ppy/chat-transcript_unedited_20120126a.txt Chat transcript from room: summit_20120126 2012-01-26 GMT-08:00 --------------- [09:06] PeterYim: Welcome to the = OntologySummit2012: Session-03 - Thu 2012-01-26 = Summit Theme: OntologySummit2012: "Ontology for Big Systems" Track (1&2) Title: Ontology for Big Systems & Systems Engineering Session Topic: Ontology for Big Systems & Systems Engineering - I : The Systems and Systems Engineering Problem Space Session Chairs: Dr. Matthew West Panel Briefings: * Mr. JackRing (OntoPilot, US) - "Toward a Unified Ontology for Systemists" * Mr. AnatolyLevenchuk (TechInvestLab, RU) - "Ontology Engineering for Systems Engineering" * Professor GiancarloGuizzardi (Federal University of Espírito Santo, BR) - "An Engineering Approach to Ontology Engineering in Complex Environments: the role of Foundational Theories and Ontological Patterns" * Dr. MatthewWest (Information Junction, UK) - "Model-based System Engineering" Session page: http://ontolog.cim3.net/cgi-bin/wiki.pl?ConferenceCall_2012_01_26 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: == . [09:25] anonymous morphed into TomTinsley [09:26] anonymous2 morphed into TimDarr [09:27] anonymous3 morphed into MattHettinger [09:27] anonymous2 morphed into RogerBurkhart [09:29] LeoObrst: Hi, all! [09:29] anonymous1 morphed into ChristopherSpottiswoode [09:29] anonymous1 morphed into NicolaGuarino [09:31] anonymous1 morphed into DougFoxvog [09:31] anonymous morphed into JosehSimpson [09:31] anonymous2 morphed into Ernani Santos [09:32] JosehSimpson morphed into JosephSimpson [09:33] anonymous morphed into Martin Gladwell [09:33] anonymous morphed into ReginaldFord [09:34] anonymous morphed into Anatoly Levenchuk [09:35] JoelBender1 morphed into JoelBender [09:36] NicolaGuarino: Note that if you are using Skype you HAVE to call the nickname "joinconference". Otherwise if you call a telephone number with Skype you cannot unmute yourself (this is what I discovered) [09:41] JackRing1: I am unmuted and speaking [09:42] JackRing1: I am on Skype [09:46] CoryCasanave: Skype "joinconference" does not seem to have a way to enter the conference code - no keypad. [09:47] PeterYim: @Nicola - my experience with skype is that you *can* do the mute/unmute with the "dial pad" (under the "call" dropdown menu) .. .amybe we are running different versions of skype ... they are definitely running different revision levels of their software depending on what platform - linux, mac, pc, ipad, etc. you are on [09:48] PeterYim: @Cory - can't find Skype Dial pad? ... it's under the "Call" dropdown menu as "Show Dial pad" [09:48] LeoObrst: What is "POSIWID"? On slide 3. [09:49] DougFoxvog: Why should simply adding cognates migrate a system from being deterministic to being non-deterministic? [09:49] AliHashemi: "The purpose of a system is what it does" [09:49] MatthewWest: @Leo: Purpose Of System Is What It Does [09:49] CoryCasanave: It seems to work ok dialing the phone # [09:49] AliHashemi: cf - http://en.wikipedia.org/wiki/The_purpose_of_a_system_is_what_it_does [09:50] NicolaGuarino: @Ali & Matthew: what it does or what it is intended to do? [09:50] AliHashemi: (if you follow the wiki link, it provides a very high level overview) [09:50] MatthewWest: What it does! [09:50] Anatoly Levenchuk: Purpose of a system what is does (what service it have) as a subsystem in upper level system. [09:50] JoelBender: @Nicola - what is does, maybe not very well :-) [09:51] NicolaGuarino: @Joel: hmmm.... [09:53] NicolaGuarino: Is a broken system a system? [09:54] CoryCasanave: Sounds similar to "SEMAT" http://www.semat.org started by Ivar Jacobson [09:54] Anatoly Levenchuk: @Nicola -- Yes, there is life cycle. Broken system on maintenance stage of it life cycle. [09:56] Anatoly Levenchuk: When we speak about what it does, that usually mean operation stage of it life cycle. And enactment of it service at this stage. [09:56] CoryCasanave: Also, thas may be applicable to: OMG RFP "A Foundation for the Agile Creation and Enactment of Software Engineering Methods" [09:57] Anatoly Levenchuk: @Cory: I prefer ISO 24744 instead of this OMG RFP. It more ontologically correct. [09:59] NicolaGuarino: @Jack: I don't understand the comparison between formal ontology and an algorithm [09:59] CoryCasanave: How can a request be correct or not? It would be the response that would be correct. [09:59] LeoObrst: I don't think there is a 1-1 relation between an ontology and an algorithm. [10:00] CoryCasanave: @Leo - Agree [10:00] LeoObrst: Oops: same question, Nicola. [10:00] AliHashemi: An algorithm could represent an implemented and operational ontology given some inference rules. [10:01] AliHashemi: could represent --> is analogous* [10:01] NicolaGuarino: formal ontology is a discipline [10:02] Gary Berg-Cross: With so many terms/concepts thrown around in the talk it would be nice to have atop-level, context diagram for what jack or others are proposing as this conceptual space. [10:02] AliHashemi: @Nicola - Jack's formulation was "situated ontology" [10:02] anonymous morphed into LinePouchard [10:03] AliHashemi: An algorithm represents the commitment of the programmers to what they believe exists in the scope of its execution. If the procedures within the program further represent operations and transformations (constraints on what is assumed to exist), then an algorithm can viewed as an operational ontology under some inference rules, no? (informal, implicit ontology) [10:04] Gary Berg-Cross: On the ontology - algorithm front, one could create an ontology to represent an algorithm as a process, as I believe that John Sowa has pointed to. [10:05] LeoObrst: An algorithm, by definition, specifies "how", whereas an ontology (like logic) specifies "what". I think that the algorithm must closely correspond to the semantics expressed by the ontology(ies), which indeed is hard to accomplish. Perhaps generating an algorithm from the semantics is the way to go, but of course is very hard. [10:05] NicolaGuarino: @all: let's just list the problems now, devoting a few chat interactions to each, otherwise we miss the whole picture presenters are trying to convey [10:06] AliHashemi: @Leo, one quick point.. In specifying the "how" you (implicitly) commit to the what. [10:08] NicolaGuarino: @Anatoly: I appreciate very much the contrast between ontology engineering and traditional mathematical tools [10:09] MatthewWest: Yes, Engineers are generally interested in mathematical rather than logical models. [10:10] GiancarloGuizzardi: @Matthew: the additional point is that they are frequently interested in mathematical models which are insensitive to true ontological notions [10:10] NicolaGuarino: @Anatoly: just to understand, is a method a *way* to achieve a certain function (e.g. cutting some materials by using lasers or scissors) [10:11] anonymous morphed into EvanWallace [10:11] MatthewWest: @Giancarlo: I agree with Ali on this. Mathematical moels have implicit ontological commitments. They may not be the ones that ontologists would wish them to make. [10:13] JackRing1: Which ontology? An ontology will be embedded in a system. Another will be embedded in the SE human activity system. A third ontology will be embedded in the SE learning environment. In a swarm of autonomous systems all three are inside the system. [10:15] NicolaGuarino: @ Anatoly: very interesting distinction between counterintuitive and folk ontologies. Still the objective in my opinion is being able to capture the actual language engineers use... [10:16] MatthewWest: I think you will find Anatoly and I would disagree with you there. What is more important is to have an ontology that is as acurate as possible to how things are, rather than to accurately reflect how people talk about things. [10:16] JackRing1: For example a software package Kollabnet prowls around in CAD files and extracts the terms and operands, etc., then helps organize a cross reference (semantic web) that shows the relationships and opportunities for parsimony. [10:18] NicolaGuarino: @Matthew: yes, but accuracy with respect to how things are is the goal of physics, photography, and so on... :-) [10:18] JackRing1: IN fact, one of the contributions of SE is to identify and resolve the gap between how things are and what things should be. Beer's POSIWID must be revealed. [10:18] LeoObrst: I still have an issue with "counterintuitive": perhaps it is naively counterintuitive, but doesn't at least some of the ontology become intuitive to the expert? [10:18] MatthewWest: @Nicola: And also ontology. [10:19] AliHashemi: @Leo, I think it raises an interesting question - how long does it take for counter-intuitive insights to become common sense? (I think this is what Anatoly was emphasizing.) [10:20] RexBrooks: Slide needs to be advanced. [10:20] JackRing1: As Will Rogers said, it isn;t what we don't know that hurts us, it is what we do know --- that ain't so. [10:20] AliHashemi: And it does point to important human factors issues in creating a system with high fidelity to reality, but also manageable for the end users. [10:21] JackRing1: Any ontology must be vetted as fit for purpose. [10:21] PeterYim: @10:20pst - we are on Anatoly's slide#8 now [10:21] NicolaGuarino: @Matthew: if we limit ourselves to describe (accurately) what things ARE we have no way to express how we want to use them for specific purposes [10:22] MatthewWest: @Ali: That is a good point. When I first came across 4D ontologies, I understood it, but found it very difficult to put into words. These days I hope I can speak about it more or less as it was an everyday idea. It takes time. [10:22] Gary Berg-Cross: What is formal pragmatics? Need more of a sense of this and an example. [10:22] MatthewWest: @Nicola: Intentions are also something we can talk about in terms of what they are. [10:23] DougFoxvog: When you need knowledge at different levels of granularity, why not use different ontologies for the different levels? Some ontologies would be far more stable than others. [10:23] AmandaVizedom: Nicola: IME, one very significant division of ontology applications falls long whether they (are required to) model (a) some slice of the world, (b) information artifacts about some slice of the world, or (c) both. In all three cases, the ontology models the thing, support reasoning about the thing, and supports manipulation of the thing in various ways and degrees. IMHO, a great many cases are of type (c), but developers think in terms of modeling (a) or (b), and not always the right one, and different requirements and methods fit those two objects. [10:24] JackRing1: The ontology of units of measure is traceable to standards and basic science. It can be considered 'truth' at least to earth-bound users. In contrast the term "vigorously" in an ontology is moderated by situation (we must accommodate Zadeh's fuzzy logic). [10:24] DougFoxvog: If you can model a heuristic, you can ontologize it. If you can't define the heuristic, then you can't ontologize it. [10:26] MatthewWest: @Jack: Actually the ontology of units is surprisingly shakey. It turns out that the standards can be interpreted (deliberately) in more than one way to avoid significant differences of opinion aboiut what units are and how they are used. Fortunately no buildings will fall down as a result of these differences. [10:26] AliHashemi: @Nicola, I would suggest those are two distinct issues. What we want (intention), vs perhaps common but inaccurate views of the system. I suspect Anatoly's point emphasizing counter-intuitive-ness is about the latter. [10:26] LeoObrst: @Anatoly: Category theory indeed is focused on structure, as opposed to set theory, and provides you ways of relating structures more generally, but multiple logics (and both their syntaxes and semantics) can be represented. Perhaps that is what you mean? [10:27] JackRing1: Anatoly: Is TRIZ an ontology? [10:27] NicolaGuarino: Nice idea of extending enterprise service bus to systems engineering [10:27] JackRing1: Matthew: Whether buildings fall down the fact was that a spacecraft crashed on MARS. [10:28] NicolaGuarino: (but I would drop the "smart" adjective, to many smart things are being advertised nowadays...) [10:28] MatthewWest: @Jack: that was simply not using the same units in different system. A much simpler problem (ontologically) than what a unit of measure is in the first place. [10:32] anonymous morphed into Victor Agroskin [10:33] MatthewWest: Welcome Victor [10:34] PeterYim: @Giancarlo - when you get a chance, please supply me with a slide deck on which slide are numbered (so I can swap it in). Thanks. [10:36] JackRing1: MatthewWest: Not different system. It was using an attribute value in one program that was expecting the number to be in the English system but was given a number in the metric system. An ontology spanning both systems would have noted the difference [10:37] LeoObrst: @Jack: I think TRIZ could be formalized as an ontology. [10:39] JackRing1: @Leo, I tried to do this in 1992 with RDD-100 Software Engineering tool but got swamped with other tasks. [10:41] anonymous morphed into Victor Groskin [10:42] Victor Groskin morphed into Victor Agroskin [10:42] JackRing1: SE must presume that two or more people constructed the system model and that they did not have a coherent weltanschaaung or even lexicon. Also, that Model(x) of one system and Model(y) of another system must be harmonized if you intend to make these subsystems of a third system. [10:44] MatthewWest: @Jack: Why more than one? [10:46] Victor Agroskin: Some ontology can be deduced from TRIZ. But the major value of TRIZ is a method, thus you have to choose some method ontology (like ISO 24744) and combine it with domain ontology - if you want to have a formal model of TRIZ. [10:49] JackRing1: Giancarlo's patterns are equvalent to my modularizations. [10:49] EvanWallace: Jack: I think that Matthew's point was that it wasn't an understanding of the notion "unit" that was a problem, but rather false assumptions about which units were being used. Yes. These sorts of false assumptions happen when you don't make units an explicit part of your model. So many would agree that there is value in defining and using an ontology of quantities, units, and measures, but the problem that Matthew mentioned about the ambiguity of the references for these things makes it more difficult to get consensus on *one* such ontology. [10:49] MatthewWest: @Jack: That is a good link to make. [10:50] JackRing1: @MW: The human mind cannot discrimminate reality from illusion. Takes two the untangle. [10:50] CoryCasanave: The use case being presented by Giancarlo is the subject of an OMG RFP: http://www.omg.org/cgi-bin/doc?ad/11-12-10, of which Giancarlo is a participant. [10:51] Gary Berg-Cross: Agree on the point of conceptual models being improved by by formal ontology distinctions. [10:52] anonymous morphed into Victor Agroskin [10:52] JackRing1: Isn't this Panel scrubbing concepts into a 'formal' ontology? [10:52] MatthewWest: @Jack: So if I am on a desert island, I don't know if it is real or a dream? [10:53] CoryCasanave: @Matthew, perhaps some people have more trouble with reality :) [10:54] DougFoxvog: @Matthew: If you think you are on a desert island, it may be real, a dream, or some other misconception. [10:56] PeterYim: @Matthew - please watch the clock [10:56] JackRing1: @MW. We have been over this before. Pls explain why witnesses to an event describe it differently. Pls explain why design reviews of system concepts always find fundamental logic or referent errors. [10:57] NicolaGuarino: @Giancarlo: distinguishing modeling patterns from analysis patterns sounds intriguing (and new), but I am not sure I understand what analysis patterns are, in practice [10:57] NicolaGuarino: I have to leave now, bye everybody. Great session! [10:58] Gary Berg-Cross: Also have to leave now... [10:59] MatthewWest: @Peter: Would it be better to drop my presentation in order ot make time for discussion? [11:00] ChristopherSpottiswoode: Bye from me too - thanks to all! [11:01] PeterYim: @Matthew - that's a thought but, it would be your call ... picking up from next session (with Henson presenting his bit is not a bad idea) [11:02] MatthewWest: @Peter: Yes that makes sense. I have one story to tell, and I can do that on the list. [11:02] PeterYim: @Matthew - since you cannot be with us next week, I definitely would want to hear your portion of the presentation [11:03] JackRing: @Evan. Not quite. The error was in not addressing units in the design model. The presumption "...false assumptions about which units were being used" is not correct because there was not consciousness of 'which' [11:03] JackRing3: @Giancarlo, For the enterprise ontology let's start with "objective" and "goal" [11:04] anonymous morphed into ReginaldFord [11:05] LeoObrst: @Giancarlo and all: I've always found some confusion between domain specific languages and ontologies. I personally think that ontologies need to provide the semantics for those DSLs, no? [11:06] CoryCasanave: Perhaps we should support "multiple inheritiance" of track topics [11:06] GiancarloGuizzardi: @Jack: these are very important and interesting notions. I have been interested in them for a while myself and have done some work in that direction. If you are interested, I would be happy to shared them with you [11:06] CoryCasanave: @Giancarlo, please post reference to the group & seminar you mentioned. [11:07] DeborahMacPherson: Great presentations! No questions but fascinating presentations [11:07] GiancarloGuizzardi: @Leo: yes, fully agree. In the ideal case, the metamodel (representing the worldview) behind a DSL should be isomorphic to the ideal domain ontology of the domain [11:08] LeoObrst: Will design patterns, analysis patterns, etc., be ontological constructs (with rules)? Are there as yet repositories for these? [11:10] TerryLongstreth: My principal concern about the combining of tracks 1 and 2 is the loss of discussion of emergent behaviors (since they are in my opinion, by definition, un-engineered) We've tried to finesse this question by expanding the notion of engineering to include any system with sentient inputs into its manifestations, but that seems to be to be a copout. [11:12] GiancarloGuizzardi: @Todd: if you are interested, I can send you the references to UFO. [11:12] LinePouchard: @everyone: I am collecting ontologies for units at present. If anyone would like to send me links, I'd be happy to examine them. I'd like in particular ontologies of units in OWL or that can me translated into OWL. Thanks [11:13] LeoObrst: @Todd: can you place your question in the chat room? So we have a textual record? Thanks! [11:13] LinePouchard: I forgot to say, you can mention them here or send me private email. [11:13] PeterYim: @LinePouchard - see: http://ontolog.cim3.net/cgi-bin/wiki.pl?UoM [11:14] AliHashemi: I'm curious to know the response to Gary's question re "Formal Pragmatics" [11:14] BobbinTeegarden: @Giancarlo, please send refs to UFO to all [11:16] Anatoly Levenchuk: Formal pragmatics -- http://en.wikipedia.org/wiki/Universal_pragmatics [11:17] Martin Serrano - DERI: Bit elaboration on finding out Federation of systems and Information modeling will be healthy to get into the real meaning.. True is Federation is more than a logic or instrumentation for modelling methods, [11:17] CoryCasanave: @Line, OMG has adopted but not yet published a date/time ontology which includes units. The ontology is expressed in OWL, UML and SBVR. [11:17] GiancarloGuizzardi: @BobbinTeegarden: The foundational work of the structural part of UFO can be found in www.inf.ufes.br/~gguizzardi/OFSCM.pdf [11:18] GiancarloGuizzardi: @BobbinTeegarden: this has been used as a foundational for the modeling language which now has been dubbed OntoUML [11:18] LinePouchard: @Cory: do you have a time frame for when it's available? [11:18] LinePouchard: @Peter: thank you [11:20] GiancarloGuizzardi: @BobbinTeegarden: parts of the Event and Social fragments of UFO can be found in http://www.inf.ufes.br/~gguizzardi/Archimate's%20ME%20semantic%20investigationTR20110610%5bCameraReady%5d.pdf (analyzing the goal modeling extension of Archimate), http://www.inf.ufes.br/~gguizzardi/Using%20a%20Foundational%20Ontology%20for%20Reengineering%20a%20Software%20Process%20Ontology_cameraready%20(1).pdf [11:21] BobbinTeegarden: @Giancarlo: Thank you, more on OntoUML? [11:21] GiancarloGuizzardi: @BobbinTeegarden: the last one is an example of its use in analyzing a Software Process Domain Ontology [11:21] CoryCasanave: @Line, very soon - I can provide the document which is being prepared for publication. [11:21] DougFoxvog: @LineP: http://forge.morfeo-project.org/wiki_en/index.php/Units_of_measurement_ontology#Measurement_Units_Ontology_.28MUO.29 [11:23] LeoObrst: @Anatoly: I agree that formal pragmatics (presuppositions, implicatures, speech acts, etc.) is needed, i.e., interpretation of the semantics in context and with respect to use, although I am not sure about Habermas and his "Universal Pragmatics". Also, epistemology must figure in: different belief stances. [11:23] CoryCasanave: @Line, the lead on the date/time ontology is Mark Linehan, IBM: emaiL: mlinehan at us dot ibm do com [11:24] FabianNeuhaus: @Cory, Line the OMG date/time ontology also contains CLIF axioms [11:24] CoryCasanave: @Fabian, sorry for the ommission? [11:27] FabianNeuhaus: @ Cory, I just thought that I mention it since the CLIF axioms are probably better suited to understand the underlying model than OWL [11:28] Anatoly Levenchuk: @Leo -- formal pragmatics (that is slightly after Universal Pragmatics that is more philosophycal by nature) is more about logic than linguics. While my friends linguists wonder that contemporary logic branch of it is differ from linguistic one, while inherit most of terminology :-) [11:28] TerryLongstreth: @Matthew - JPL = NASA Jet Propulsion Laboratory? [11:28] GiancarloGuizzardi: @BobbinTeegarden: Sorry for the delay...An approach based on OntoUML used at a systems engineering department at the US DOD is described in (http://www.omgwiki.org/architecture-ecosystem/lib/exe/fetch.php?media=dmg_for_enterprise_ldm_v2_3.pdf). [11:28] GiancarloGuizzardi: @BobbinTeegarden: I will send more information in a second... [11:30] CoryCasanave: @Fabian, yes - the CLIF is very precise in date/time. [11:31] LeoObrst: @Anatoly: I am a linguist/semanticist originally and think of formal pragmatics mainly from that perspective. In the ontology world, this is mostly addressed via formalized use cases, competency questions, which I admit is really just the beginning. [11:31] GiancarloGuizzardi: @BobbinTeegarden: I have input a lot of information on specific parts of OntoUML in the following OMG SIMF forum:http://www.omgwiki.org/architecture-ecosystem/doku.php?id=uml_based_data_modeling_for_an_enterprise_data_model (see the lower part on discussions) [11:32] Dickson Lukose: thank you! [11:32] AliHashemi: thank you all! take care. [11:32] LeoObrst: Thanks, Matthew and all! [11:32] Dickson Lukose: bye [11:32] GiancarloGuizzardi: Thanks a lot Peter, Henson and Matthew. very interesting discussions [11:32] GiancarloGuizzardi: bye everyone [11:34] PeterYim: @TerryLongstreth - ref. the change of Track labeling should not affect the conversation (at least not the way we have seen conversations direct themselves on the mailing lists all along) ... I think combining the tacks helps people who are confused as to what track their conversation belonged to, in the first place [11:35] DougFoxvog: @Henson: For the strange life of system components, an ontology could represent the model of the system, the physical components that fill the roles of the different components of the model, and temporary and permanent IDs for the physical components. With such an ontology, the various aspects you referred to on slide 7 could be [11:38] Anatoly Levenchuk: @Leo -- formal pragmatics is branch of philosophical logic, ontology is another branch. They are siblings on knowledge tree :-) [11:34] PeterYim: -- session ended: 11:33am PST -- -------------