loud music here :(

IanH: hand over chair to pfps

<Uli> not here

<IanH> bye

<Uli> bye

pfps: go over issues list

issue 3: two proposals. most current one from bmotik

bmotik: allow bnode in funtional spec
... could use freely just like in rdf

<ivan> Boris' proposal

<bijan> "distinct" not "unique"

bmotik: regarding semantics. could interpret them as unique, or through existential

need help with scribe ...

alanr: isn't this what we had before?
... what is the difference

<alanr> but not wrt anonymous individuals?

<Zakim> jjc, you wanted to articulate some difficulties

<Uli> I understand Boris as saying that the "skolem" proposal is far better than people might have thought...

<pfps_> boris: divergence between DL and Full

jjc, can you please put a summary in IRC?

<pfps_> jjc: syntax bnodes are existentials, object bnodes are skolems?

<scribe> ACTION: jjc send out email clarify [recorded in http://www.w3.org/2008/02/20-owl-minutes.html#action01]

<trackbot-ng> Sorry, couldn't find user - jjc

<MarkusK> +1 to Boris, Skolmes just are a weakened semantic way for DL-tools to work on it

<jjc> ACTION: jeremy to summarise problem with bnodes in ISSUE-3 vs bnodes in OWL list [recorded in http://www.w3.org/2008/02/20-owl-minutes.html#action02]

<trackbot-ng> Created ACTION-90 - Summarise problem with bnodes in ISSUE-3 vs bnodes in OWL list [on Jeremy Carroll - due 2008-02-27].

issue 91

want to defer because of the action item?

<bmotik> Could Jeremy actually summarize whast the problem is/

<alanr> Can michael take either of these?

<alanr> schneider

issue 95

<m_schnei> i will think about this... later :)

issue 95 No compatibility restrictions between the datatype being restricted and the facets in the DatatypeRestriction construct

<MichaelSmith> +1 to boris's suggestion restricting datarange to just datatypeURIs

<bijan> +1

bmotik: DatatypeRestriction takes data types
... change synatx a little bit

<Uli> +1

pfps: looks like this can be resolved next week.

<bmotik> ACTION: bmotik2 to Update the spec to resolve ISSUE-95 [recorded in http://www.w3.org/2008/02/20-owl-minutes.html#action03]

<trackbot-ng> Created ACTION-91 - Update the spec to resolve ISSUE-95 [on Boris Motik - due 2008-02-27].

bmotik: can I update it now and send out an email for the resolution

issue 68

issue 68: mapping rules are non-monotonic

<alanr> :

<alanr> :) good

pfps: I don't quite understand.

<alanr> I also thought it was not being understood

<Carsten> have to leave, sorry. bye

<m_schnei> and for me

<Uli> bye

<m_schnei> and for owl-full

<alanr> problem for monotonicity principle

jjc: is a problem for Jena

<Uli> Jeremy, what is the problem due to?

bijan: a proposal may solve the problem

pfps: this is not about QCR,

<alanr> +1 property punning is the issue

<bmotik> Could Jeremy explain why this is a problem for Jena?

<alanr> a solution that works is extremely acceptable :)

<alanr> write it up?

<alanr> motivation was backwards compatibility, as I understand it

lost m_schnei?

<sandro> m_schnei, try pressing 41#

<sandro> okay try pressing 61#

<bijan> Is the relevant jeremey email linked from the tracker page?

<bijan> This one? http://lists.w3.org/Archives/Public/public-owl-wg/2008Jan/0262.html

pfps: to m_schnei, this is a different issue

<alanr> effectively dropping triples

pfps: QCR is a separate issue

<Zakim> m_schnei, you wanted to clarify what the problem is in owl-full

<pfps_> ack [IPcaller]

jjc: why it is a problem for Jena

<Zakim> jjc, you wanted to try and explain why this is a problem for Jena

jjc: knowledge is stored in triples

<bijan> It must! Or editors can't be based on Jena!

jjc: as you add more knowledge to Jena model, you don't retract triples

<sandro> JJC: the problem here is that as you add triples, some of the old triples are supposed to be changed -- and that's not how Jena wants things to work.

jjc: it is a fundamental design

<alanr> reasoning != editing

bmotik: if Jena is a triple based, do we really need to retract triples?
... I don't see why these mapping rules will require you to retract triples?

jjc: not clear

<bijan> alanr... I don't understand the relevance of that comment

pfps: bijan proposed to do something?

<MichaelSmith> I was going to agree with Bijan, which is what I understand boris to have done as well

<alanr> can answer bijan, if necessary

bijan: has no clue about the problem. not sure which email is relevant.

<alanr> great!

pfps: take an action to write up an email summarizing

<bijan> I also don't understand the assertion that Jena can't delete

pfps: my understanding of this problem

action 16

<bmotik> ACTION: pfps to Write up his understanding of ISSUE-68 (nonmonotonic mapping rules) [recorded in http://www.w3.org/2008/02/20-owl-minutes.html#action04]

<trackbot-ng> Sorry, couldn't find user - pfps

<sandro> trackbot-ng, list users?

<bmotik> ACTION: patelschneider to Write up his understanding of ISSUE-68 (nonmonotonic mapping rules) [recorded in http://www.w3.org/2008/02/20-owl-minutes.html#action05]

<trackbot-ng> Sorry, couldn't find user - patelschneider

<sandro> trackbot-ng, list users

<alanr> in the sense that pellet (used to?) be unable to retract without doing a lot of work

<bijan> TopQuadrent is based on Jena and handles OWL 1.1 (by extending Jena)...isn't this an existence counterproof?

<bmotik> ACTION: peter to Write up his understanding of ISSUE-68 (nonmonotonic mapping rules) [recorded in http://www.w3.org/2008/02/20-owl-minutes.html#action06]

<trackbot-ng> Sorry, amibiguous username (more than one match) - peter

<trackbot-ng> Try using a different identifier, such as family name or username (eg. ppatelsc, phaase)

<alanr> pre: incremental reasoning

<bmotik> ACTION: patel-schneider to Write up his understanding of ISSUE-68 (nonmonotonic mapping rules) [recorded in http://www.w3.org/2008/02/20-owl-minutes.html#action07]

<trackbot-ng> Created ACTION-92 - Write up his understanding of ISSUE-68 (nonmonotonic mapping rules) [on Peter Patel-Schneider - due 2008-02-27].

pfps: defer it

issue 69

issue 69 punning is incompatible with OWL Full

<alanr> how so?

<bijan> And: http://jena.sourceforge.net/javadoc/index.html

<bijan> I meant: http://jena.sourceforge.net/javadoc/com/hp/hpl/jena/rdf/model/Model.html#remove(com.hp.hpl.jena.rdf.model.Statement)

<bmotik> I do

<bijan> How is it that Jena can't delete triples?

<jjc> Jena can delete triples, but not as part of the process of adding other triples

bmotik: it is incompatible with owl full semantics.
... in owl 1.0, DL and FULL are not completely aligned.

<alanr> property punning is the only issue that I see

<pfps_> note for scribe - the IRC messages should be moved into the previous topic

<alanr> cardinality

<bijan> I really don't understand why that does...

<bijan> er...matters

bmotik: don't see it as a showstopper

<bmotik> Alan, this issue is much more than about cardinality.

<Uli> also, since punning makes more ontologies falls into OWL 1.1/DL, we might also lose some more theorems?

<Uli> Alan, can you write this comment into IRC?

bijan: have all sorts of punning in pellet

<pfps_> alanr: property punning causes more problems than other punning

bijan: encountered no problems so far

<jjc> that's a nice point

<Uli> ok, but what was the point about cardinality?

<alanr> I've used pellet a lot. I would be surprised about cardinality issues. ymmv.

bijan: able to handle more graphs is really important
... don't think punning is a really issue

<jjc> zakim aaff -s me

<jjc> zakim aaff is me

<m_schnei> as boris said: there is only one "compatibility" aspect in 1.0, and this will be maintained in 1.1 with punning: 1.1-Full will have every 1.1-DL entailment (for legal 1.1-DL onts)

alanr: change the way owl dl works.

<m_schnei> see sec 5.4 in AS&S

<m_schnei> +1 to close/rejected

<bijan> +1

<bmotik> +1

<MarkusK> +1

pfps: close to our time

<sandro> pfps: I don't know the way forward on this. I leave it to the chairs to make a decision.

<bmotik> I wanted to say something as well about punning.

<pfps_> jjc: bijan's comments about different compatability useful, will consult

thanks peter

<bijan> I'll have to miss that

<bijan> because I'm teaching all day long

<alanr> ok

<bijan> After that I'm free

pfps: defer that UFD meeting.

<Uli> hooray

<bmotik> Bye

<Rinke> ok, thanks

<ivan> bye everyone

<Rinke> bye

<Uli> bye

<Achille> bye

<Elisa> bye

<MarkusK> bye

<sandro> zhe, hold on just a moment and I'll have the wiki minutes ready

- DRAFT -

SV_MEETING_TITLE

20 Feb 2008

Attendees

Contents

accept previous minutes

pending actions review

Due and overdue actions

Fragments and Conformance (Continued)

action 16

issue 69

Summary of Action Items

Scribe.perl diagnostic output