IRC log of owl on 2008-04-04

Timestamps are in UTC.

00:04:20 [Zakim]
Zakim has left #owl
00:07:33 [alanr]
alanr has joined #owl
00:54:11 [alanr]
rrsagent, bookmark
00:54:11 [RRSAgent]
00:54:25 [alanr]
rrsagent, make log public
01:18:43 [IanH]
IanH has joined #owl
01:27:14 [IanH_]
IanH_ has joined #owl
01:41:28 [sandro]
sandro has joined #owl
02:28:10 [Carsten]
Carsten has joined #owl
04:45:01 [alanr_]
alanr_ has joined #owl
05:29:37 [alanr]
alanr has joined #owl
13:08:26 [RRSAgent]
RRSAgent has joined #owl
13:08:26 [RRSAgent]
logging to
13:08:27 [trackbot-ng]
RRSAgent, make logs public
13:08:29 [trackbot-ng]
Zakim, this will be OWLWG
13:08:29 [Zakim]
ok, trackbot-ng; I see SW_OWL(F2F)9:00AM scheduled to start 8 minutes ago
13:08:30 [trackbot-ng]
Meeting: OWL Working Group Teleconference
13:08:30 [trackbot-ng]
Date: 04 April 2008
13:08:38 [sandro]
Zakim, room for 5 for 480 minutes?
13:08:39 [uli]
uli has joined #owl
13:08:40 [Zakim]
ok, sandro; conference Team_(owl)13:08Z scheduled with code 26632 (CONF2) for 480 minutes until 2108Z
13:08:41 [jjc]
DAY 2!
13:08:42 [ekw]
ekw has joined #owl
13:08:53 [baojie]
baojie has joined #owl
13:10:06 [baojie]
baojie has joined #owl
13:10:17 [sandro]
Ian: Thanks again to Clark & Parsia for dinner last night, and to NIST for local organizating.
13:10:19 [kendall]
kendall has joined #owl
13:10:39 [m_schnei]
m_schnei has joined #owl
13:11:25 [thomassch]
thomassch has joined #owl
13:11:29 [bijan]
bijan has joined #owl
13:11:37 [bijan]
scribenick: bijan
13:12:27 [Achille]
Achille has joined #owl
13:12:45 [bijan]
Topic: RDF Mapping
13:13:23 [bijan]
IanH: First up is fixing a bunch of issues by eliminating data/object property punning. Boris has a lil presentation on a proposal.
13:14:59 [Zhe]
Zhe has joined #owl
13:15:29 [bijan]
boris: Two basic roots to many issues in the mapping: 1) object/data/annotation punning and 2) typing occurances of terms (2 gets driven by 1)
13:16:00 [bijan]
alanr: But we require typing in OWL 1.0?
13:16:20 [bijan]
boris: But in 1.0 it's not quite worked out.
13:16:46 [bijan]
[And in 1.0 you can get away with *global* typing whereas with punning you need *local* typing --- bijan]
13:17:04 [bijan]
jjc: Do we need typing for annotations?
13:17:46 [bijan]
boris: yes; and there is object/data punning in annotation properties (even in 1.0)
13:17:48 [bcuencagrau]
bcuencagrau has joined #owl
13:18:49 [bijan]
boris: Propsoal: disallow property punning and require a type for every (property) URI
13:19:37 [MarkusK]
MarkusK has joined #owl
13:19:52 [bijan]
boris: Changes to the specs. Use declarations in the struc spec/functional spec (but map them to type triples) (see slide 3 for details)
13:20:06 [bijan]
alan: Do we type ontology properties as well?
13:20:23 [bijan]
boris: With the current ontology properties there's no need since we have a list of them.
13:20:46 [bijan]
jjc: And that was deliberate...ontology properties were a closed set so you can just look for them.
13:21:07 [bijan]
m_schnei: What about typing vocabulary?
13:21:26 [bijan]
boris: I'm talking about at the level of the structural spec so no type triples
13:21:43 [bijan]
m_schnei: I mean the typing in the quantifers, e.g., objectSomeValuesFrom
13:21:46 [bijan]
boris: getting to it
13:22:37 [bijan]
jjc: What about typing inference?
13:23:50 [bijan]
boris: I deal with that. But i think at the structural spec level we should require declaration. But what your are talking about is "repair" and we should allow that. But the tool should generate the type triples internally.
13:23:57 [bijan]
bijan: are we specing the repair rules?
13:24:06 [bijan]
chairs: We have a later agenda item on it.
13:24:43 [sandro]
boris: Have a separate notion of "repair" for using ontologies which omit declarations.
13:25:09 [ivan]
ivan has joined #owl
13:26:02 [bijan]
boris:UML diagrams stay the same. We augment the grammar for functional style syntax with some annotations indicating the required global type.
13:27:57 [bijan]
...this means that the diagrams match the grammar, but we don't have in the explicit syntax any typed terms but we do have productions that indicate what the type of the construct is (e.g., no typed terminals, but the non-terminals can be typed so there is not a large change from the current spec or apis)
13:29:39 [bijan]
Discussion about whether tools should add types in all files where there are missing types. A lot of discussion about whether to make this recommondation. This behavior is highly undesired for some people.
13:30:10 [sandro]
Sandro: the implication of "ontologies SHOULD put declarations at the top of the file" is "OWL tools SHOULD NOT use triples stores." [[ or maybe not... ]]
13:30:19 [sandro]
Evan: "Suggested Practice"
13:30:22 [bijan]
Discussion is actually about whether to recommende putting htings at the tope of the store.
13:31:11 [bijan]
alan: Suggests adding more context making clear that devianting from this advice is ok, but has costs.
13:31:13 [sandro]
jjc: "declare before use"
13:31:21 [msmith]
+1 to jeremy
13:31:28 [MarkusK]
+1 to jeremy
13:31:42 [sandro]
q+ sort-order of triples in rdf serialzation
13:31:53 [m_schnei]
what does "declare before use" mean in a *set* of axioms?
13:32:13 [bijan]
alan: Declare before use still suggests a per document bias.
13:32:31 [MarkusK]
"declare before use" refers to the serialisation, not to sets of axioms
13:33:08 [bijan]
sandro: This falls into the "if you put triples in a particular order you can do well"
13:33:33 [bijan]
bijan: But we could incorporate this in the "nice serialization" stuff that I've suggested we offer earlier.
13:33:47 [sandro]
wide acclaim to bijan there.
13:34:28 [bijan]
boris: Changes to RDF mapping; this fixes tons of problems, e.g., non-mon gone, remove duplicate vocabualry etc.
13:35:02 [bijan]
boris: We should make clear that repair isn't part of the spec....type validation is based on the explicit set of triples
13:35:15 [sandro]
boris: (and agreed by JJC loudly) the doc should be clear you don't do RDFS reasoning around declarations
13:35:35 [bijan]
boris: "declaredAs" will go away in RDF
13:36:09 [bijan]
boirs: Given cyclic imports and typing triples occuring anyway, the parsing algorithm has to be two pass.
13:36:17 [bijan]
13:36:31 [sandro]
boris: parsing has to be two pass --- (1) imports closure, (2) look for declarations, etc. The document should be more clear about this.
13:36:32 [bijan]
jjc: While we can express this, we shouldn't say it is *required*
13:36:47 [sandro]
JJC: "Streaming OWL-DL" by me --- you can do it in one pass.
13:36:56 [bijan]
boris: yes, but I think *speccing* it as two pass will be clearer...obvious people can optimize better.
13:37:10 [sandro]
JJC: so the spec should no say you MUST do it in two pass.
13:37:24 [sandro]
alan: "Notes to Implementors"
13:37:30 [bijan]
alan: This is another notes to implementors. We should have an additional section
13:37:53 [sandro]
ian: Do we need an Implementor-Facing-Document Task Force? :-) :-) [[laughter]]
13:38:24 [bijan]
boris: importing files of different syntaxes?
13:38:50 [bijan]
...But now we can formulate things at the object model so mixed imports (and typing) becomes clean.
13:39:10 [alanr]
alanr has joined #owl
13:39:24 [bijan]
boris: Changes to XML Syntax: Drop typed quantifiers as in the functional syntax. Regularizes all the syntaxes and makes the parsing (at a high level) exactly the same.
13:39:25 [sandro]
boris: by having declarations at the syntactic-object level, we are clear about how to handle import-closures which include OWL serialized in different syntaxes.
13:39:34 [sandro]
(lots of nods)
13:39:43 [bijan]
jjc: secondes the proposal
13:40:22 [bijan]
m_schnei: Can we defer some other issues?
13:40:32 [bijan]
Generally people wanted to do things on a case by case basis.
13:41:07 [bijan]
sandro: We can make decisions and if we find problems later then we can reopen the decision.
13:42:00 [bijan]
Topic: ISSUE-17
13:42:42 [bmotik]
bmotik has joined #owl
13:42:50 [bijan]
IanH: This is a non-issue if we accept the boris proposal.
13:43:09 [bijan]
...Thus this issue is resolved by this proposal.
13:43:53 [bijan]
jjc: One use case from the "Full/RDF" domain is dc:creator, but I don't believe it is address by punning so I don't think we harm that use case *more*.
13:44:38 [bijan]
pfps: I'd like to mark that we don't agree with the lack of use cases.
13:45:14 [bijan]
IanH: So we'll say that we technically don't know how to do it so we closed it.
13:46:22 [msmith]
to alan, the dc use case is mentioned briefly at
13:46:59 [bijan]
ekw: about boris's proposal: The only place the proliferation of typed vocabulary occurs is in the struc spec?
13:47:17 [bijan]
boris: Yes.
13:47:31 [bijan]
IanH: But only in non-terminals. Terminals are all type free.
13:47:55 [bijan]
bijan: But that could impact the OMG UML mapping
13:48:02 [Carsten]
Carsten has joined #owl
13:48:02 [alanr]
mike: search for creator on that page finds nothing
13:48:09 [bijan]
boris's proposal:$6d8508f0$2a12220a@wolf
13:48:15 [sandro]
PROPOSED: Close ISSUE-17 as resolved, saying we forbid any property punning, as per Boris' proposal ($6d8508f0$2a12220a@wolf) with a note that the reason is that technically we don't know how to do it.
13:48:29 [sandro]
13:48:31 [bijan]
13:48:32 [Rinke]
13:48:34 [bmotik]
+1 (Oxford)
13:48:35 [MarkusK]
13:48:35 [bcuencagrau]
13:48:36 [Achille]
13:48:37 [baojie]
13:48:38 [uli]
13:48:38 [Zhe]
13:48:39 [msmith]
13:48:40 [IanH]
+1 (Oxford)
13:48:42 [ekw]
+1 (NIST)
13:48:43 [alanr]
+1 (Science Commons)
13:48:47 [jjc]
+1 (HP)
13:49:48 [pfps]
+1 (ALU)
13:50:14 [sandro]
NOTE that Boris doesn't mention AnnotationProperties here, but they are understood as being parallel to DataProperties and ObjectProperties
13:51:05 [sandro]
RESOLVED: Close ISSUE-17 as resolved, saying we forbid any property punning, as per Boris' proposal ($6d8508f0$2a12220a@wolf) with a note that the reason is that technically we don't know how to do it. NOTE that Boris doesn't mention AnnotationProperties here, but they are understood as being parallel to DataProperties and ObjectProperties
13:51:29 [msmith]
alanr, search for dublin core
13:51:30 [JLUCIANO]
JLUCIANO has joined #owl
13:51:31 [bijan]
Topic: ISSUE-18
13:52:18 [bijan]
jjc: I'm not sure who did ISSUE-18 though my name is on it.
13:52:30 [bijan]
...So I withdraw it.
13:52:52 [sandro]
ISSUE-18 is closed, JJC doesn't support it any more, no one else does either.
13:52:56 [bijan]
Topic: ISSUE-65
13:53:17 [bijan]
IanH: The new proposal addresses this explicitly.
13:54:25 [sandro]
Bijan: let's just merge ISSUE-65 into that previous resolution.
13:55:39 [bijan]
jjc: I'm happy as long as after the edits the terms are actually fully reduced.
13:55:43 [sandro]
JJC: I expect that issue-65 will be addressed by Boris' proposal
13:56:08 [bijan]
IanH: We can always revisit if things turn out unexpectedly.
13:56:18 [bijan]
13:56:46 [bijan]
jjc: I believe the proposal addresses this, but obviously I need to check.
13:57:44 [bijan]
Topic: Issue-89 and Issue-19
13:57:52 [bijan]
IanH: Proposal fixes this.
13:59:17 [bijan]
m_schnei: There is an issue. It's not clear but in OWL 1 some type triples are entailments and now if they have no semantics we have a backward compatibility issue.
13:59:23 [bijan]
boris: it's not about entailment.
13:59:55 [sandro]
Matthew: Are declarations still regarded as Axioms?
13:59:58 [sandro]
Boris: Yes
14:00:00 [bijan]
jjc: The entailment still holds.
14:00:30 [bijan]
alan: There was some old wording saying that declarations don't have semantics.
14:00:42 [bijan]
boris: and it *doesn't* have semantics.
14:02:39 [bijan]
alan: Simple proposal: let's remove the wording (about no semantics) but leave things as it is
14:03:47 [bijan]
markus: You cannot ask that something a subclass of thing you have to priorly said that it's a class, so in some sense it changes the semantics but not in the ordinary sense we mean when dropping an axiom.
14:04:16 [bijan]
pfps: Where do we *say* in the document that declarations have no semantics.
14:04:20 [bijan]
...No where.
14:04:40 [bijan]
IanH: it all works out. yay.
14:05:43 [bijan]
alan (strikes back): Finds some text that seems to be problematic.
14:05:50 [bijan]
pfps and jjc: that seems false
14:06:10 [bijan]
IanH: Ok, that's a separate issue and seems largely editorial.
14:06:22 [bijan]
ekw: How do we handle this editorial/whatever issue.
14:06:52 [sandro]
PROPOSED: Close ISSUE-65, ISSUE-68, ISSUE-89, and ISSUE-19 as resolved, as per Boris' proposal ($6d8508f0$2a12220a@wolf), amended to include AnnotationProperties in parallel to DataProperties and ObjectProperties.
14:08:22 [bijan]
m_schnei: What is the status of the semantics of declarations? I
14:08:29 [bijan]
...'m still confused.
14:10:46 [sandro]
m_schnei: about the mapping of c rdf:type owl:Class ... it seems like there is a semantic difference here
14:11:03 [sandro]
bijan: I think this is a change to the presentation of the semantics, but not to their meaning.
14:11:29 [sandro]
PROPOSED: Close ISSUE-65, ISSUE-68, ISSUE-89, and ISSUE-19 as resolved, as per Boris' proposal ($6d8508f0$2a12220a@wolf), amended to include AnnotationProperties in parallel to DataProperties and ObjectProperties.
14:11:46 [bmotik]
+1 (Oxford)
14:11:48 [Achille]
14:11:48 [jjc]
+1 (HP)
14:11:49 [MarkusK]
+1 (FZI)
14:11:52 [bijan]
+1 (Manchester)
14:11:53 [pfps]
+1 ALU
14:11:54 [Rinke]
+1 (Amsterdam)
14:11:55 [alanr]
+1 (Science Commons)
14:11:57 [IanH]
+1 (Oxford)
14:11:57 [sandro]
+1 (W3C)
14:11:57 [baojie]
14:11:58 [Zhe]
+1 (Oracle)
14:12:01 [bcuencagrau]
14:12:06 [ekw]
+1 (NIST)
14:12:12 [msmith]
+1 (C&P)
14:12:55 [sandro]
RESOLVED: Close ISSUE-65, ISSUE-68, ISSUE-89, and ISSUE-19 as resolved, as per Boris' proposal ($6d8508f0$2a12220a@wolf), amended to include AnnotationProperties in parallel to DataProperties and ObjectProperties.
14:13:46 [bijan]
Topic: Issue-66 & Issue-94
14:13:52 [kendall]
I demand someone dial-in so that Mike and I didn't lug the conference phone up here for no reason; to this end, I shall dial-in if I have to! :>
14:14:10 [sandro]
14:14:38 [sandro]
Think of all the poor reserved Zakim ports, sighing miserably because they have no purpose in life.
14:15:07 [bijan]
IanH: In the RDF Mapping boris added the following text: """More precisely, let O be any OWL 2 DL ontology in functional-style syntax, let RDF(O) the set of RDF triples obtained by transforming O into RDF triples as specified in Section 2, and let O' be the OWL 2 DL ontology in functional-style syntax obtained by applying the reverse-transformation from Section 3 to RDF(O); then, O and O' are logically equivalent (i.e., they have exactly the same set of models)."
14:15:07 [bijan]
14:15:35 [bijan]
IanH: This resolves the issue.
14:15:45 [bijan]
jjc: I accept it resolves the issue, but it's editorially nasty.
14:16:46 [bijan]
sandro: Let me understand: Roundtripping gives you the same models, but not necessarily the same syntactic mapping. It would be nice to mention this.
14:17:57 [bijan]
PROPOSAL: Resolve issue-19 by adding to the RDF Mapping the following text: "More precisely, let O be any OWL 2 DL ontology in functional-style syntax, let RDF(O) the set of RDF triples obtained by transforming O into RDF triples as specified in Section 2, and let O' be the OWL 2 DL ontology in functional-style syntax obtained by applying the reverse-transformation from Section 3 to RDF(O); then, O and O' are logically equivalent (i.e., they have exactly the same
14:17:57 [bijan]
set of models)."
14:18:30 [bijan]
Achille: We've asserted this, but what's the status of proving it? Do we require a formal proof?
14:18:50 [bijan]
IanH: Formal proofs for mapping seem a bit different than formal proofs for semantics etc.
14:19:22 [bijan]
jjc: This resolution resolves things by making any falsity in the roundtripping a bug not a feature. How we detect and fix such bugs is a different issues.
14:20:13 [bmotik]
+1 (Oxford)
14:20:14 [bijan]
IanH: The semantics equivalence claim in the profiles document has a different status
14:20:15 [Achille]
+1 (IBM)
14:20:16 [alanr]
+1 (Science Commons)
14:20:17 [pfps]
+1 ALU
14:20:20 [bijan]
+1 (Manchester)
14:20:21 [jjc]
+1 (HP)
14:20:22 [Zhe]
+1 (Oracle)
14:20:22 [baojie]
+1 (RPI)
14:20:22 [msmith]
+1 (C&P)
14:20:23 [Rinke]
+1 (Amsterdam)
14:20:25 [ekw]
+1 (NIST)
14:20:25 [MarkusK]
+1 (FZI)
14:20:26 [IanH]
+1 (Oxford)
14:20:28 [uli]
14:20:30 [bcuencagrau]
14:20:33 [jluciano]
+1 (MITRE)
14:21:00 [sandro]
+1 (W3C) understanding that various word-smithing may happen later, and it's really ISSUE-66.
14:21:05 [bijan]
14:21:43 [bijan]
Topic: Issue-86
14:22:28 [bijan]
IanH: MikeS has a solution to mapping inverse property expressions: <>
14:23:10 [bijan]
msmith: Instead of trying to use a bnode in property position, reverse the arguments.
14:24:01 [bijan]
m_schnei: I have a solution and it's a more general problem. Inverse expressions are a symptom.
14:24:31 [msmith]
m_schnei email
14:24:33 [bijan]
...The problem is more general...can we tackle the more general problem
14:24:44 [bijan]
boris: we should fix the concrete case.
14:25:15 [sandro]
m_schnei: This solution is a hack. Let's solve it properly, so that other s-p-o things are serialized in RDF consistently.
14:25:25 [sandro]
bmotik: yes, it's a hack, but that's fine.
14:25:48 [m_schnei]
14:25:50 [sandro]
bijan: Also, RDF may change. Let's not go to much effort to work around this (perhaps temporary) shortcoming of RDF.
14:26:44 [sandro]
pfps: msmith's solution is to make sure that on the RDF side we have no b-node there. m_schnei's solution is if it's not a property name, then translate it to an object restriction.
14:27:14 [bijan]
jjc: One problem with hasValue restrictions is that it uses a very expressive construct for a less expressive bit.
14:27:17 [sandro]
jjc: introducting hasvalue restrictions is a problem, because as Uli pointed out yesterday, it confuses the expressivity of the ontology.
14:27:39 [bijan]
IanH: And hits roundtripping hard.
14:28:07 [bijan]
Boris: It's a simple simple solution that fits in with the rdf etc. etc.
14:28:10 [sandro]
bmotik: msmith's solution is so easy. it fits well with rdf. it doesn't introduce nominals. (Uli: and it keeps the graph beautiful) (Ian: It keeps the data part as data)
14:28:59 [sandro]
m_schnei: Is it CERTAIN that this is the only case where this situation occurs?
14:29:12 [sandro]
Ian: YES. If you discover otherwise, we will revisit this issue.
14:29:34 [bijan]
m_schnei: My problem is that I dismissed it so, I didn't understand it.
14:29:55 [bijan]
markus: If we encounter new problems, we'll have to revisit the hack anyway.
14:30:21 [sandro]
MarkusK: If some other occurance of this situation occurs, we'll have to come up with something. we'll have a new issue, and we can apply that solution back to this, if necessary.
14:30:54 [sandro]
Uli: This is not a hack. It's standard re-writing for conjunctive queries. There's nothing going wrong here, nothing weird here.
14:30:55 [bijan]
Uli: It's not a's a standard technique, used in rewriting of conjunctive query, etc.
14:31:00 [msmith]
this was proposal #2 in
14:47:40 [thomassch]
thomassch has joined #owl
15:05:30 [baojie]
baojie has joined #owl
15:06:40 [sandro]
scribenick: Achille
15:06:55 [IanH]
Proposed: resolve issue-86 as proposed in; it is believed that this is the only place that this problem arises.
15:07:47 [IanH]
Proposed: resolve issue-86 as per proposal 2 in; it is believed that this is the only place that this problem arises.
15:07:59 [msmith]
+1 (C&P)
15:08:01 [pfps]
+1 ALU
15:08:02 [bijan]
+1 (Manchester)
15:08:03 [jjc]
+1 (HP)
15:08:04 [Rinke]
+1 (Amsterdam)
15:08:04 [MarkusK]
+1 (FZI)
15:08:04 [alanr]
+1 (Science Commons)
15:08:05 [baojie]
+1 (RPI)
15:08:06 [bmotik]
+1 (Oxford)
15:08:06 [Achille]
+1 (IBM)
15:08:06 [Zhe]
+1 (Oracle)
15:08:07 [IanH]
+1 (Oxford)
15:08:08 [bcuencagrau]
15:08:09 [uli]
15:08:12 [ekw]
+1 (NIST)
15:08:18 [sandro]
+0 swapped out
15:08:24 [evrensirin]
evrensirin has joined #owl
15:08:27 [jluciano]
+1 (MITRE)
15:08:46 [IanH]
RESOLVED: resolve issue-86 as per proposal 2 in; it is believed that this is the only place that this problem arises.
15:09:19 [Achille]
topic: Issue 94: Issues surrounding roundtripping
15:10:11 [sandro]
issue 94 rejected / withdrawn.
15:10:21 [Achille]
alan: let's reject it
15:10:26 [sandro]
(as per boris and ian and alan)
15:11:01 [Achille]
bijan: it might be nice to regularize it
15:11:11 [sandro]
Bijan: It would be nice to regularize this
15:11:53 [sandro]
peter: the only n-ary construct in OWL 1 was allDifferent
15:12:46 [sandro]
bijan: so now we've discovered that there are lots of these....
15:12:49 [Achille]
alan: do we want a resolution
15:13:08 [Achille]
bijan: if we reject it I will hjave another one with disjoint
15:13:40 [sandro]
? PROPOSED: Resolve ISSUE-94, introducing disjointProperties
15:15:01 [jjc]
Jeremy: Disjoint properties is N^2 without the special class AllDisjointProperties, this is one item to avoid an N^2 rule, so it is probably worth it.
15:15:02 [Achille]
m_scheni: n-ary construct map to a lot of binary statement in rdf
15:15:33 [Achille]
... this is not very nice, It will address the problem of annotating n-ary constructs
15:15:46 [MarkusK]
s /m_scheni/m_schnei/
15:16:16 [Achille]
jjc: we are not solving the annotation issues in this metting
15:16:30 [sandro]
jjc: reification-annotations got parked for furthur reflection. I suggest we re-open something relative to annotations.
15:16:45 [sandro]
ian: close should say what we add, but not what we rule out.
15:16:59 [Achille]
jjc: we do not rule out them for now
15:16:59 [sandro]
jjc: yes, open world assumption :-)
15:18:15 [Achille]
ian: add it for disjointProperty and close this issue
15:18:19 [sandro]
PROPOSED: Resolve ISSUE-94, introducing disjointProperties (and note that disjointClasses is already in OWL2), handled like allDifferentFrom was in OWL 1
15:18:31 [m_schnei]
sandro, you can also have a look in my Full proposal <>
15:18:48 [pfps]
+0 ALU
15:18:50 [alanr]
+1 (Science Commons)
15:18:51 [jjc]
+1 (HP)
15:18:52 [Achille]
+1 (IBM)
15:18:53 [sandro]
+1 (W3C)
15:18:53 [bijan]
+1 (Manchester)
15:19:01 [ekw]
+1 (NIST)
15:19:01 [IanH]
+1 (Oxford)
15:19:03 [msmith]
+1 (C&P)
15:19:04 [bcuencagrau]
+1 (oxford univ)
15:19:07 [jluciano]
+1 (MITRE)
15:19:14 [MarkusK]
+1 (FZI)
15:19:14 [Rinke]
+1 (Amsterdam)
15:19:26 [Zhe]
+1 (Oracle)
15:19:36 [Achille]
peter: we are adding more vocabulary
15:19:46 [sandro]
m_schnei, I hope to look at it at some point, yeah.
15:20:00 [sandro]
RESOLVED: Resolve ISSUE-94, introducing disjointProperties (and note that disjointClasses is already in OWL2), handled like allDifferentFrom was in OWL 1
15:20:17 [Achille]
topic: Issue 96 OWL-1.1 vocabulary naming in RDF mapping is not consistent
15:20:33 [Achille]
ian: inconsistency in the vocabulary naming
15:21:57 [Achille]
m_schnei: disjointObjectProperties instead of disjointObjectPropertyWith
15:23:37 [sandro]
maybe "disjointWithProperty" ?
15:24:05 [Achille]
alan: disjointObjectPropertiesWith is not good english
15:24:25 [sandro]
m_schnei: the two-class case has a special serialization on RDF, so it should for property as well.
15:24:25 [Achille]
alan: let's change then all to make it better english
15:24:56 [Achille]
... all = (disjointClasses and disjointProperties)
15:25:39 [Achille]
evren: disjointProperties as we have equivalentProperties
15:26:09 [sandro]
Bijan: the problem is that "disjointWith" was too grabby.
15:26:22 [Achille]
even: disjointProperty as we have equivalentProperty
15:27:03 [sandro]
how about: p PropertyDisjointWith q
15:27:13 [Achille]
m_schnei: onely one case with singular
15:27:42 [sandro]
JJC: we could optionally (as per alan) just not have the binary special case.
15:27:48 [Achille]
jjc: a note to signal a potential alternative design
15:28:42 [Achille]
jjc: get rid of the binary case
15:29:48 [sandro]
PROPOSED: we do not have special vocabulary for pair-of-disjoint properties, but we include an editor's note asking for feedback from OWL Full community about whether they want PropertyDisjointWith or whatever.
15:30:31 [Achille]
m_schnei: we have case where we have binary only, n-ary and binary, and n-ary only
15:31:00 [Achille]
bijan: this is not a major issue (i.e. the lack of regularity in the language)
15:31:17 [Achille]
m_schnei: in this case I would like to reject the issue
15:31:22 [sandro]
? PROPOSED: resolve ISSUE-96, moving forward without a special vocabulary for pair-of-disjoint properties, but we include an editor's note asking for feedback from OWL Full community about whether they want PropertyDisjointWith or whatever.
15:32:00 [sandro]
m_schnei: i would rather have the vocabulary as is than the term gone.
15:32:56 [sandro]
? PROPOSED: resolve ISSUE-96, using owl:propertyDisjointWith for the binary case of disjoint properties
15:33:09 [jluciano]
+1 (MITRE)
15:33:21 [sandro]
PROPOSED: resolve ISSUE-96, using owl:propertyDisjointWith for the binary case of disjoint properties
15:33:23 [ekw]
0 (NIST)
15:33:24 [alanr]
+1 (Science Commons)
15:33:27 [MarkusK]
+1 (FZI)
15:33:27 [baojie]
0 (RPI)
15:33:28 [Rinke]
+1 (Amsterdam)
15:33:29 [jluciano]
+1 (MITRE)
15:33:31 [bijan]
+0 (Manchester)
15:33:33 [Achille]
0 (IBM)
15:33:33 [Achille]
0 (IBM)
15:33:35 [sandro]
+0 (W3C)
15:33:35 [bcuencagrau]
15:33:36 [jjc]
+0.5 (HP)
15:33:38 [Zhe]
0 (Oracle)
15:33:44 [pfps]
15:33:48 [IanH]
+1 (Oxford)
15:34:10 [msmith]
15:34:16 [sandro]
RESOLVED: resolve ISSUE-96, using owl:propertyDisjointWith for the binary case of disjoint properties
15:34:56 [sandro]
ian: That's all the RDF Mapping issues!
15:35:32 [sandro]
Bijan: Jeremy, do you think this will make HP happier with OWL 2? HP had concerns about the mapping in the OWL 1.1 submission.
15:35:53 [Achille]
jjc: i will report to hp that many of the key issues have been addressed
15:36:03 [Achille]
topic: Issue 3 & Issue 46 anonymous individuals / Unnamed Individual Restrictions
15:36:31 [Achille]
boris: allow bnodes in the structural spec
15:36:44 [Achille]
boris: interpret them as anonymous individuals
15:37:04 [sandro]
bmotik: expand structural spec to add bnode-flag on individuals. also, DL will require that assertions are tree-like.
15:37:10 [Achille]
boris: mandate that the assertion are tree like to gurantee decidability
15:37:25 [sandro]
bijan: this gets us back to the OWL 1 status quo.
15:37:51 [sandro]
bijan: This relates to the skolemize-bnodes case.
15:38:29 [sandro]
Ian: this is intended to close these issues. you're free to raise an issue re: skolem.
15:38:40 [Achille]
jjc: the experience is that owl implementaters have interpreted them as skolem
15:38:50 [sandro]
bijan: skolem would allow non-tree-graphs.
15:39:32 [bmotik]
I've just implemented the resolution of two issues w.r.t. property mapping: I've renamed disjointObjectProperties into propertyDisjointWith and have added the n-ary variant for disjoint object properties
15:39:54 [Achille]
bijan: by this change we will make all owl tools non-conformant
15:40:16 [sandro]
Bijan: This makes all implementations non-conformant.
15:40:17 [Achille]
bijan: i do not see it as a good idea
15:40:31 [Achille]
bijan: i would prefer not having them at all
15:40:48 [sandro]
BijanL ... I would rather NOT HAVE BNodes, if they are going to have existential semantics.
15:40:53 [Achille]
evren: + 1 for bijan
15:41:10 [Achille]
.. but I do not think it makes all tools non-conformant
15:41:22 [Achille]
... I am talking about OWL 1.0
15:41:52 [Achille]
jjc: I agree with evren. OWL 1.0 test cases are about consistency check
15:41:55 [sandro]
JJC: the OWL1 entailment tests did not *TECHNICALLY* affect conformance.
15:42:14 [sandro]
JJC; (even though they were "Normative")
15:42:15 [bmotik]
ACTION: bmotik2 to Modify entire spec according to the proposal from
15:42:15 [trackbot-ng]
Created ACTION-131 - Modify entire spec according to the proposal from [on Boris Motik - due 2008-04-11].
15:43:20 [sandro]
Alan: this proposal differs from OWL 1 in that the tree-like requirement is no longer enforced by the semantics.
15:43:27 [Achille]
bijan: I would liek to break the backward compatibility in the spec
15:44:20 [sandro]
m_schnei: Cyclic property assertion _:x p _:x
15:44:24 [Achille]
m_schnei: cyclic property assertions ?
15:44:42 [jjc]
this is not treee shaped
15:44:43 [Achille]
ian: it would be illegal as not being tree shape
15:45:02 [sandro]
ian: This proposal would rule that out. But it's also ruled out in OWL 1 because it's not in the DL syntax.
15:46:22 [Achille]
ian: if we resolve this according to boris's proposal, it is up to bijan to introduce the skolem semantics
15:47:01 [Achille]
jjc: tree possibilities: 1) no anonymous individuals, 2) individuals as skolem
15:47:11 [Achille]
3) as per boris's proposal
15:47:14 [sandro]
STRAWPOLL: (1) no anon individuals (2) anon indivs as per Boris (3) anon indiv with skolem semantics
15:47:36 [sandro]
Uli: option 4 -- something else?
15:47:40 [jjc]
prefer (2), oppose (1) and (3)
15:48:03 [sandro]
Alan: we could have ADDITIONAL skolem semantics.
15:48:15 [ekw]
prefer (3)
15:48:28 [sandro]
Bijan: that doesn't meet my requirements of having a good enterpretation of EXISTING graphs.
15:48:38 [sandro]
Ian: separate issue.
15:48:56 [sandro]
15:49:13 [jjc]
15:49:16 [bijan]
15:49:19 [m_schnei]
-0.5 to option 1
15:49:26 [Zhe]
15:49:26 [pfps]
0 or option 1
15:49:27 [sandro]
15:49:36 [MarkusK]
15:49:38 [jluciano]
15:49:38 [alanr]
First vote on (1)
15:49:39 [ekw]
15:49:39 [Rinke]
-0 for option 1
15:49:43 [Achille]
15:49:44 [evrensirin]
15:49:48 [alanr]
-1 for (1)
15:49:50 [IanH]
15:49:51 [bcuencagrau]
15:49:56 [uli]
15:50:09 [sandro]
OPTION-2: (2) anon indivs as per Boris
15:50:13 [jjc]
15:50:14 [bijan]
15:50:18 [bcuencagrau]
15:50:19 [baojie]
15:50:20 [alanr]
15:50:26 [Rinke]
15:50:27 [ekw]
15:50:27 [evrensirin]
15:50:29 [pfps]
15:50:30 [IanH]
15:50:32 [sandro]
0 confused
15:50:36 [Achille]
15:50:37 [m_schnei]
-0.25 to 2
15:50:43 [Zhe]
15:50:50 [MarkusK]
15:50:51 [uli]
15:50:56 [jluciano]
15:51:01 [sandro]
OPTION-3: (3) anon indiv with skolem semantics
15:51:23 [bijan]
15:51:23 [evrensirin]
15:51:25 [alanr]
15:51:28 [MarkusK]
15:51:28 [jjc]
15:51:29 [sandro]
15:51:30 [pfps]
15:51:31 [Achille]
15:51:31 [Rinke]
15:51:32 [ekw]
15:51:34 [m_schnei]
+1 for 3 (skolems)
15:51:43 [jluciano]
15:51:44 [Zhe]
15:51:54 [uli]
15:51:56 [IanH]
15:51:58 [bcuencagrau]
15:52:43 [sandro]
jjc: I might be able to go back to HP and get a different answer.
15:53:00 [Achille]
m_schnei: i am interested in an example where the existential semantics is used
15:53:07 [Achille]
bijan: it is never used
15:53:15 [sandro]
Bijan: The only place existential semantics are used is in separate operations.
15:53:33 [Achille]
m_schnei: the statement was for DL
15:54:21 [sandro]
Bijan: This is a request on Jeremy to present some evidence....
15:54:25 [Achille]
m_schnei: could jeremy provide some evidence of this use in DL ?
15:55:00 [Achille]
sandro: we want to add in the draft that we understand that it is non consensual
15:55:21 [sandro]
Sandro: If we make this change, we defiitely want to highlight it, saying we're looking for evidence if it causes any real problem.
15:56:08 [sandro]
? PROPOSED: We go with Boris's proposal on the syntax of bnodes, and open a new issue on the semantics of bnodes.
15:56:22 [Achille]
evren: how about give the skolem semantics for DL and existential for Full
15:56:22 [sandro]
Evrin: maybe differeent semantics in DL and Full.
15:57:02 [Achille]
ian: let's not discuss evren's proposal now
15:57:58 [Achille]
bijan: i will write a proposal so that we can use it as the starting point for further discussion
15:58:40 [alanr]
Action: Bijan to put consolidated list of OWL visible differences between bnode semantics as existential versus skolem
15:58:40 [trackbot-ng]
Created ACTION-132 - Put consolidated list of OWL visible differences between bnode semantics as existential versus skolem [on Bijan Parsia - due 2008-04-11].
15:58:49 [bmotik]
The e-mail with the proposal:
15:59:31 [sandro]
PROPOSED: Resolve ISSUE-46, accepting Boris's proposal ( only in terms of the syntax of bnodes, and open a new issue on the semantics of bnodes, including editorial note in owl2-syntax about semantics being still open.
15:59:54 [sandro]
PROPOSED: Resolve ISSUE-3 and ISSUE-46, accepting Boris's proposal ( only in terms of the syntax of bnodes, and open a new issue on the semantics of bnodes, including editorial note in owl2-syntax about semantics being still open.
16:00:10 [Achille]
+1 (IBM)
16:00:12 [bmotik]
+1 (Oxford)
16:00:12 [IanH]
+1 (Oxford)
16:00:14 [Zhe]
+1 (Oracle)
16:00:15 [bcuencagrau]
+1 (oxford)
16:00:16 [alanr]
+1 (Science Commons)
16:00:17 [baojie]
+1 (RPI)
16:00:18 [sandro]
+1 (W3C)
16:00:22 [MarkusK]
+1 (FZI)
16:00:41 [Rinke]
+1 (Amsterdam)
16:00:52 [msmith]
+1 (C&P)
16:00:53 [ekw]
+1 (NIST)
16:01:17 [sandro]
RESOLVED: Resolve ISSUE-3 and ISSUE-46, accepting Boris's proposal ( only in terms of the syntax of bnodes, and open a new issue on the semantics of bnodes, including editorial note in owl2-syntax about semantics being still open.
16:01:17 [jjc]
+1 (HP)
16:01:20 [pfps]
+1 ALU
16:01:22 [bijan]
+1 (Manchester)
16:01:43 [bmotik]
ACTION: bmotik2 to Update the structural spec to add anonymous individuals; no mention about semantics so far
16:01:43 [trackbot-ng]
Created ACTION-133 - Update the structural spec to add anonymous individuals; no mention about semantics so far [on Boris Motik - due 2008-04-11].
16:01:53 [Achille]
topic: Issue 16 entity annotations
16:02:04 [Achille]
ian: we have a proposal from boris
16:03:06 [alanr]
16:03:10 [pfps]
Boris's proposal may be:
16:04:16 [Achille]
boris: Only annotations . The target of annotation will be either entity or annotation
16:05:17 [Achille]
bijan: i request it to be postponed because I need to think about it more in the context of my annotation proposal
16:05:54 [Achille]
pfps: another proposal (much simpler) from me
16:05:57 [Elisa]
Elisa has joined #owl
16:06:07 [Achille]
pfps: what is the status of entity annotation?
16:06:42 [Achille]
pfps: I would vote to reject the issue because there is nothing wrong
16:06:45 [Zakim]
Team_(owl)13:08Z has now started
16:06:52 [Zakim]
16:07:14 [sandro]
Hold on, Elisa
16:07:18 [pfps]
my solution
16:07:44 [dlm]
dlm has joined #owl
16:09:08 [Zakim]
16:09:11 [Zakim]
16:09:13 [Zakim]
16:10:35 [pfps]
zakim, ??p3 is meetingRoom
16:10:35 [Zakim]
+meetingRoom; got it
16:11:54 [Achille]
jjc: the issue is about metaannotation
16:12:30 [Achille]
jjc: the ability to annotate the annotation
16:12:39 [Achille]
pfps: my proposal does that
16:13:28 [Achille]
bijan: this is an important issue for debra
16:13:41 [Elisa]
We, also, in work for JPL and other DoD projects have a serious need for meta-annotations
16:13:52 [Achille]
achille: this is also an important issue for IBM
16:14:05 [Achille]
ian: we should carry that discussion on emails
16:14:44 [kendall]
+1 to this, one of our users, NCI, really really really wants meta-annotations
16:15:07 [Achille]
ACTION: jjc to drive this issue forward to resolution
16:15:07 [trackbot-ng]
Sorry, couldn't find user - jjc
16:16:48 [Achille]
topic: Issue 72 Annotation Semantics
16:16:59 [jjc]
16:17:05 [Achille]
ian: the current spec gives no semantics to annotation
16:17:16 [Achille]
and it is not 100% compatible with OWL 1
16:17:19 [jjc]
jeremy asked about adding rdf:source as per member submission to allow named graphs to support annotations
16:17:31 [jjc]
peter replied
16:17:31 [sandro]
pfps: if you're proposing that the underlying transfer mechanism be something other than RDF triples (as you are), then you are opening the floodgates of clear water which will wash away a whole lot of mess. I will jump with glee. But you will still be opening flood gates....
16:17:55 [Achille]
alan: the case I am aware of is if you have 2 individuals that are same as
16:18:13 [Achille]
then annotations in one is also annotations on the other
16:18:33 [Achille]
m_schnei: I have also found a bug in the semantics
16:18:51 [Achille]
m_schnei: I will send an email to the WG because I do not remember
16:18:57 [Achille]
... the details
16:19:26 [Achille]
alan: by which mechanism the annotation remains with the annotated object
16:20:09 [Achille]
jjc: this example with sameAs. the annotation is about the syntax (i.e the uri)
16:20:26 [Achille]
jjc: annotations are always about the syntax not the semantics
16:20:58 [Achille]
bijan: I agree with jjc for the most part
16:21:43 [Achille]
bijan: tools can address this issue by presentating same individual together
16:21:45 [Achille]
for example
16:22:04 [Achille]
bijan: it should be handled at the level of tools
16:24:15 [Achille]
bernado: are we talking about a backward compatibility issues or are we talking about annotations in general
16:24:34 [Achille]
bernardo: we should focus on backward compatibility only
16:24:45 [alanr]
cats are friendly, annotation "Alan believes this". cat's are friendly, annotation "Uli believes this". Alan believes same thing as ULI
16:25:47 [alanr]
Bijan asks, does alan believe no cats are unfriendly (if that's what ULI says?). Answer: yes.
16:25:54 [Achille]
evren: second jeremy's point: annotations are on the syntax
16:26:19 [MarkusK]
markus: many use cases for AnnotationProperties could now be addressed by punning
16:26:38 [MarkusK]
markus: (referring to Class-Individual, Property-Individual punning)
16:26:46 [Achille]
bijan: I agree with Markus
16:27:04 [Achille]
MarkusK, thanks for your script assist
16:27:33 [Achille]
16:27:36 [m_schnei]
The discussion between Alan and me about 1.0-DL semantics for annotations: <>
16:28:35 [Achille]
ian: it seems to me that we should accept that the semantic is not backward compatible
16:28:46 [sandro]
16:28:59 [Achille]
alan: my concern is that the semantics of annoation are not understood
16:29:01 [jjc]
16:29:37 [Achille]
alan; my experience tells me that tools do not have a standard way to deal with annotations
16:29:44 [bcuencagrau]
16:30:19 [Achille]
ian: I am saying that the resolution that they do not have a semantics leads to a backward compatibility issue
16:30:56 [Achille]
bijan: two things:
16:31:12 [Achille]
.. 1) what are the semantics?
16:31:27 [Achille]
... 2) the documentation is not perfect
16:31:56 [Achille]
... we should do a better jobs on documentating annotations this time around
16:32:45 [Achille]
alan: we are not committed to backward compatibity on annotation
16:33:14 [Achille]
alan: we will clarify its semantics even with the risk of breaking backward compatibity
16:33:40 [Achille]
sandro: owl 2 is a superset of owl1 with some bugs fixed
16:34:05 [Achille]
... maybe I should add that we may also break compatibility on annotations
16:35:09 [Achille]
alan: my concern is that by saying that it has no semantics, the typical tool may simply ignore them
16:35:24 [Achille]
pfps: what does protege do?
16:35:56 [sandro]
Alan: I don't only author ontologies, I use them! :-)
16:35:57 [Achille]
mathew: protege does not throw them away
16:37:34 [Achille]
evren: 1) OWL 1 has semantics for annotation but tools still ignore them (for example reasoners)
16:38:00 [Achille]
evren: it will be fixed soon in pellet
16:38:30 [Achille]
PROPOSED: we are not committed to backward compatibility with respect to annotations
16:38:35 [sandro]
Possible wording: The OWL Working Group intends to make OWL 2 be a superset of OWL 1, except that some "bugs" may be fixed and annotation semantics may be different.
16:38:57 [alanr]
evren: What I'll be curious about is how queries on syntax are made - looking forward to seeing what you guys come up with
16:39:00 [jjc]
(note HP will abstain, personally I am in favour)
16:39:05 [Achille]
PROPOSED: we are not committed to backward compatibility with respect to annotation semantics
16:43:00 [alanr]
PROPOSED: Close Issue-72 saying we are not committed to backward compatibility with respect to annotation semantics, given that the WG will attempt to improve annotation support in OWL 2
16:43:26 [bijan]
+1 (Manchester)
16:43:29 [Achille]
+1 (IBM)
16:43:30 [msmith]
+1 (C&P)
16:43:33 [bcuencagrau]
+1 (Oxford)
16:43:34 [jjc]
0 (HP)
16:43:34 [MarkusK]
+1 (FZI)
16:43:34 [Rinke]
+1 (Amsterdam)
16:43:35 [alanr]
+1 (Science Commons)
16:43:35 [Zhe]
+1 (Oracle)
16:43:39 [IanH]
+1 (Oxford)
16:43:39 [baojie]
+1 (RPI)
16:43:40 [ekw]
+1 (NIST)
16:43:41 [sandro]
m_schnei: The point is that we want OWL 1 documents with annotations to have the same annotations in OWL 2
16:43:50 [jluciano]
+1 (MITRE)
16:43:56 [sandro]
0 (W3C)
16:43:56 [pfps]
+1 ALU
16:44:07 [alanr]
michael: I don't see anything here that would negatively impact that
16:44:52 [alanr]
RESOLVED: Close Issue-72 saying we are not committed to backward compatibility with respect to annotation semantics, given that the WG will attempt to improve annotation support in OWL 2
16:45:44 [Achille]
topic: Task Force Report: Imports and Versioning
16:46:33 [Achille]
alan: assumptions: 1) ontologies will be versioned
16:47:18 [Achille]
alan: 2) terms may not be versioned , 3) different versions will coexist at different locations
16:49:00 [Achille]
alan presentation will be sent to the WG mailing list
16:49:16 [Achille]
s/ alan presentation / alan's presentation
16:50:44 [IanH_]
IanH_ has joined #owl
16:50:56 [Achille]
alan: after the import closure is done, all the version requirements are checked
16:51:22 [Achille]
alan: in case of violations, tools can request users intervention
16:52:30 [Achille]
jjc: this is a generic pb
16:52:40 [Achille]
jjc: it is not owl specific.
16:52:49 [Achille]
16:53:19 [Achille]
alan: there is an implication that OWL addresses this issue
16:54:19 [sandro]
Bijan: we should be sure to reference David Orchard's work on web versioning in the TAG, and be prepared to discuss differences.
16:58:37 [Achille]
alan: an external mapping file specifies the location of the OWLontologies
17:01:54 [Achille]
alan: this is not a report from the task force
17:02:05 [Achille]
alan: it is just a proposal from me
17:03:27 [Achille]
alan's proposal is at :
17:05:47 [m_schnei]
m_schnei has joined #owl
17:07:34 [Zakim]
17:42:42 [baojie]
baojie has joined #owl
17:51:27 [evrensirin]
evrensirin has joined #owl
17:56:53 [Rinke]
Rinke has joined #owl
17:57:12 [Zakim]
17:57:14 [Zakim]
17:57:14 [Zakim]
17:59:34 [jjc]
jjc has joined #owl
18:03:50 [IanH]
IanH has joined #owl
18:05:04 [thomassch]
thomassch has joined #owl
18:06:36 [alanr]
alanr has joined #owl
18:06:45 [m_schnei]
m_schnei has joined #owl
18:07:02 [ekw]
ekw has joined #owl
18:07:47 [kendall]
Elisa: my pleasure
18:08:00 [bcuencagrau]
alanr: explains his proposal for imports and versioning
18:08:27 [uli]
uli has joined #owl
18:08:32 [MarkusK]
MarkusK has joined #owl
18:08:36 [bcuencagrau]
alanr: ontology documents have a single ontology header
18:09:02 [bcuencagrau]
pfps: there are RDF graphs that are not OWL Full ontologies
18:09:56 [Zakim]
18:09:58 [Zakim]
18:09:59 [Zakim]
Team_(owl)13:08Z has ended
18:10:00 [Zakim]
Attendees were Elisa_Kendall, meetingRoom
18:10:20 [sandro]
zakim, room for 4 for 300 minutes?
18:10:21 [Zakim]
ok, sandro; conference Team_(owl)18:10Z scheduled with code 26631 (CONF1) for 300 minutes until 2310Z
18:10:27 [ekw]
Elisa we are redialing
18:10:36 [sandro]
Elisa, we're now on conf cod 26631
18:10:43 [sandro]
Elisa, we're now on conf code 26631
18:10:51 [bmotik]
bmotik has joined #owl
18:10:54 [sandro]
sandro has changed the topic to: CONFERENCE CODE 26631
18:11:28 [Zakim]
Team_(owl)18:10Z has now started
18:11:35 [Zakim]
18:11:45 [bcuencagrau]
alanr: requirement: imprt by name has to have a precise meaning
18:12:40 [bcuencagrau]
boris: in RDF we can have multiple RDF triples per ontology. If there are many, then they cannot take part in an import
18:13:23 [bcuencagrau]
matthew: taking the xml base of the document would suffice
18:13:34 [bcuencagrau]
sandro: better not use xml:base
18:14:06 [bcuencagrau]
sandro: the RDF mapping should not have anything specific to the concrete serialization used
18:14:49 [sandro]
sandro: (early) and RDFy way to do this is: <> owl:sameAs <...the ontology uri...>
18:14:49 [bcuencagrau]
alanr: the purpose of the proposal is to specify how the tools should work
18:15:07 [sandro]
sandro: or owl:sameOntologyAs if you like.
18:15:17 [bcuencagrau]
alanr: namely, to avoid situations where the tool would not know what to do
18:15:36 [bcuencagrau]
alanr: for this purpose, introduce a special kind of document
18:16:21 [bcuencagrau]
mschneider: some tools like TopBraid composer has such a mapping internally, similar to what the special document mentioned by alan would do
18:17:19 [bcuencagrau]
bmotik: in short, when we import an ontology we should be able to check the version plus a redirection mechnism to go to the right version
18:18:05 [bcuencagrau]
alanr: the main purpose is to identify obvious problems, where before we were not able to detect such problems
18:18:12 [ekw_]
ekw_ has joined #owl
18:18:33 [bcuencagrau]
alanr: this is just to detect more problems than before
18:18:51 [bcuencagrau]
matthew: is there a notion of a ``session''?
18:19:19 [bcuencagrau]
alanr: it does not. It only refers to the simplest case
18:20:26 [bcuencagrau]
bmotik: having one mapping file per session would be a reasonable thing to do in the OWL API
18:20:47 [bcuencagrau]
jluciano: better leave those issues out of the spec
18:21:06 [bcuencagrau]
bmotik: these kinds of issues should be decided by the tool builders
18:21:15 [bcuencagrau]
bijan: this is big
18:21:17 [sandro]
zakim, who is on the phone?
18:21:17 [Zakim]
On the phone I see Elisa_Kendall
18:21:36 [bcuencagrau]
bijan: we are trying to solve quite a huge problem
18:22:01 [Zakim]
18:22:02 [Zakim]
Team_(owl)18:10Z has ended
18:22:02 [Zakim]
Attendees were Elisa_Kendall
18:22:03 [bcuencagrau]
bijan: the scope is probably too large for our schedule
18:22:15 [sandro]
+1 Bijan -- this is at least as big as easy keys, etc.
18:22:36 [bcuencagrau]
msmith: please clarify what aspect of it is too big
18:22:42 [bcuencagrau]
bijan: all aspects of it
18:23:32 [bcuencagrau]
jjc: probably it would be good to have a special vocabulary to deal with the redirections. May not be normative
18:23:32 [sandro]
JJC: One use case is to standardize relation between tools. It could be informative/suggested.
18:23:47 [sandro]
Alan: I want to know when there are errors.
18:24:19 [sandro]
Bijan: owl:imports gets used in RDF. If we propose something like this, ...
18:24:24 [ekw_]
ekw_ has joined #owl
18:24:59 [bcuencagrau]
bijan: RDF does not have an inclusion mechanism. If we propose something like this we should try to reach some consensus even outside the group
18:25:29 [bcuencagrau]
ianH: we are not going to solve this today. We are just presenting proposals
18:26:15 [bcuencagrau]
sandro: there were a couple of formal objections in the old OWL WG concerning imports
18:26:29 [bcuencagrau]
sandro: some people didin't want to include it in the language
18:26:40 [bcuencagrau]
ianH: another proposal by Peter
18:27:00 [alanr]
alanr has joined #owl
18:27:27 [bcuencagrau]
ianH: time wise, could we continue later than 4:30 or 5:00?
18:27:29 [sandro]
no objection to going past 4:30
18:27:32 [sandro]
no objection to going past 5:00
18:27:50 [bcuencagrau]
ianH: we will try to finish around 5
18:27:58 [sandro]
jluciano: I'm probably going to leave around 3.
18:28:20 [sandro]
pfps: the projector is leaving around 5:30.
18:28:44 [bcuencagrau]
peter: imports means import from location
18:29:16 [bcuencagrau]
pfps: if you have versions, then use the same simple mechanism, but then you generate import-specific locations
18:29:48 [bcuencagrau]
pfps: one could also have an annotation property to specify the version
18:29:56 [bcuencagrau]
pfps: does not need to be a number
18:30:34 [bcuencagrau]
pfps: if one imports different versions of a different ontology, one could detect that
18:30:58 [bcuencagrau]
pfps: different versions have same names, but different version numbers
18:31:53 [bcuencagrau]
pfps: it is exactly like the W3C versioning system works
18:32:39 [bcuencagrau]
bmotik: is there a relationship beteen the URI of the version and the annotation?
18:32:45 [bcuencagrau]
pfps: not necessarily
18:33:13 [bcuencagrau]
pfps: one should not import multiple versions of an ontology
18:33:47 [bcuencagrau]
pfps: the annotation is available for pragmatic purposes, but it is meant to be descriptive
18:34:28 [bcuencagrau]
zhe: is there anything in the proposal that deals with version compatibility?
18:34:42 [bcuencagrau]
pfps: no word about version compatibility. That is hard
18:35:10 [bcuencagrau]
bmotik: great, but we need an explicit statement saying what an implementation can do concerning redirection
18:35:50 [bcuencagrau]
pfps: this all is completely outside the semantics of OWL
18:36:51 [bcuencagrau]
mschneider: this is a best practice approach and I like it. Is there anything in alan's proposal that is not covered by peter's?
18:37:03 [bcuencagrau]
alan: yes, the ability to repair
18:37:58 [sandro]
Sandro: I import D1 and D2. D1 contains the fact that it's name is D2, version 1. D2 contains the fact that its name is D1, version 1.
18:37:59 [bcuencagrau]
sandro: with Peter's approach, the provide can open different branches and label them as they want
18:38:09 [bcuencagrau]
pfps: but all this is outside the semantics
18:39:01 [bcuencagrau]
pfps: the annotations can only talk about the ontology it is in
18:39:12 [bcuencagrau]
pfps: not about other ones
18:39:41 [Zhe]
Zhe has joined #owl
18:39:48 [bcuencagrau]
mschneider: clarification required from alan concerning repair
18:40:50 [bcuencagrau]
jjc: request from alan writing what his solution provides compared to peter's
18:41:28 [sandro]
Peter: Ontology-versioning is a good reason to violate the SHOULD about name=location.
18:42:23 [bcuencagrau]
bmotik: in the imports task force, the goal was for each ontology to specify waht particular version is under consideration
18:42:28 [sandro]
+1 Boris -- include in the ontology its "version-name", which is the location at which you can, in the future, get this version.
18:43:06 [bcuencagrau]
bmotik: tools should be able to rename the locations
18:43:31 [sandro]
boris: What you import must have its name or version-name the same as its location.
18:43:34 [bcuencagrau]
alanr: if an ontology is missing or has a bug, there is no way to fix the imports in such situation
18:43:43 [bcuencagrau]
bmotik: you can
18:44:29 [bcuencagrau]
bmotik: one can overwrite globally for all ontologies in the import closure
18:44:38 [kendall]
+1 to moving on, please
18:44:52 [kendall]
+50, in fact
18:45:20 [bcuencagrau]
alanr: in my approach the overwritting is portable whereas in peter's it depends on the tool
18:45:54 [bcuencagrau]
rinke: in peter's proposal import goes by location, whereas in alan's it goes by name
18:46:00 [alanr]
kendall: We had a several week task force, and this is an important part of our meeting.
18:46:19 [bcuencagrau]
rinke: in law applications one would want to do it by name. This is an issue
18:46:24 [kendall]
yes, where "our" includes me, and that's my preference, which I've a right to express here.
18:46:33 [kendall]
this is very much in the weeds IMO
18:46:58 [bcuencagrau]
jjc: the overwriting mechanism already covers this
18:47:47 [bcuencagrau]
pfps: in my proposal it is possible to say that an ontology must import a particular other one
18:48:32 [sandro]
q+ to say a URN is a location.
18:48:39 [bcuencagrau]
bmotik: now we are saying that import is by location, but location and name can be different providing we can overwrite
18:48:51 [bcuencagrau]
rinke: who overwrites the location?
18:48:56 [bcuencagrau]
bmotik: the application
18:49:29 [bcuencagrau]
sandro: the notion of URN is flawed
18:49:49 [bcuencagrau]
sandro: URN refers to a name that refers to some content
18:49:54 [sandro]
18:50:17 [sandro]
Sandro: So when we say "location" here, we really mean any URI which has associated content on the web.
18:50:29 [bcuencagrau]
alanr: people perceive that my proposal allows the user to do too much
18:51:02 [alanr]
clarifying (I was asking this question - I don't think it does)
18:52:00 [bcuencagrau]
bmotik: we want the mapping to depend on the application
18:52:04 [bcuencagrau]
alanr: no way
18:52:15 [bcuencagrau]
alanr: that is not what we want
18:52:39 [bcuencagrau]
bijan: I am a real user, and I don't like writing mapping files
18:53:09 [kendall]
kendall has joined #owl
18:53:15 [bcuencagrau]
inaH: now people are kind of clear what the proposals are about
18:53:38 [bcuencagrau]
sandro: the proposals are the same at the core, but alan's adds a resolver on top of it
18:54:08 [bcuencagrau]
evren: two differences: alan's has the mapping files;
18:55:19 [bcuencagrau]
second: the mapping files can overwrite the imports of another ontology in your ontology
18:55:50 [bcuencagrau]
bmotik: both proposals are quite similar
18:56:19 [bcuencagrau]
bmotik: the difference is that alan's is more expressive because one could detect version incompatibility
18:57:05 [sandro]
Boris: the extra component is whether we have a Portable Way To Specify Location-Overrides.
18:57:34 [bcuencagrau]
bmotik: there is another disagreement. Alan thinks that we should have a portable way of resolving locations across tools
18:57:54 [bcuencagrau]
bmotik: this is a different issue that is independent of what proposal we take
18:58:00 [bcuencagrau]
bmotik: we should split the issues
18:58:26 [bcuencagrau]
jjc: the imports task force has disagreements
18:59:10 [sandro]
jjc: lets give input to the TF, via strawpoll.
18:59:14 [bcuencagrau]
bijan: we should go step by step. Agree with Boris
18:59:26 [thomassch]
thomassch has joined #owl
18:59:41 [sandro]
Bijan: Chunk 1 : name -vs- location. Chunk 2 : Version Management in Imports. Chunk 3: Portable Location Overrides
19:00:37 [bcuencagrau]
alanr: there are problems with Peter' proposal that does not address my user needs
19:00:40 [sandro]
JJC: I like Bijan's proposal better than mine.
19:01:01 [sandro]
JJC: Second proposal that peter's "Basic Idea" is an improvement over the past.
19:01:30 [bcuencagrau]
alanr: Peter's proposal is part of mine but not enough
19:01:41 [jie]
jie has joined #owl
19:01:49 [bcuencagrau]
pter: is it an advance over the current situation?
19:01:53 [bcuencagrau]
alanr: marginal
19:02:04 [bcuencagrau]
pfps: this means yes
19:02:35 [bcuencagrau]
IanH: let's have a straw poll
19:02:42 [pfps]
Basic idea: Imports(U) means import from the location U.
19:02:44 [pfps]
+ the usual overrides to allow caching and local storage
19:02:45 [pfps]
The ontology at U SHOULD have name U
19:03:12 [sandro]
[that was the "basic idea" we have been talking about, from Peter]
19:03:13 [bcuencagrau]
IanH: part 1 on the IRC
19:03:24 [bcuencagrau]
Part2: versioning
19:03:36 [bcuencagrau]
jjc: Peter's or alan's?
19:03:49 [bcuencagrau]
jjc: Part 3: portable mechanism for overwriting
19:04:14 [bcuencagrau]
IanH: part2: should we do versioning?
19:04:42 [bcuencagrau]
jjc: three options: no versioning, peter's versioning, or alan's versioning
19:04:55 [bcuencagrau]
mschneider: there is no versioning in Peter's
19:05:14 [bcuencagrau]
pfps: this is actually an important part: only very minor support for it
19:05:25 [bcuencagrau]
IanH: wave hands
19:05:32 [bcuencagrau]
IanH: Part1
19:05:51 [bcuencagrau]
Ianh: Should the imports task force shoul do it?
19:05:56 [sandro]
NONBINDING-PROPOSAL: We should incorportate peter's "basic idea"
19:06:04 [sandro]
(pretty much unanmous.)
19:06:06 [bcuencagrau]
IanH: clear yes
19:06:14 [bcuencagrau]
IanH: versioning
19:06:32 [bcuencagrau]
Rinke: still not happy with the location part
19:06:34 [baojie]
baojie has joined #owl
19:06:59 [bcuencagrau]
msmith: let's close Issue 21
19:07:20 [sandro]
cf ISSUE-21
19:07:30 [bcuencagrau]
IanH: second part, versioning
19:07:57 [bcuencagrau]
IanH: option 1: do nothing; Option 2: minimal Option 3: rich version
19:08:10 [sandro]
NONBINDING-PROPOSAL: (1) don't do anything about versioning, (2) minimal (Peter's) versioning, (3) substantial (Alan's) versioning
19:08:56 [bcuencagrau]
Option 1: 2 people
19:09:23 [bcuencagrau]
mschneider: we had versioning in OWL 1.0
19:09:25 [sandro]
1 - no change from OWL 1: 2 (bijan, baojie)
19:09:36 [Elisa]
+1 to Peter's
19:09:36 [bcuencagrau]
Option 2: peter's
19:09:45 [sandro]
2 - Peter's proposal: 10 people
19:09:45 [bcuencagrau]
IanH: 10 people
19:09:53 [bcuencagrau]
IanH: alan's
19:09:59 [bcuencagrau]
IanH: 3 people
19:10:00 [sandro]
3 - Alan's: 3 people
19:10:30 [bcuencagrau]
jluciano: is peter's a subset of alan's?
19:10:40 [bcuencagrau]
IanH: let's not go there now
19:11:03 [sandro]
Ian: advice to TF would be to focus in very lightweight versioning
19:11:09 [bcuencagrau]
jluciano: Peter's is a starting point; then we could build on top of it
19:11:35 [bcuencagrau]
Bijan: peter's does not break alan's
19:11:45 [sandro]
Uli: Peter's thing doesn't stop us from doing heavier versioning later.
19:11:49 [sandro]
Alan: well, it might
19:11:52 [bcuencagrau]
IanH: Part 3
19:11:55 [sandro]
Boris: it should not.
19:12:29 [bcuencagrau]
jjc: we could have an informative option
19:12:38 [jluciano]
+1 (MITRE)
19:13:10 [sandro]
Topic: Should we specify a Portable Location Override format?
19:14:37 [sandro]
Spec shouldn't say anything about it: 6
19:15:00 [sandro]
Spec should say something non-normative: 9
19:15:03 [bcuencagrau]
bcuencagrau has joined #owl
19:15:13 [sandro]
Spec should include normative mechanism: 1
19:15:33 [bcuencagrau]
sandro: what does non-normative mean?
19:15:45 [kendall]
non-normative is an expensive trial balloon
19:16:08 [bcuencagrau]
bmotik: means that one recommends implementors what to do
19:16:30 [bcuencagrau]
msmith: it gives a kind of base
19:16:31 [sandro]
-1 to "informative"
19:16:42 [bcuencagrau]
bmotk: there is value in that
19:17:24 [bcuencagrau]
Bijan: we shoul first engage people like TopBraid composers people
19:17:43 [bcuencagrau]
sandro: clarification: it could be normative but optional
19:18:17 [kendall]
kendall: there's nothing OWL-specific here; ergo, this is the wrong WG to do anything (or anything more than very minimal) about this general problem.
19:18:19 [bcuencagrau]
jjc: in an informative version, we could get it wrong
19:18:32 [sandro]
jjc: I want "informative" because we're likely to make mistakes. This way people can adopt it if it's right, or not if it's wrong.
19:18:49 [bcuencagrau]
mschneider: an informative version could become a de facto standard
19:18:52 [sandro]
Okay -- from that angle, I'm okay with it.
19:19:13 [bcuencagrau]
bmotik: we have some chances of getting it right
19:19:41 [bcuencagrau]
bmotik: to make it normative, we should be more precise about details
19:19:51 [Elisa]
fyi, we do cache locally in our tool, which is particularly important for our government customers, but I'm happy to have an informative approach in the spec
19:19:51 [bcuencagrau]
bmotik: too many gaps
19:20:04 [bcuencagrau]
bmotik: going with informative is safer
19:20:12 [bcuencagrau]
bmotik: you can leave many things open
19:20:19 [jjc]
I agree strongly with boris
19:20:59 [jluciano]
take care everyone! (Signing off now)
19:21:18 [sandro]
It sounds a lot like a "Project" file contains the mapping file....
19:22:02 [Elisa]
+1 to Kendall
19:22:08 [bcuencagrau]
Kendall: this is a very general problem. Why are we doing anything there? this is not OWL-specific
19:22:18 [bcuencagrau]
kendall: not an OWL problem
19:22:44 [bcuencagrau]
IanH: let the task force go away and come up with a proposal given the feedback
19:23:12 [bcuencagrau]
Bijan: we should get mopre support from other implementors
19:23:38 [bcuencagrau]
IanH: see if we can close issue 21 on the basis that we believe we have agreement
19:23:47 [bijan]
Scribecorrect: I said that *if* we had explicit support from the implementors, then I would be more inclined to support doing something.
19:24:29 [kendall]
kendall: For example,
19:25:35 [bmotik]
Peter's proposal for resolution of ISSUE-21:
19:25:47 [bmotik]
Imports(U) means import from the location U.Imports(U) means import from the location U.
19:25:52 [bmotik]
+ the usual overrides to allow caching and local storage
19:25:57 [bmotik]
The ontology at U SHOULD have name U
19:26:39 [sandro]
PROPOSED: Resolve issue-21, to say we're importing by location and that, modulo overrides, modulo versioning, the ontology at U SHOULD have name U.
19:27:53 [bcuencagrau]
bmotik: I do not understand what we are voting for
19:28:01 [IanH]
IanH has joined #owl
19:28:28 [sandro]
PROPOSED: Resolve issue-21, to say we're importing by location and that, modulo overrides, modulo versioning, the ontology at U SHOULD have name U.
19:29:09 [sandro]
Evan: as long as URNs are considered location, as Sandro claims, then I'm fine with this. I don't want to outlaw URNs.
19:29:19 [sandro]
Alan: I agree, this protects URNs.
19:29:33 [bcuencagrau]
ewallace: what does location really mean? Are we talking specifically about URLs?
19:29:53 [uli]
uli has joined #owl
19:30:50 [bcuencagrau]
bmotik: location might not match the name
19:31:17 [bcuencagrau]
bmotik: but the ontology location should match the imports
19:31:37 [sandro]
PROPOSED: Resolve issue-21, to say we're importing by location and that, modulo versioning, the ontology imported from U SHOULD have name U. There may be some kind of override of location, but not the name.
19:31:44 [bcuencagrau]
bmotik: modulo versioning
19:34:32 [bcuencagrau]
bmotik: the proposal implies import by location with a possibility to overwrite the location
19:34:33 [sandro]
PROPOSED: Resolve issue-21, to say that import is fundamentally by location, and that what you retreive from that location should give its name as that import-location. There may be overrides along the way to retreiving that content, but the import-text and the final-name text SHOULD be the same. All of this may be changed to accomodate versioning..
19:36:10 [kendall]
next on the agenda: OWLWG solves the problems with SVG and XML Schema!! :>
19:36:24 [bcuencagrau]
Bijan: this is not really imports by location
19:36:34 [bcuencagrau]
IanH: let us move forward
19:36:48 [bcuencagrau]
IanH: let's postpone this to email discussion
19:37:02 [bcuencagrau]
mschneider: what does it mean import by name?
19:37:06 [sandro]
(Boris says he completely agrees with my phrasing in the PROPOSAL)
19:37:30 [bcuencagrau]
bmotik: you never say where the location is. What I propose is that the name and the location should match
19:38:00 [bcuencagrau]
sandro: let us make a straw poll
19:38:18 [sandro]
PROPOSED: Resolve issue-21, to say we're importing by location and that, modulo versioning, the ontology imported from U SHOULD have name U. There may be some kind of override of location, but not the name.
19:38:19 [alanr]
19:38:20 [jjc]
19:38:21 [bmotik]
+1 (Oxford)
19:38:23 [sandro]
19:38:23 [MarkusK]
+1 (FZI)
19:38:25 [Zhe]
19:38:28 [bcuencagrau]
19:38:29 [Rinke]
19:38:31 [alanr]
don't understand
19:38:32 [msmith]
19:38:34 [Elisa]
19:38:39 [bijan]
19:38:41 [baojie]
19:38:53 [uli]
need to think more
19:38:54 [ekw]
19:38:55 [pfps]
+1 ALU
19:39:05 [sandro]
Ian: Okay, moving along!
19:39:12 [Rinke]
(need to think more as well... )
19:39:21 [bcuencagrau]
IanH: no consensus. Let's take it offline
19:39:23 [sandro]
evan: I buy Bijan's argument that this ISN'T imports-by-location.
19:39:41 [Rinke]
I buy that argument as well
19:39:48 [Rinke]
... I think...
19:40:03 [bcuencagrau]
IanH: no user-facing documents discussion
19:40:27 [bcuencagrau]
19:41:04 [sandro]
ian: after break -- Syntax-As-Reference, or EasyKeys, RichAnnotations, NAry.
19:41:21 [bcuencagrau]
ianh: straw poll
19:41:35 [sandro]
syntax-as-reference - 0
19:41:36 [bcuencagrau]
ianh: who wants to do easy keys?
19:41:39 [sandro]
easykeys - 10
19:41:39 [bcuencagrau]
19:41:42 [Elisa]
+1 for easy keys
19:41:43 [sandro]
rich annotation - 3
19:41:46 [bcuencagrau]
19:41:47 [bcuencagrau]
19:41:48 [sandro]
nary - 3
19:41:56 [bcuencagrau]
nary datatypes?
19:41:58 [bcuencagrau]
19:41:59 [sandro]
Ian: Come back at 4pm for EasyKeys.
19:43:20 [sandro]
zakim, who is on the call?
19:43:20 [Zakim]
apparently Team_(owl)18:10Z has ended, sandro
19:43:21 [Zakim]
On IRC I see uli, IanH, bcuencagrau, baojie, thomassch, kendall, Zhe, alanr, ekw, bmotik, MarkusK, m_schnei, jjc, Rinke, evrensirin, dlm, Elisa, Carsten, bijan, RRSAgent, Zakim,
19:43:23 [Zakim]
... sandro, pfps, msmith, trackbot-ng
20:07:01 [evrensirin]
scribenick: evrensirin
20:07:17 [evrensirin]
ian: admin material
20:07:35 [evrensirin]
ian: scribes should clean up the minutes they scribed as usual
20:08:07 [evrensirin]
ian: next f2f scheduled with iswc
20:08:19 [sandro]
NOT "next" Rather, "A Future"
20:08:42 [evrensirin]
ian: we'll setup a poll about f2f date
20:09:02 [evrensirin]
ian: people should consider about hosting
20:09:35 [evrensirin]
ian: make a proposal if you want to host
20:10:22 [evrensirin]
ian: informal telecon next week for people who weren't in the f2f to chat with people who were in f2f
20:11:15 [ekw]
Note that a number of us will be at a meeting 23 - 27 June.
20:11:32 [ekw]
Elisa Can you hear us?
20:11:45 [sandro]
ACTION: sandro set up F2F calendar poll
20:11:45 [trackbot-ng]
Created ACTION-134 - Set up F2F calendar poll [on Sandro Hawke - due 2008-04-11].
20:11:52 [evrensirin]
sandro: put a note saying "before you comment look at the wiki to see if issue is addressed"
20:12:16 [evrensirin]
jjc: add a editorial note at the top of each document saying vocabulary is simplified
20:12:28 [sandro]
Sandro: Any problem with me linking to wiki, instead of including review comments, ?
20:12:39 [evrensirin]
alanr: this is reasonable, any objections?
20:12:42 [sandro]
(general agreement -- no objection)
20:12:55 [evrensirin]
boris: I'll do it now
20:13:01 [evrensirin]
ian: continue with easy keys
20:13:03 [sandro]
Topic: Easy Keys
20:13:27 [jjc]
jjc has joined #owl
20:13:33 [bijan]
20:13:43 [evrensirin]
uli: two goals: 1) do not prevent future extensions
20:14:01 [evrensirin]
...2) an easy extension
20:14:16 [evrensirin]
bijan: this is n-ary datatypes
20:14:41 [evrensirin]
bijan: I'll explain easy keys
20:14:54 [evrensirin]
bijan: having inverse functional datatype properties in full generality is hard
20:15:07 [evrensirin]
bijan: we have an idea how to implement but not efficiently
20:15:26 [evrensirin]
bijan: datatype reasoner needs to consider all literals globally
20:15:33 [evrensirin]
... but keys are very useful
20:15:35 [kendall]
"they" said that about QCR, too :)
20:16:02 [evrensirin]
bijan: we can do explicit keys on explicit individuals
20:16:14 [evrensirin]
... if there are two individuals with same key
20:16:21 [evrensirin]
... infer they are owl:sameAs
20:16:41 [evrensirin]
... you can write this using DL-safe rules
20:17:29 [evrensirin]
... basic rule is if x and y are related to z by the key they are same
20:18:04 [evrensirin]
... people think this kind of functionality is good enough
20:18:22 [evrensirin]
... but writing these rules is not easy or intention-revealing
20:18:34 [evrensirin]
... they keys can be restricted to a certain class, too
20:18:46 [evrensirin]
... you can go further and consider compound keys
20:19:31 [evrensirin]
... e.g. having both same SSN and same age is a key
20:19:50 [evrensirin]
jjc: does this only for named individuals or include bnodes?
20:20:00 [evrensirin]
bijan: no bnodes, it is hard to handle
20:20:26 [evrensirin]
jjc: what are other things I would miss with easy keys vs full keys?
20:20:29 [bmotik]
I've just added an editorial note to all the documents about the vocabulary.
20:21:03 [evrensirin]
alanr: how does this interact with punning?
20:21:34 [evrensirin]
bijan: if you are fine with sameAs between individuals it is ok
20:22:07 [evrensirin]
bijan: regarding jeremy's question, missing keys is ok in easy keys
20:22:22 [evrensirin]
... these are not integrity constratins
20:22:36 [evrensirin]
... having more than one key should be forbidden
20:22:46 [evrensirin]
... you can add a cardinality restriction
20:23:24 [baojie]
have to leave, wish a fruitful session on easy key :)
20:23:29 [evrensirin]
bernardo: afaik, the difference between proper keys is easy keys only affect data (ABox)
20:24:00 [evrensirin]
... proper keys can affect TBox subsumption
20:24:10 [evrensirin]
carsten: modulo nominals
20:24:17 [evrensirin]
bijan: yes
20:24:40 [evrensirin]
jeremy: this looks more like rules
20:24:55 [evrensirin]
bijan: yes but we do not want to tell people to use rules
20:25:14 [evrensirin]
... have special syntax for keys
20:25:52 [evrensirin]
... the syntax lists one or more properties and the class for which the key is defined
20:26:18 [evrensirin]
... more than one property is for compound keys
20:26:31 [evrensirin]
... you are free to mix object and datatype properties here
20:26:40 [evrensirin]
jjc: can I use property chains here?
20:27:03 [evrensirin]
uli: technically we can add them
20:27:22 [evrensirin]
bijan: we didn't think of them but we can add
20:27:55 [evrensirin]
... you can write it in DL-safe rules so it is possible
20:28:11 [evrensirin]
... but for that complexity we might tell people to use rules
20:28:18 [evrensirin]
... but I'm flexible on this issue
20:29:21 [MarkusK]
markus: property chains allow for non-safe (anonymous) middle parts
20:29:23 [evrensirin]
bijan: you can also add property chains by a new property axiom giving the chain name
20:30:00 [evrensirin]
boris: the semantics refers to Hu which is not in anywhere else
20:30:07 [MarkusK]
markus: so with property chains, one can declare a new property that is then used in an easykey
20:30:29 [MarkusK]
markus: this works only for ObtjectProperties
20:30:47 [MarkusK]
s /ObtjectProperties/ObjectProperties/
20:30:50 [evrensirin]
bijan: we introduce O predicate that is true for all individuals
20:30:57 [evrensirin]
bijan: that is how you get safety
20:31:08 [evrensirin]
... Hu does the same for literals
20:31:28 [evrensirin]
boris: OWL is two-sorted so you would need two Hu's
20:32:21 [evrensirin]
uli: O is the name of the ontology and use Hu for objects and literals
20:32:39 [evrensirin]
uli: we can have Hu1 and Hu2 separately
20:32:47 [evrensirin]
boris: yes you have to have two
20:33:27 [evrensirin]
bijan: lexical difference is an issue
20:34:01 [evrensirin]
bijan: also you can have restriction >16 and <18 which means the value is exactly 17
20:34:07 [evrensirin]
jjc: how do you write that?
20:34:25 [evrensirin]
bijan: using dataranges and restrictions on facets
20:34:52 [sandro]
bijan: the difficulty is not lexical variants, .... it's cases where you infer the properties value with facets.
20:34:53 [evrensirin]
boris: we have unique interpretation and if we drop it it wouldn't be easier
20:35:33 [evrensirin]
jjc: if I have an inferred value for an easy key for a named individual does the key apply?
20:35:37 [evrensirin]
bijan: yes
20:36:07 [evrensirin]
alanr: but not if the subject is inferred
20:36:49 [evrensirin]
bijan: boris, you can search for all literals and then lexically normalize them
20:37:18 [evrensirin]
boris: from every constant we need to uniquely interpret it
20:38:07 [evrensirin]
bijan: counter example is addition of a new fact might merge two existing individuals that weren't merged before
20:39:05 [evrensirin]
bijan: you can get literals with existential restrictions
20:39:16 [evrensirin]
boris: but it is not in Hu
20:39:25 [evrensirin]
bijan: no, it would be in Hu
20:40:00 [evrensirin]
uli: this might make implementations a little harder but otherwise there are other complications
20:40:26 [evrensirin]
jjc: are there other things that I should be worried about other than bnodes?
20:40:33 [evrensirin]
bijan: not that I can think of
20:40:50 [evrensirin]
ian: what if the restriction specifies a range?
20:41:21 [evrensirin]
... the range can be a finite range then it is not that easy any more as Boris says
20:41:29 [evrensirin]
uli: it is as easy as DL-safe rules
20:41:48 [evrensirin]
bijan: you can always build a disjunction to denote that range
20:42:01 [evrensirin]
boris: in OWL I can say greater than 3 which is infinite
20:42:08 [evrensirin]
... so what do I put in Hu?
20:42:19 [evrensirin]
... I'm not sure if this is decidable
20:42:37 [evrensirin]
... also how about complements?
20:43:12 [evrensirin]
bijan: there was another choice to make it more easier
20:43:24 [evrensirin]
... we discussed that possibility
20:43:39 [evrensirin]
... but then sometimes you get surprising results
20:44:03 [evrensirin]
... when you mention one literal explicitly you get new inferences
20:44:43 [evrensirin]
alanr: I can make a class whose age is 17 is equal to nothing
20:44:54 [evrensirin]
... is this still easy with your proposal?
20:45:09 [evrensirin]
uli: mention two surprising results
20:45:41 [evrensirin]
... saying john's age is 17 and saying john belongs to class of people whose age is 17 would give different results
20:45:55 [evrensirin]
jjc: is there rdf/xml syntax for this?
20:46:11 [evrensirin]
bijan: we didn't do the work without knowing if wg would like the proposal
20:46:29 [evrensirin]
boris: as feedback, I like the general idea but I don't understand details
20:46:44 [evrensirin]
... without seeing more details I don't understand this is decidable
20:46:49 [evrensirin]
bijan: fair enough
20:47:09 [evrensirin]
m_schnei: this is trivial but where is the datatype defined?
20:47:36 [evrensirin]
bijan: could be with a range restriction
20:48:29 [evrensirin]
markusk: since this interacts with datarange facets there are no implementations
20:49:11 [evrensirin]
ian: I agree that I want to see a more formal presentation of semantics and proof for decidability
20:49:42 [evrensirin]
boris: the herbrand universe should be finite for decidability
20:49:49 [evrensirin]
uli: not necessarily
20:50:15 [evrensirin]
uli: before going with semantics I want to learn how people feel about oddities
20:50:35 [evrensirin]
alanr: how about 1^^xsd:int and 1^^xsd:double as key value?
20:50:50 [evrensirin]
msmith: int and double are disjoint
20:52:12 [evrensirin]
alanr: so the first question is are we happy with easier one with oddities and how much work is needed for complicated version?
20:52:24 [evrensirin]
boris: I cannot say without more details
20:52:38 [evrensirin]
pfps: we are happy to see the proposal and would like to see more
20:53:07 [evrensirin]
jjc: I have an issue with simpler case
20:53:56 [evrensirin]
m_schnei: can you put something in the wiki about not-easy keys
20:54:15 [evrensirin]
alanr: we are 5min to 5
20:54:28 [evrensirin]
uli: I want three-fold indication for easy keys
20:55:28 [evrensirin]
uli: you can introduce proper idfp if you don't use the key in restriction
20:55:47 [evrensirin]
uli: option 1 - very easy keys with oddities
20:55:55 [evrensirin]
... option 2- as we presented
20:56:38 [evrensirin]
... options 3 - introduce proper idfp without using in restrictions or non-finite dataranges
20:56:50 [evrensirin]
alanr: anybidy who doesn't want either?
20:56:54 [evrensirin]
alanr: noone
20:57:14 [evrensirin]
alanr: large part of wg want at least very easy keys
20:57:29 [evrensirin]
alanr: who wants option 2 with uli and Bijan providing prrof?
20:57:36 [evrensirin]
alanr: fairly good support
20:58:01 [evrensirin]
alanr: who likes option 3?
20:58:09 [Elisa]
I would be interested at least 1 and 2
20:58:29 [evrensirin]
markusk: you can still have easy keys and proper ifdp
20:58:32 [evrensirin]
bijan: yes
20:58:51 [evrensirin]
alanr: thanks for coming
20:59:08 [pfps]
pfps has joined #owl
20:59:13 [sandro]
carsten: thanks to chairs
20:59:15 [sandro]
21:00:57 [thomassch]
thomassch has left #owl
21:01:13 [sandro]
RRSAgent, pointer?
21:01:13 [RRSAgent]
21:24:22 [Carsten]
Carsten has joined #owl
21:32:32 [IanH_]
IanH_ has joined #owl
21:43:36 [Zakim]
Zakim has left #owl
22:18:53 [alanr]
alanr has joined #owl