chat-transcript_20100114f_edited.txt // PeterYim: . Welcome to: OntologySummit2010: Panel Session-2 - "Training Content for the Future Ontologists" - Thu 14-Jan-2010 * OntologySummit2010 Theme: "Creating the Ontologists of the Future" * OntologySummit2010 Panel Session-2 Topic: "What's in the Brain of an Ontologist -- The Strawman's Proposal" * Co-chairs: Professor MichaelGruninger (University of Toronto, Canada) & Dr. LeoObrst (MITRE, US) * Panelists: o Professor BarrySmith (University at Buffalo, US) o Professor ChristopherMenzel (Texas A&M University, US) o Dr. PierreGrenon (KMi, Open University, UK) o Dr. RobertStevens (University of Manchester, UK) ''(in absentia)'' o Dr. AldoGangemi (ISTC-CNR, Italy) . please refer to session page at: http://ontolog.cim3.net/cgi-bin/wiki.pl?ConferenceCall_2010_01_14 . anonymous1 morphed into EdDodds anonymous morphed into JessicaTurner anonymous morphed into ChuckTurnitsa anonymous1 morphed into MatthewWest anonymous morphed into BradleyShoebottom anonymous morphed into ElizabethFlorescu FabianNeuhaus: Peter I am on the phone anonymous morphed into TimWilliams anonymous morphed into BobbinTeegarden MichelleRaymond: I've called in twice. Very garbled. Are others having difficulty with Audio? FabianNeuhaus: Michelle, audio seems fine to me anonymous1 morphed into JoanneLuciano Lisa Zilinski morphed into LisaZilinski MichelleRaymond: After Peter mutes us, I'll see if it improves and will then retry if neeeded. SteveRay: @Michelle - you are loud and clear on my end as well. MichelleRaymond: Whew, third time's the charm. Just called back and all clear. Thanks. anonymous morphed into SusanMatney PeterYim: Michael is taking us through the Strawman Proposal at: http://ontolog.cim3.net/cgi-bin/wiki.pl?ConferenceCall_2010_01_14#nid26HW anonymous morphed into JulitaBermejoAlonso ArturoSanchez: @(Leo+Michael): first recommendation is to clarify if the goal is to create a Body of Knowledge or Curricular Recommendations, or both. As mentioned before, the recommendations put forward by ACM/IEEE-CS contain examples of how to put together both. ArturoSanchez: @(Leo+Michael): question ... it is not clear to me why Prolog is presented as a knowledge representation language (not a big deal) ... I'd just like to know the rationale ... RaviSharma: Domain of Applications or Domain of Ontologies RaviSharma: What is Enterprise ontology? is it a structure of organization or all cells of Zachman framework? JoanneLuciano: Ontology Evaluation? BradleyShoebottom: Yes, on computer science. I am a historian SteveRay: @Arturo, for the record could you provide a link to the ACM recommendations? RaviSharma: Why not FGDC Metadata and also OGC (geolocation standards) related ontologies such as Geophysical and geo databases, areas e.g. Dr. Rob Raskin's interests areas at JPL, MSFC etc. GiancarloGuizzardi: @(Michael+Leo): following Arturos proposal. One very interesting source of inspiration is the "Guide to the Software Engineering Body of Knowledge (SWEBOK)" - http://www.computer.org/portal/web/swebok. ArturoSanchez: @Giancarlo: tante grazie! MatthewWest: I like everything I see on this list, but I think at least one thing is missing, and that is something of philosophical ontology. What I think is needed is something of the basic ontological distinctions and commitments, e.g. between classes and particulars, between 3D and 4D individuals, modal logic and possible worlds. I don't think you need to be able to practice philosophical ontology, but you do need an appreciation. GiancarloGuizzardi: @Arturo: prego MatthewWest: Sorry I have to leave now. I will try to come back later. ArturoSanchez: @SteveRay: the ACM/IEEE-CS Recommendations can be seen at http://www.acm.org/education/curricula-recommendations JulitaBermejoAlonso: Personally, I miss the use of ontologies for software engineering, as part of the Applications in the proposal RaviSharma: Giancarlo it takes us to IEEE Education homepage?? PeterYim: the person who just type his/her input into the "hand-queue space" ... you can take that out by pressing the "hand" button again ... and please type your input into the field left of the "send" button (and then click "send") ToddSchneider: If we're also addressing the engineering of ontologies, then the strawman list is deficient w.r.t. engineering. GiancarloGuizzardi: @RaviSharma: its supposed to take us directly to the SWEBOK homepage (inside IEEE) JulitaBermejoAlonso: Thanks, my mistake JoanneLuciano: I'm wondering if there it wouldn't be good (tagging onto Barry's points) a module on ontology communities of practice / logistics / ontology development in practice in the communigty. ArturoSanchez: @(Leo+Michael): within Ontological Engineering, I wonder how the actual incremental and cooperative building of ontologies with 'domain experts' is reflected ... RaviSharma: giancarlo: but it does not, can you provide the direct link GiancarloGuizzardi: perhaps this one: http://www.computer.org/portal/web/swebok/htmlformat GiancarloGuizzardi: or better: http://www.computer.org/portal/web/swebok/html/contents GiancarloGuizzardi: does it work? ArturoSanchez: @RaviSharma: there is an extra period in the URL that does not belong ... http://www.computer.org/portal/web/swebok GaryBergCross: Would like to see more on the lifecycle methodology to build and maintain "ontologies". RaviSharma: Integration: So where are we with respect to ariving at integration standards, methods, examples of ontology integration) ArturoSanchez: @(Leo+Michael+ChrisMenzel): maybe we need to include 'Interoperability', which includes 'Integration' EdDodds: fwiw - If folks use their wikiname name format here in the soaphub chat when Peter imports the transcript into the wiki after the event, your name will take folks directly to your profile when clicked RaviSharma: Integration: There are also iter-and- intra domain (application)integration methodologies. ToddSchneider: FOL is a knowledge representation language for Foundation of Mathematics. JoanneLuciano: Agree with need for uses/application of ontolgies (integration) This will feed into Evaluation later. Also, in applicaitons - how to determine if one can re-use an ontology for a new appplication RaviSharma: Another suggestion, are we aiming at further drill down in the content of the strawman? This way we could create onto-education profiles and types of courses and contents as well as examples? JoanneLuciano: (I need to go to the CSHALS organizers meeting) Conference on Semantics in Healthcare and Life Sciences (CSHALS) http://www.iscb.org/cshals2010 AmandaVizedom: @Michael: regarding how to incorporate feedback: I'm thinking that this discussion should be reflected in the Requirements Survey, in addition to today's discussion and the post-analysis you and Leo may be planning. Some of it already is, but I'm picking up some additional questions and choices to add. Not to the fine-grained level of specific languages, but to the question of what knowledge and skills ontologists need now. GiancarloGuizzardi: Regarding foundations, I believe that we should have represented in a curriculum issues such as Meta-Level Criteria for classifiying ontological choices such as: Revisionary vs. Descriptive, 3D vs. 4D. But also basic issues such as the epistemological foundations behind ontological foundations (realism, moderate constructivism, radical constructivism), as well as basic notions such as Ontological Commitment ToddSchneider: What is the role of the ontologist, as being considered in these discussions, with respect to the development of systems (that would use their work product)? RexBrooks: I'm wondering why, if we are focusing on how to develop the ontologists of the future, we are not outlining the values we would like them to have; such as focus on accuracy, ability to prioritize the development of ontologies for practical uses, etc? The point is what values do we want these ontologists to reflect in their work once they learn the things outlined in the strawman? RaviSharma: Giancarlo: these would fall into categories such as reasoning, ontology models? GaryBergCross: Ontologies can help with data integration, but also better knowledge engineering so its connections to those methods should be part of the training and not just the knowledge representation part of KE. AmandaVizedom: Meant to add, for all: this brings up the point that some of the questions and topics in the surveys and discussions may appear to overlap, but here we can see how they differ and complement each other. Michael and Leo have carefully scoped this strawman and discussion to view things through lens of training content, and the survey segment now out also views things through this lens. The Requirements survey (coming soon!) will take a POV more situated in the ontology development and use world, asking what training is needed from that perspective. Both POVs will inform the full-scope collaboration of figuring out what training for ontologists should cover. GiancarloGuizzardi: @RaviSharma: the first are basic meta-level criteria for classifying (and choosing) between foundational ontologies. The latter are general basic foundations on the field. These are not orthogonal though AmandaVizedom: @Leo: music to an epistemologist's ears! ToddSchneider: It may be necessary to come to a consensus as to the roles and responsibilities of someone who undergoes the proposed training. RaviSharma: Leo: I would suggest similar to what I did in the reply to Giancarlo a while ago on Epistemology, words that explain Axiology and if you permit from other philosophies (non-Greek) notions of wisdom, knowledge, understanding, cognition-verification, etc. EdDodds: Mind-mapping tools maybe... ArturoSanchez: Need to step out for a while ... GaryBergCross: Yes, we should including methodfs across the ontology lifecycle - knowledge elicitation, structuring and ontology development, validation etc, . The structuring procedure is the one people usually focus on in ontological engineering. anonymous1 morphed into JimDisbrow AmandaVizedom: @Todd: In case Michael and Leo don't have chance to catch up with chat comments real-time: because that would be such an enormous task, and probably not one we could do while respecting the multidisciplinary nature of the field, the idea here instead is to come up with a fairly complete set of content-bins -- for example, things that might be turned into modules in variuos ways for various purposes. RaviSharma: Leo and mike: How do we balance and parse out the various portions of the Strawman? FOr example into categories of future ontologists such as : 1. Philosophical fouindations of Ontology, 2. Ontology languages. 3. IT tools for ontology 4. Domain specialist ontologist etc. RexBrooks: @Ravi: Let's not leave out the application and societal results (benefits). RaviSharma: All: We need to discuss when we can reach ontology reasonability test? For example if I have deevlped triples and serach criteria that give me the short list better than taxonomies give, is that enough of a result or we have better pass or filtering ideas? AmandaVizedom: @Fabian: Yes! I might suggest, correspondingly: Classic approaches to big modeling issues, especially competing approaches, from where they have grown, and the tradeoffs of each. MichelleRaymond: @ Rex and Amanda: The skills developed though addressing the training content must produce both 1)capabilities of an ontologist (model generation, understanding of semantic and syntactic choices on ontology development, mapping, ...) and 2) characteristics of an ontologist (produces consistent information structures, keys in on relevant existing ontologies for reuse or extension, results both human and machine readable...). RaviSharma: Rex: those were in 4.? however left out Societal results that will probably provide new ways of collective knowledge, analyses etc. yes important. RexBrooks: @ Ravi: That's the what I was reaching for with my comment on the values we want to develop in our future ontologists. MichaelUschold: On the topic of uses or applications of ontologies, see this paper: http://www.cs.manchester.ac.uk/~horrocks/Teaching/cs646/Papers/uschold99.pdf RexBrooks: Thanks MichaelUschold, I just downloaded the paper. FabianNeuhaus: Good point by Giancarlo: ontology life cycle management should be added to the list ToddSchneider: Amanda, having a complete list of topics and how they relate is essential, but the "purpose" also needs to be explicated. MichaelUschold: In addition to epistomology, noted by Leo, a carriculum should also include content on semiotics, distinguishing between types, tokens and meanings. PierreGrenon: @Fabian: the initial curriculum should reflect ontology life cycle RaviSharma: I suggest adding UML (ODM) to the languages and Financial Services to the Applications list in the Strawman RaviSharma: Great link Michael it also supports my sugestion on languages: Ontology representation languages- (e.g., UML, Express, Ontolingua, XML) Knowledge interchange languages: (e.g., KIF, PIF[7], CDIF) Translation tools: (e.g., Ontolingua translators, CDIFtools, StepTools, ... (lots!)) Distributed Objects: (e.g., CORBA, COM) RexBrooks: Semiotics yes! EdDodds: What about developing some social media mechanism which nudges institutional shareholders to ask (during quarterly analysts concalls) if C-Suites actually know what an ontology is and whether they use them in their enterprise? JulitaBermejoAlonso: I suppose all you already know a book on Ontological Engineering that reviews most of the topics in the Proposal. A good starting point... at least for me from scratch AmandaVizedom: On earlier point regarding evaluation: complementary but slightly different-angled point: Content is thin here, but it is still worth covering: how to develop operational ontology tests. That is, taking Fabian's point about staring at axioms and not understanding, testing and evaluation are not only ontology-internal matters. Where ontology is developed as part of a system (including not only the technology, but also the processes to be supported and the people involved in those processes), the effects of ontology modeling choices on the function of the system should be tested. This area is underdeveloped, but *thinking* about it can nevertheless be taught, along with examples where available. RaviSharma: MichaelUschold this could be required reading or reference as it deals with integration: and many other areas Figure 4: Data Access via Shared Ontology GiancarloGuizzardi: Another point regarding "domains of applications"...one point I see missing there is the ontological Analysis, (Re)design and integration of Modeling Languages and Reference Models (e.g., ISO standards). This has been one of the first application areas of (philosophical) ontology in computer science and there is still a very active community on the use of ontological theories as foundations for conceptual modeling languages and methods AmandaVizedom: @Todd: Can you say more about what you mean? I am thinking that the purpose would be more specific to a particular program or curriculum, whereas this would be more raw material for forming such a thing to serve a purpose. Is that different from what you mean by "purpose?" anonymous1 morphed into AldoGangemi AldoGangemi: I'm in, still on time for contrib? PeterYim: Community input are solicited through: (1) participation in virtual panel sessions like this, (2) through the [ontology-summit] mailing list, (3) through the wiki pages marked "Community Input", and (4) by responding to the surveys that are coming out ... see details in the Resources section on today's session page at: http://ontolog.cim3.net/cgi-bin/wiki.pl?ConferenceCall_2010_01_14#nid26H4 ... and the OntologySummit2010 Homepage at: http://ontolog.cim3.net/cgi-bin/wiki.pl?OntologySummit2010 RaviSharma: Michael:when did you publish this valuable paper? AmandaVizedom: Thanks for paper, Michael, very helpful! GaryBergCross: The analogy to SW engineering also provides the idea of a difference between a "programming" course oten using a particular language and a more general SW Engineering course where one learns more abstract concepts. The same very rough course distinction may apply here. RaviSharma: Amanda -I would like to continue to echo the same. ToddSchneider: Amanda, this goes back to a request I made of Barry, What's an ontology for training ontologists? The inclusion of a subject or module should include a justification, and in that justification a purpose, goal or intent of including the module I would expect to be embedded. MichelleRaymond: http://ontolog.cim3.net/cgi-bin/wiki.pl?OntologySummit2008_QualityAndGatekeeping 2. The ontology is expressed in a formal language with a well-defined syntax. (1E8F) 3. The authors of the ontology provide the required metadata. (1E8G) 4. The ontology has a clearly specified and clearly delineated scope. (1E8H) 5. Successive versions of an ontology are clearly identified. (1E8I) 6. The ontology is adequately labeled. MichelleRaymond: My previous cut-paste comments are from OntologySummit2008 and identify characteristics of an ontology. This can lead us to what must ontologist training include. RaviSharma: Leo: one of the topics we discussed was that for OOR? similarly Staging and E-Commerce etc. Does this get employed in Enterprise ontologies? GaryBergCross: A tutorial on ONTOLOGICAL ENGINEERING by Asunción Gómez-Pérez had these topics: Reuse and Sharing Problems in Building KBS from Scratch Problems when you reuse/share knowledge in KBS The Knowledge Sharing Initiative Definitions of Ontologies Ontological Commitments Components of an Ontology Types of Ontologies Libraries of Ontologies What does an explicit ontology look like? Principles for the Design of Ontologies Ontologies versus knowledge bases Uses of Ontologies RaviSharma: MichaelGruninger: UML and XML? FabianNeuhaus: I would suggest to restrict the languages to languages with a model theoretic semantic MichaelUschold: How to include content on ontology applications? Have a module that examines a small number of diverse case studies. For each, consider the original competency questions for the ontology, what services the ontology provides, where it fits into the architecture of the production software, how the ontology evolved over time, etc. This is a big picture view. Another way to include content on ontolgoy uses, is to take a larger number of case studies and analyze the different benefits that the ontology supported. RaviSharma: Gary: Real good List certainly candidate for courseware and also specialization. RexBrooks: Do we want to include the likes of SPARQL, when thinking about practical uses. We're planning to use it in specifying and querying emergency management policies for distribution of messages as well as within various SOAs. AmandaVizedom: @Rex: Good question, especially as many people are now using SPARQL and SPIN to work around limitations in OWL. RaviSharma: Leo: Model theoretic Semantics is another area. AmandaVizedom: Under Ontology Lifecycle > Design, and in tune with Barry's point about considering the humans involved: Existing training sometimes includes the formal aspects of requirements analysis, but I've seen too many "trained" ontologists with no idea how to gather requirements from human users or for an application context that includes human users. [I mean this separately from the topic of knowledge elicitation techniques and skill, though there is a corresponding lack of training there). RaviSharma: MOF and RDF are two meta-concepts that map into the laguage as well as interoperability considerations? anonymous1 morphed into BartGajderowicz RaviSharma: Aldo: I agree with your expressions. [ ppy comment: believe Ravi was referring to Todd and not Aldo ] RaviSharma: We definitely could be helped if we had a decision tree that is ontologydriven (reasoning based on interest of the learner-student)? FabianNeuhaus: Suggested answer to Todd's question: after the training the person should be able to deploy a high-quality ontology (in cooperation with domain experts) AldoGangemi: how to unmute? PierreGrenon / MichaelGruninger: *3 AldoGangemi: i still cannot unmute, maybe skype is an issue? GiancarloGuizzardi: Aldo: probably it wont work on skype AldoGangemi: i cannot digit the star on skype MichaelUschold: Aldo: i had this problem, get a new version of skype, it handles this. GiancarloGuizzardi: Aldo: I had the same problem in the past RexBrooks: When using Skype you may need to try to get the numberpad in the Skype dispay, you can use your cursor to press *3. AldoGangemi: i am using the cursor, but skype does not do anything after a * pressing AldoGangemi: and i have a quite recent version of skype, maybe the mac version has issues RexBrooks: That was my best suggestion, sorry. If the computer numberpad or numbers don't work. AldoGangemi: ok i'll do the chat, better than nothing AldoGangemi: My main points: AldoGangemi: the strawman is mostly irrelevant for what we want to know about the mind of an ontologist PierreGrenon: Aldo, it's about the brain RaviSharma: Aldo: mind is a whole new area and yes it does relate to cognition at very deep theoretical and philosophy-logic levels. PeterYim: @MichaelGruninger ... I'll unmute ALL for Aldo RexBrooks: @Aldo; this is in line with my earlier comments on values. I think we need an ontology of values which could be studied by the future ontologist early on as a way to navigate the bewildering mass of categories in the strawman. AldoGangemi: yes, that's what i wanted to say MatthewWest: I'm afraid I need to leave now. Good discussion. PeterYim: thanks, Matthew! GaryBergCross: Developing an ontology of the ontology topic is too hard. Let's just build a taxonomy for it. MichelleRaymond: @Gary, I think we're starting with just a List. MichelleRaymond: Hierarchy is what is forthcoming JulitaBermejoAlonso: @Todd: I had gone through a on-my-own training on most of the topics in the proposal. As a result, an awful lot of information and background that had to be filtered and traded off to accommodate the domain ontology I wanted to develop. Most of the topics suggested in the proposal provided the means. Not the actual expertise to develop the ontology. AmandaVizedom: I like that suggestion very much. GaryBergCross: A curriculum view seems approapriate to organize "Training Content for the Future Ontologists" but it doesn't say what courses we would divide the content into and how to sequence them. MichelleRaymond: Crossing Automated Reasoning, Ontological Engineering, KR Languages, and Existing Ontologies -- Use case: Ontology technology is critical in development of a situation understanding system for interactive real-time reporting using open standards. Providing key data for situation understanding requires: gathering relevant information, formatting information for exchange, transferring information to users/systems, and presenting information to users. Continued situation awareness requires: data updates, sharing information, associating new data inputs. For example, "Effective emergency response requires organized notification and up to the minute situation awareness throughout the emergency. An added challenge is consistent presentation of building geometry and semantic information using accessible, broadly adopted, easy to understand, open formats that are accurate and maintainable over time." - MacPherson, Raymond Ontology training should address meeting the needs of this use case, (and certainly many others.) GiancarloGuizzardi: I am afraid I have to leave now. Thanks for the great discussion. PeterYim: thank you, Giancarlo GiancarloGuizzardi: bye, Peter SteveRay: I also must leave early. Thanks. PeterYim: bye, Steve RaviSharma: Amanda: I would be happy to work with you for such a survey although we can also work with Arturo? RaviSharma: Amanda: I see the difference between your and arturo's areas. AmandaVizedom: @Rex: I'd agree that many (especially OWL-focused) newer ontologists focus too much on classes of objects and not enough of relationships (not so true of those who've worked in more expressive languages, but that's not the majority). However, I'd argue that seeing this in terms of verbs and nouns leads to more problems, as it conflates the linguistic objects and the conceptual objects. IMHO, separating those is *essential* to having an ontology that can support uses / users across languages. In fact, it's essential even to reusability across community-specific dialects. RaviSharma: next advance will be strengths of relationships and some relevance ranking for the desired outcome. GaryBergCross: ON the relation to language DOLCE aims at "capturing the ontological stands that shape natural language and human cognition." They are clear on the assumption/commitent that the surface structure of "natural language and the so-called commonsense have ontological relevance. As a consequence, the categories refer to cognitive artifacts more or less depending on human perception, cultural imprints and social conventions." RaviSharma: especially Leo and Jim PierreGrenon: @Amanda second this PierreGrenon: which is also why teaching ontology should not start with OWL or worse Protege.. AmandaVizedom: @Pierre: indeed! PierreGrenon: @amanda: and only those with security clearance ought to get taught linguistics RexBrooks: @Amanda: I don't disagree. MichelleRaymond: @Pierre: agreed! That is certification of a tool or language NOT knowledge of "Ontology" AldoGangemi: i start my tutorials by teaching linguistics and semiotics it works AldoGangemi: thanks all PeterYim: thank you all ... great session ... bye PeterYim: - conference call session ended 12:26pm PST - AmandaVizedom: I take Gary's point about respecting the cognitive significance of language. I think, though, that these are not in conflict if the support for concept-to-linguistic object (multiple) mapping is rich enough. RaviSharma: Thanks to all. GaryBergCross: Aldo Thanks for the detail on how you start the course. can you provide an Outline? PierreGrenon: @aldo: you mean you manage to teach them linguistics PierreGrenon: @amanda: linguistics ought to be the ontology of language JoanneLuciano: I've had to be on another call and didn't realize this was still going on. GaryBergCross: There's a very practical side to some "linguistic analysis" that ilustrates underlying concepts we have about rality. Even Object-oreirnted analysis used these. JoanneLuciano: Has there been any discussion of an 'agile' or incremental approach to teaching ontology engineering? PierreGrenon: not enough about the incremental EdDodds: Thanks all! The Best Buy reference was a helpful commercial case study GaryBergCross: Joanne, do you mean the use of an incremental method to build ontologies? JoanneLuciano: yes Gary JoanneLuciano: And along those lines, in some efforts I've been involved in, we've been using "minimal information...." to define the first set of terms to implement so that the ontology can be used in short order GaryBergCross: Well I agree with importance of incremental scoping and focused improvments. But we left such details out along with top-down versus bottom up approaches and much more on the methodology side. JoanneLuciano: I would like to see a ciriculum that covers theory and practice - and practice in more than one setting GaryBergCross: I can imagine an intro course that combines these and then advancd courses for each. GaryBergCross: Have to go, bye all. -- end of chat session -- //