ppy/OntologySummit2011-Launch_chat-transcript_edited_20110120b.txt SteveRay: Welcome to the OntologySummit2011 Launch Event - Thu 2011-01-20 Topic: OntologySummit2011: "Making the Case for Ontology" Co-chairs: Dr. SteveRay & Dr. NicolaGuarino anonymous morphed into YuLin anonymous morphed into David Price David Price morphed into DavidPrice SteveRay: Early birds! SteveRay: See details on the session page at http://ontolog.cim3.net/cgi-bin/wiki.pl?ConferenceCall_2011_01_20#nid2L7T anonymous morphed into DavidEddy anonymous1 morphed into JulitaBermejoAlonso anonymous morphed into PierreGrenon anonymous morphed into MichelleRaymond anonymous1 morphed into AlStevens anonymous morphed into LaurentLiscia OASIS MikeDean1 morphed into MikeDean LaurentLiscia OASIS: I'm delighted to be here (albeit lurking in the wings) LaurentLiscia OASIS: Congratulations to you all as you launch into your new summit. anonymous1 morphed into EdDodds anonymous2 morphed into BobbinTeegarden LaurentLiscia OASIS: And let me state that the case for ontology does not actually need to be made: anyone in standards knows that ontology is the way to go. anonymous morphed into ElizabethFlorescu NicolaGuarino: *3 to unmute, right? FabianNeuhaus: Yes you are right MikeBennett: Did Peter just mute himself? PeterYim: got disconnected just now ... I'm back in Ramdsriram: Steve: How do you un-mute here. FabianNeuhaus: @Ram: *3 Ramdsriram: @Fabian: Thanks anonymous1 morphed into AliHashemi EdDodds: Would folks mind listing any of their Twitter or Identica IDs on the chat? @ed_dodds @conmergence fwiw MikeBennett: Twitter ID: @MikeHypercube AmandaVizedom: Twitter ID: ajvizedom Gary Berg-Cross: Twitter ID garybcross AlStevens: Twitter ID astevens SteveRay: My Twitter id: steveraysteve DavidPrice: Twitter ID : davidpricenet AmandaVizedom: Relatedly, last year we regretted not having selected a nice, short #hashtag for the conference. Anything for this year, organizers, or is the door open for suggestions? EdDodds: #ontologysummit2011 works for me AmandaVizedom: @Ed complaints last year about using 19 char for a hashtag EdDodds: #ontsum2011 SteveRay: Understand the long hashtag problem, and yet we benefit from the tweets showing up under general searches if we don't abbreviate. anonymous1 morphed into JohnSowa anonymous2 morphed into JVermeer EdDodds: We could target the creation of open courseware for ontology related matter (DSPACE [http://www.dspace.org], Open Courseware [http://www.ocwconsortium.org/], curriki [http://www.curriki.org], etc.) since this kind of material could prosper in the distance education world. EdDodds: RE: Education - James H. Shelton III, Assistant Deputy Secretary for Innovation and Improvement http://www2.ed.gov/news/staff/bios/shelton.html might be enlisted as an ontology ally anonymous1 morphed into JeffreyAbbott anonymous2 morphed into ArturoSanchez ToddSchneider: @Nicola, is the IAOA web site using https yet? JohnBateman: @ToddSchneider: Nicola, is the IAOA web site using https yet? An SSL-based layer will soon be brought online for membership registration. ToddSchneider: @John, excellent. I've been waiting to join. anonymous2 morphed into TaraAthan EdDodds: FWIW Real-time user case example: Drupal 7 - http://groups.drupal.org/node/120684 Semantic Web: RDF instance data (and ontology) management on a Sesame Triple-Store MikeBennett: Thanks Ed. SteveRay: JohnSowa said: Steve Jobs doesn't use metrics. We can't measure where we want to go. This is a problem. ToddSchneider: @John, developing a 'wow factor' for infrastructure has always been a challenge. anonymous1 morphed into Mary Parmelee ArturoSanchez: @SteveRay and @Nicola: second that! -- Peter is instrumental! ArturoSanchez: @MichaelGruninger: Re: Track 2: I suggest to use "Ontology-Driven Software Engineering", in lieu of "Ontology-Driven Software Design" ... ToddSchneider: @Arturo, @Michael, suggest replacing 'software' with 'system'. ToddSchneider: @Michael, it sounds like you'll be able to develop architectural patterns for the uses of ontologies/semantic technologies. ArturoSanchez: @ToddSchneider: "O-D Systems" >> Refers to a class of software systems (I presume); "O-D Software Engineering" >> Refers to the practices associated with building "O-D Systems". So, it seems we want to discuss both concepts ... NicolaGuarino: I support Todd very much. We should not be limited to software systems. Indeed, the general perspective concerning the future use of ontologies is much broader: from information systems to socio-technical systems. MichaelGruninger: @Arturo -- I agree, we should refer to "Ontology-driven Software Engineering" NicolaGuarino: AlanRector said: we should clarify relationship between ontologies vs.data models and data structures PeterYim: @Alan and all ... if you can capture your main points on this chat board before or after your verbal remarks, that would be great (as the chat-transcript will be captured and become searchable later) EdDodds: Is OMG.org represented here by anyone -- model driven is their thing MikeBennett: We are liaising closely with OMG in our work at the EDM Council and are using their ODM metamodel. ArturoSanchez: @EdDodds: I am not from OMG, but have been associated with "Domain-Specific Software Development" (e.g., ongoing workshop at OOPSLA) ToddSchneider: @Ed, OMG's model-driven == UML Ramdsriram: @Todd Not sure I understand that what you mean by model-driven == UML. I believe UML is used as a scheme to represent the model-driven architecture concepts. EdDodds: @ToddSchneider True but do a quick search on UML & Ontology; quite a little material out there ToddSchneider: OMG's model-driven approaches are based on UML. EdDodds: Yes MikeBennett: For now AlanRector: Agree that UML is a poor way to develop ontologies; but many of the ontologies we develop have to have, as at least one of their outcomes, UML models / data models or at least provide constraints on them. Furthermore, thre are far more people trained in UML/MDA than ontologies. I look forward to a longer discussion on 3 Feb. DavidPrice: @MichaelGruninger : Seems slide 3 is 'What roles do ontologies play in applications?'and suggest adding'What are the advantages of the use of ontologies in those roles?' JulitaBermejoAlonso: There has been some developments on ontologies and UML (even suggesting to move forward to OMG's SysML), mostly for agent-based systems and software engineering RexBrooks: Just to let you know I'm listening, The Open Group attempted to use a form of UML for a SOA Ontology and the result was neither especially useful, understandable or successful for a number of reasons, but UML is not a great fit for ontologies unless one is very, very careful. EdDodds: @RexBrooks Thanks RexBrooks: @Ed Dodds You're welcome. RexBrooks: Provided we provide an adequate disclaimer, it would be interesting to see how well an application based on the TOG SOA Ontology performs according to the criteria being discussed. DavidPrice: @ToddSchnieder Another core technology behind OMG MDA is QVT ... what's MDA without a transformation engine. Ontology-driven should/could drive a similar need. ToddSchneider: David, What does QVT stand for? DavidPrice: Query/View/Transform Gary Berg-Cross: Mike G makes a point (in passing) the Ontologies provide value (the beef) for something like model-based development. MichaelGruninger: @Todd: interesting idea to think of the framework as providing architectural patterns; perhaps this will guide us in how detailed we need to make the framework ToddSchneider: Michael G., Alan's point suggests a decomposition of 'uses' of ontologies in systems development. AlanRector: On applications: a) Important to clarify relationship between ontologies and data structures / data models. AlanRector: b) Links to standard software engineering methodologies, UML, MDA, etc AlanRector: c) To be specific on the added value of "ontology driven architectures" Ramdsriram: @MichaelGruninger: There is a framework called Zachman Framework. Have you hear of that. May be useful in organizing your track. MichaelGruninger: @DavidPrice: I saw the discussion of the advantages of ontologies being the focus of Track 3; the framework would be used to help understand how different applications can be compared to each other. DavidPrice: @MichaelGruninger OK, makes sense there. JimRhyne: It is hard to argue that something computational can only be done with an ontology. Almost anything I can do with, e.g. OWL/Pellet I can also do with Java and a database. The real difference is how easy it is to create and maintain the behavioral rules for an application in an ontology. A similar argument has been made for rule systems. The problem with rule systems is the difficulty of debugging without some kind of consistency checker. DavidPrice: @JimRhyne I doesn't have to be about 'can only be done', can simply be 'can be done better, faster, cheaper' JimRhyne: @DavidPrice - yes, Mike retracted his earlier remarks on slide 2 NicolaGuarino: @MikeBennett: not just "the best that they have", but perhaps also "the worst that they have"... Learning from failures in applying ontologies might be very useful... MikeBennett: @Nicola that is a good idea. I think what we want for the summation at the Face to Face would be what people did that worked and what they would do differently next time. I'm not sure we'll attract people to present on their failures though PeterYim: @Nicola ... that said, we still want to build a repertoire of "Best" cases that people can point others to when they are "making a case for ontology" DavidEddy: ...we're 60+ years into software & haven't quantified it or applied metrics. Why bother with Ontology? DavidPrice: @MikeBennett slide 5 : Why only business case for ontology as a whole? Why not allow ontology in combination with other technologies or aspects of technologies? Don't understand what that limitation is included. MikeBennett: @DavidPrice that was weak phrasing on my part. Ontology as a whole range of formal models of reality, not just one type like OWL or RDF. Not ontology in isolation. Will look at how these delivered some value, which is almost always in connection with some application. BrandNiemann (Semantic Community)1: Re "While the field of ontology, in the information science sense, ... I suggest we consider ontology, in the data science sense, - see http://semanticommunity.info/Data_Science BrandNiemann (Semantic Community)1: Mills will probably speak to the value proposition next week, which suggests we broaden out to building knowledge-centric systems, not IT centric systems, in which ontology may or may not be needed - see Knowledge-Centric Paradigm: A New World of IT Solutions @ http://semanticommunity.info/@api/deki/files/8282/=BrandNiemann01112011.ppt PeterYim: @Todd - slide#3 ... [ref. Todd's remark about concentrating only on monetary value] sure we will focus on "value" wrt to ROI (return on investment), but thought we will *also* address other intangible values as well, like quality improvement, strategic impact, and others that have been brought up on the [ontology-summit] list earlier Brian Haugh: The proposed metrics focus appears to neglect key metrics used in evaluating the quality of results, such as precision and recall in search. SteveRay: I agree with Peter's point, and Brian's suggestion. The metrics can be monetary, or other types like quality, performance, capability... BillHogan: I would say that how ontologies facilitate/improve/affect all these aspects of your overall architecture are more important than these things as attributes of the ontology itself. ToddSchneider: Brain, The metrics I suggested are needed to to bolster the case to decision makers. They need to be simple and related to the primary interests of the decision makers. Your example suggests a particular usage and the metrics for that case may need to stress performance metrics. ToddSchneider: @Peter, intangible values rarely are of interest to 'decision makers'. However, there may be a useful way to connect these to more base ROI metrics. I leave to the community to help solve this. PeterYim: [subsequently added] @Todd: for C-level executives and Policy makers, strategic value (mostly intangible) would be very pertinent NicolaGuarino: @Todd: To understand value metrics and value models, maybe it would be useful to develop an ontology of value and value models... See http://www.vmbo2011.ugent.be/VMBO2011/Welcome.html ToddSchneider: @Nicola, Yes that occurred to me, but was hesitant to introduce that notion due to the required work and constrained time line. In principal whatever is developed for track 3 will provide a basis for such an ontology. AlanRector: @RexBrooks & ToddSchneider: How does the argument for ontology in general relate to the ontology spectrum in slide 4 ? ToddSchneider: @Alan, Slide 3, Ontology Spectrum, was only to suggest that there may need to be either multiple metrics or value sets for the metrics. RexBrooks: @Todd We should contact Kurt Conrad on the value ontology (ontology of value types) and/or value model ontology (ontology of models associated with various value types). This is, of course, directly related to architectural models, too, hence an NCOIC connection. AmandaVizedom: Must drop off. Thanks to presenters; looking forward to continuing sessions. DavidPrice: @ToddSchnieder I find it hard to make sense of 'ROI for an ontology'. I understand ROI for an application as that's what affects an organization, but not for particular components of that application. I hope the track can help answer this question. MikeBennett: @DavidPrice re ROI for an ontology. If someone tried to solve some problem using technical means alone and spent a lot of time and money, and then spent some time creating some ontology and addressed the same problem in less time, there's an ROI. ToddSchneider: @David, You're correct about 'ROI of ontology'. Hence the focus on system. Systems use ontologies and semantic technologies; Systems have an understandable ROI. DavidPrice: @ToddScheider OK - I'll be interested in seeing how it goes. Getting ROI info is notoriously difficult. ToddSchneider: @David, I also interested in seeing how this goes. Lots of questions, confusion. RexBrooks: Also, since I am directly involved with using UML is SOA, specifically for the OASIS SOA Reference Architecture Foundation (SOA_RAF), I will be creating an ontology from the SOA-RAF and ensuring that it works, but I will inevitably restrict/qualify it as specific to the SOA-RAF and not SOA at large. Brian Lucas: Another aspect to consider is that to be truly interoperable, I believe the ontologies must themselves be re-usable and cross-referenced (and, ideally, reconciled into upper-level ontologies). I have recently come to the conclusion that upper-level ontologies are very necessary to interoperating lower-level ontologies, if we wish to have any hope of reconciling the separately-developed, domain-specific ontologies. And the connection of these "non-IT" ontologies may help drive the actual IT implementations (UML or otherwise). Perhaps the Grand Challenge track is the place for this thinking? YuLin: agree with Brian MikeBennett: +1 agree with Brian SteveRay: +2 on Brian's remark. I am arguing this very point in integrating the 70+ standards being developed to support the smart grid interoperability in the US. DavidPrice: @Brian Lucas : Upper-or-not is a huge debate and I'd be concerned about it being a on the critical path wrt a good way to 'Make the Case for Ontology' in the larger world. I've worked in both worlds and find strong advocates that disagree completely. Brian Lucas: @DavidPrice : I do not believe it should be on the critical path either. It is a personal interest of mine in the organization space, and I'm launching a non-profit in this domain, but I'll continue the conversation in one or more of the tracks. DavidPrice: @Brian Lucas, I'll be interested in following up on that with as part of the Summit. JimRhyne: @Nicola - unfortunately VMBO is a workshop format with unpublished proceedings. One has to commit to attending in order to benefit from the discussions. NicolaGuarino: @Jim: Yes, VMBO is just an informal workshop, but if you look at the organizers and the previous attendants publications you will find a lot of interesting material.... JimRhyne: @Nicola - agreed, been down this path. Not planning to attend this year. How can we get cooperation from the VMBO attendees to work on this problem? NicolaGuarino: @Jim: I'll talk with the VMBO organizers and let you know. I'll try to involve at least some of them (indeed I am one of them) RexBrooks: @Todd I sent you an email on my availability, just a heads up. ToddSchneider: @Rex, Thanks. I'll be in touch. DavidPrice: @SteveRay at al : Excellent topic for the Summit! MikeBennett: And I hope everyone will mark their diaries for next Thursday for the first in the Track 2 Applications and Case Studies with Mills Davis. PeterYim: Great session! LeoObrst: Thanks, folks, goodbye! JimRhyne: @Nicola - great - will follow up with you. MikeBennett: Thanks Peter and everyone. Looking forward to it all. PavithraKenjige: thank you! EdDodds: thank you all! PeterYim: -- session ended: 11:27 am PST --