ppy/chat-transcript_edited_20120329b.txt
PeterYim: Welcome to the
= OntologySummit2012: Session-12 - Thu 2012-03-29 =
Summit Theme: OntologySummit2012: "Ontology for Big Systems"
Session Topic: Organizing the 'Big' Communique
OntologySummit2012_Communique co-Lead Editors & Session co-Chairs: ToddSchneider & AliHashemi
Panelists - Track Champions and Co-editors of the Communique:
* Track-1&2: Ontology for Big Systems and Systems Engineering - MatthewWest & HensonGraves
* Track-3: Challenge: Ontology and Big Data - ErnieLucier & MaryBrady
* Track-4: Large-Scale Domain Applications - SteveRay & TrishWhetzel
* Cross-Track-A1: Ontology Quality and Large-Scale Systems - AmandaVizedom & MikeBennett
* Cross-Track-A2: Ontology for Federation and Integration of Systems - CoryCasanave & AnatolyLevenchuk
* OntologySummit2012_Symposium Co-chairs: RamSriram and MichaelGruninger
* OntologySummit2012 General Co-chairs: LeoObrst and NicolaGuarino
Session page: http://ontolog.cim3.net/cgi-bin/wiki.pl?ConferenceCall_2012_03_29
Mute control: *7 to un-mute ... *6 to mute
Can't find Skype Dial pad? ... it's under the "Call" dropdown menu as "Show Dial pad"
== Proceedings: ==
anonymous morphed into ElizabethFlorescu
anonymous morphed into ErnieLucier
ToddSchneider: Good afternoon/morning everyone.
ToddSchneider: Will be with you shortly.
anonymous morphed into TomTinsley
MikeBennett: I have to jump off just before 2 Eastern / 11 Pacific.
anonymous morphed into DougFoxvog
MatthewWest: I've had my head down in our own track, so do not have a broad view of this summit.
MikeBennett: One possible theme is the different applications of ontologies, as a technical artifact in its own right, and as a means to capture common semantics across some large engineering system. I don't know if that fits with what you are looking for here though.
AliHashemi: SteveRay points out that many systems: software, enterprises are based around "Model Driven Systems"
MatthewWest: Model driven is closely associated with semantics of course.
NicolaGuarino: Modelling is much more general than ontological modelling
MikeBennett: @Nicola agreed. And ontology has broader applications than model driven engineering (indeed, the latter has been a minority case in the Semantic Web world but has been shown to be important in the big systems context)
NicolaGuarino: @Mike agreed
TerryLongstreth: Lemma: Ontological methods can be applied to engineering models to improve depth and breadth of modeling semantics
NicolaGuarino: @Terry: I agree very much. Ontological analysis and actual engineered ontologies just complement (in a very useful way) model driven engineering (for instance model driven engineering based on systems of differential equations)
BobbinTeegarden: @Steve Are you implying, perhaps, that the (ontology)Model IS the System (as in Model Driven Architecture (MDA))?
SteveRay: @Bobbin: Yes I am.
HensonGraves: agree that we are in a model driven age. Also our modeling language are not as good as we need. Ontology is the value proposition to make the models work.
TerryLongstreth: @Henson: +1
SteveRay: I suppose my point is that ontological modeling is a better, more rigorous way of modeling in general.
MatthewWest: Engineering models are more often mathematical than logical, but there are none the less ontological elements.
MatthewWest: @Steve: does that mean you propose replacing mathematics with logic?
SteveRay: @Matthew: Not really. Logic is just a part of mathematics, right? Where it makes sense, use logic. Where a differential equation makes sense, by all means use that.
MatthewWest: @Steve: Yes, but most people see ontology as being limited to expression in logic, and not to include broader mathematical models.
SteveRay: @Matthew: Fair enough. For inherently numerical problems, I would agree that mathematics as traditionally understood is best (such as a control system for example). But for symbolic problems, ontology models are best.
SteveRay: @Matthew: So, both are models, and in fact I would submit that an ontological model provides the contextual framework in which a mathematical model operates.
MatthewWest: @Steve: Agreed.
JimKirby: Where are the slides?
MikeBennett: @Jim on the hopper (vnc server) http://vnc2.cim3.net:5800/
ErnieLucier: @Jim if you do not have access to hopper (the vnc server) then http://ontolog.cim3.net/file/work/OntologySummit2012/OntologySummit2012_Communique/2012-03-29_draft-review/OntologySummit2012_communique-drafting-I--ToddSchneider-AliHashemi_20120329.pdf
JimKirby: @Ernie Thanks!
AnatolyLevenchuk: We may at least tell that ontology is about meta-modeling part of modeling. There are many levels of meta-models and models, therefore we have difficulties in differentiating ontologizing and modeling (and programming too). Model transformations, compilation and mapping is about the same activity.
anonymous morphed into MaryBrady
RexBrooks: While I haven't come to any overarching conclusion, I am now using UML Modeling in Enterprise Architect and Owl Ontology / Ontologies in Protege, and they are quite useful when working back and forth from one to the other for specific classes, terms, systems-programs, etc. Of course having a coordinated set of ontologies and models as the end products is very handy as resources and references for getting specific kinds of information about these things as needed.
RexBrooks: I haven't gotten to the point where using these with inferencing engines or open data sources with SPARQL but I expect that to become even more useful.
MikeBennett: @Rex have you considered using the Ontology Definition Metamodel (ODM) so as to have your ontologies and logical UML models in the same tool? Mail me off list if you need to know details.
BobbinTeegarden: @Rex Enterprise Architect is just coming out with an OWL Plugin, very formative stage; and Elisa Kendall's VOM Plugin is more mature (and does follow ODM, ref by Bennett).
BobbinTeegarden: @Rex VOM Plugin is in MagicDraw, just fyi.
MikeBennett: @Bobbin agreed. Also lets one generate OWL for use in Protege tools.
RexBrooks: @Bobbin-Mary-Matthew: Thanks very much. Wish I could afford MagicDraw, but I'm glad to hear that there is a plugin on the way for EA. However, I will probably continue to use them as springboards back and forth, creating a kind of synergy I haven't had before.
RexBrooks: @Mike: I had your email on another machine that failed recently. I would like to contact you about the ODM. I was aware of it, but not this capability. My email is rexb[at]starbourne.com
AliHashemi: ErnieLucier suggests that the distinction between Current Problems and Uses is unclear.
AliHashemi: NicolaGuarino suggests that section headings should convey more meaning.
ErnieLucier: I have to leave now.
MaryBrady: @Ernie: I can stay for just a bit longer...about 1:15
anonymous morphed into GiancarloGuizzardi
AliHashemi: google-doc of the developing communique draft is at: https://docs.google.com/document/d/1OG_iNRROkfh2T76Ri0SrNzwLwVKGKo4kQOWwBKxHjy8/edit
AliHashemi: Please note - anyone with this link can edit the document (while we are in-session now)
PeterYim: @Henson, @Matthew - the figures are now in - see: http://ontolog.cim3.net/cgi-bin/wiki.pl?OntologySummit2012_BigSystemsEngineering_Synthesis
TrishWhetzel: Regrets, I need to leave the call now.
NicolaGuarino: Ontological analysis as enabler of good modeling. I endorse this very much. Very crisp statement.
NicolaGuarino: (who said that?) ... [it was HensonGraves and MatthewWest citing that as being among the track-1&2 key conclusions; the statement was reiterated by session co-chair ToddSchneider just now.]
GiancarloGuizzardi: @Nicola: Fully agree.
SteveRay: +1 on Nicola's statement
CoryCasanave: We should not differentiate modeling and ontological analysis, ontological analysis should be positioned as part of modeling and one that is emerging as best practice. The precise modeling encompassing ontological analysis is a key enabler to the model driven approach Steve identified.
MaryBrady: Regrets...I too have to leave.
PeterYim: I just want to emphasize that some statements (or recommendations) made (say, by panelists or even in the syntheses) are context sensitive. If we don't have the luxury (say, limited by document length constraints) in the synthesis write-ups and/or the communique to provide those context, we should avoid citing them out of context.
MikeBennett: Cross Track X1 (Ontology Quality), the Google Doc seems to incorporate our community input page and not our track champions' synthesis page.
HensonGraves: @amanda, there are well developed methods for validating models, e.g., but test. Presumably these methods could be used to test ontologies. also you could build on Nicola's notion of ontology correctness
AmandaVizedom: @henson, yes, and there are even techniques for unit testing, and various researchers have been developing more quantitative measures of other ontology characteristics that may or may not be applicable to particular cases... and there are many techniques for in-use testing and domain expert validation that are not well documented. That's one step; finding more ways to streamline and/or automate is another.
DougFoxvog: @Amanda: Could you provide a link to methods/tools for validating ontology quality that you were referring to? Are you referring to tools such as OntoClean?
AliHashemi: @Doug - this probably isn't the same as what Amanda suggests, but this is also relevant: http://arxiv.org/abs/1202.3602
AmandaVizedom: @Doug, I was mostly talking about the need to document approaches to evaluation, so naturally I cannot provide links. But the reference library has a good start for discovery of some of what is documented.
BobbinTeegarden: Something about quality and requirements sometimes missed: if the goal is to tune the current system, quality/requirements are important; but if the goal is to use modeling to do 'possibility' thinking, integrate newness, or get out of the box and design a future system or enhancement, it's more a creative sketching activity and 'quality' is more of an inhibitor, requirements are highly conceptual... Is this worth saying?
AmandaVizedom: @Bobbin- Conceptual requirements are still requirements! But more generally, I'd say that this is part of the way that requirements vary with usage. And as a reminder, by "quality" here we are limiting ourselves to the engineering sense: the quality of something is the degree to which it meets requirements. So, if some characteristic (computational properties, reusability, consistency with X,....) isn't a requirement of the usage, it shouldn't be part of the quality measurement for this usage. What we need is better, explicit, and well-grounded understanding of what requirements go with what usages!
CoryCasanave: don't know why my call dropped!
KathyEllis1 morphed into KathyEllis
HensonGraves: [ref. Anatoly's point about "metamodelling = ontologizing"] @anatoly, I agree with you
MikeBennett: Nicola is making a very important point here: metamodels and ontologies are not in any way the same thing.
SteveRay: Agree with Nicola. Metamodelling would refer to M2. Modeling would be M1.
HensonGraves: @steve, an auto is M0, the model is M1, and the metamodel for autos can be at M2
SteveRay: @Henson: Agreed
SteveRay: [ref. Anatoly's remark that Nicola's rejection of "metamodels=ontologies" is possibly related to "presentation versus representation"] Nicola is not talking about presentation versus representation.
HensonGraves: @nicola, the conceptualizations and patterns can be represented within metamodel. the model of a system is an instance of the metamodel of a system as a pattern
MatthewWest: Ontology is useful at each (meta) level, and in distinguishing between the levels.
NicolaGuarino: At every modeling level there is a corresponding (often implicit) ontology. Ontology does not just belong to the meta level
AliHashemi: +1 to Nicola's point
SteveRay: @Nicola: Also agreed. I don't think ontology is better suited for one metal level or another. It is orthogonal to the metal level. It's just a better way to model.
CoryCasanave: @Steve +1 - semantic modeling at all levels!!
MatthewWest: @Ali and Todd: [ref. Todd suggesting to do some real time cut-and-paste into the developing communique draft] Please do not do that. We have provided input that was roughly in the order of the outline, just take it offline.
PeterYim: +1 on what MatthewWest is suggesting - that the lead editors should just make the calls and come up with a first draft based on what the champions have turned in
MatthewWest: It looks like you already have our stuff in there.
PeterYim: +1 on Steve's remark about clarifying "Current State" as being "Current state of the practice" vs. "state of the art"
SteveRay: Absolutely agree with what Henson is saying
PeterYim: @Henson - well said - can you document that on the chat, please
AliHashemi: [documenting what Henson just said ...] Shift towards explicit semantics ... from informal modeling to modeling in formal languages ... to underpin modeling languages w/ explicit semantics ... to understand the underlying ontology of the elements of the languages
SteveRay: Eh?
AliHashemi: ?
SteveRay: Well defined semantics without knowing what the context is?
NicolaGuarino: ... and there is also a shift from just using *ontologies* (as useful engineering artefacts) toward using *ontological analysis* (as a methodology which helps understanding and disentangling the complexity of big systems)
MatthewWest: @Nicola: +1
GiancarloGuizzardi: @HensonGraves: Yes. I agree with that point. Formal characterization should reflect ontological distinctions. Formal semantics cannot guarantee quality per se. Logics (or any piece of mathematics for that matter) does not care what we do with it and, thus, cannot itself fully constrain the possible interpretations of a model (and a metamodel) to the intended ones
GiancarloGuizzardi: @Nicola: Fully agree with that.
AnatolyLevenchuk: @Nicola not ontological analysis but ontology engineering (like requirement engineering and systems architecture engineering along with requirement analyses etc. as small part of engineering thing)
MikeBennett: Apologies, I have to drop off now.
PeterYim: "inferencing" helps make sense of "big data"
PeterYim: ontological engineering helps augment humans in dealing with "big data" by off-loading a lot of the work to machines
MatthewWest: @Peter: Yes, seeing how ontology can help to automate mundane but necessary activity.
TerryLongstreth: Ontological analysis requires a canonical methodology, which may equate to ontological engineering, but I think should be broader
AnatolyLevenchuk: @peter better ontology engineering (not ontological). We then have ontology as explicit engineering artifact with life cycle, practices (like analysis, management etc.).
NicolaGuarino: Thank for you efforts, Todd & Ali!
anonymous morphed into NikolayBorgest
AliHashemi: re. "Reference List" see - https://www.zotero.org/groups/ontologysummit2012/items/collectionKey/I4QX3RT7
AmandaVizedom: re. "Reference List" content page on the wiki is at - http://ontolog.cim3.net/cgi-bin/wiki.pl?OntologySummit2012_RecommendedReading
CoryCasanave: How wide or narrow do we consider "inferencing", production of derivative information from models is done a lot, inference is more identified with FOL
PeterYim: @Cory - not necessarily, even simple inferences (say, applying modus ponens) can prove to be useful
CoryCasanave: @Peter - I agree but that may not be the interpretation of readers
PeterYim: @Cory - guess we (the lead editors) will just have to word it properly to make sure that we are looking at a spectrum of possibilities
CoryCasanave: @Peter - good, but not easy!
GiancarloGuizzardi: Folks. I have to drop off now. thanks for all the effort. bye
PeterYim: Bye, Giancarlo ... thanks for joining us today!
NicolaGuarino: I have to go as well. Bye bye folks, good session!
PeterYim: great session ... lots discussed and done!
PeterYim: Ali says: we will be publishing a draft of the communique the day before our session next Thursday
PeterYim: -- session ended: 11:23am PDT --