ppy/chat-transcript_unedited_20120202a.txt Chat transcript from room: summit_20120202 2012-02-02 GMT-08:00 ---------------------- [08:25] anonymous morphed into MarcelaVegetti [09:16] anonymous morphed into SteveRay [09:20] PeterYim: Welcome to the = OntologySummit2012: Session-04 - Thu 2012-02-02 = 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 - II: a response to the problem space and setting out the working program for this Summit Track Session Chairs: Dr. Matthew West Panel Briefings: * Dr. LeoObrst (MITRE, US; Ontolog) - "Aspects of Semantic Interoperability" * Mr. VictorAgroskin (TechInvestLab, RU) - "Big System Models" * Mr. ChrisPartridge (BORO Solutions, UK) - "Systems Engineering (SE) practice issues and SE (IT) issues ontology can help to solve" * Mr. DavidLeal (CAESAR Systems, UK) - "Ontologies and engineering analysis" * Ms. ElisaKendall (Thematix, US; OMG) - "Design For Analysis" * Dr. HensonGraves (Algos Associates, US) - "Triage on the 'Big Systems and Systems Engineering' Track" Session page: http://ontolog.cim3.net/cgi-bin/wiki.pl?ConferenceCall_2012_02_02 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: == . [09:20] anonymous morphed into MarcelaVegetti [09:23] anonymous morphed into DavidLeal [09:23] anonymous1 morphed into MarkJohnson [09:23] anonymous morphed into ElisaKendall [09:23] anonymous1 morphed into BobbinTeegarden [09:25] anonymous1 morphed into MatthewHettinger [09:25] anonymous morphed into TomTinsley [09:27] anonymous morphed into Onno Paap [09:29] anonymous morphed into ChristopherSpottiswoode [09:33] anonymous morphed into GiancarloGuizzardi [09:37] PeterYim: == LeoObrst presenting ... [09:37] FrankOlken: Leo, Yes, we can hear you. [09:39] JackRing: Enterprise has coherent purpose. Community is non-deterministic [09:46] anonymous morphed into NicolaGuarino [09:47] DougFoxvog: Re commerce. Electronic Data Interchange (EDI) systems have standardized electronic messages with thousands of defined terms. This has been done without ontologies. [09:47] SimonSpero: [Leo slide 8]: Often more communication/interoperability between similar bits of different organizations than within the organization (intraoperability?) [09:48] SimonSpero: @DougFoxvog: right - there are many EDIs [09:49] SteveRay: The more similar the contexts, the more you can do without ontology and simply rely on shared implicit meaning. So ontology really comes into its own when you integrate different contexts. [09:49] anonymous morphed into TimWilson [09:50] DougFoxvog: @SteveRay +1 [09:50] PeterYim: == VictorAgroskin presenting ... [09:50] James Odell: [Leo slide 9] - not just "swarm" agents, rather multiagent systems, in general. [09:51] SimonSpero: (incf @SteveRay): There is ontology *and* vocabulary bound into the domain. This is recognized in contract law, where terms may have different meanings to standard definitions based on usage of trade (which must be proved up) [09:51] ToddSchneider: Steve, your assertion presumes that parties involved in the communication are aware of the shared contexts. For technology mediated interactions contexts need to be made explicit. [09:52] LarryLefkowitz: @SteveRay: Allegedly shared info often reveals subtle differences that can throw significant wrenches into the works. i.e., What do two organizations (often with in the same enterprise) mean by "employee"? Do they include contractors? Part-time workers? Temp workers? Or is a "fork-lift" a vehicle? etc, etc, etc. [09:55] SteveRay: I agree with all these responses. I was pointing out the continuum: If you are in exactly the same context, you can rely on pure syntax and your shared understandings. The further apart two parties get in context, the more misunderstandings you will have without an ontology to make your respective interpretations explicit. [09:56] SimonSpero: [Agroskin slide 3]: The cat is an information object (when identified as a cat). Briet, Buckland, etc. [09:57] AmandaVizedom: RE Leo's slide 8 (the mapping nightmare): Indeed. In fact, it's much worse; single, large enterprises typically have multiple semantics within their own information and systems. One way to identify and begin to understand these is to look at functions within the enterprise. For example, the enterprise may have a focus on production of Widgets, but some sub-organizations within the enterprise will specialize in Widget quality management, others in the Widget-part supply chain, others in Widget sales, others in Widget-industry business intelligence, others in Widget customer support, others in management of Widget-specialists and other personnel... Typically, each of these sub-units has its own systems and data, and contexual assumptions are implicit. When the enterprise aims to become more efficient and effective by sharing and reusing information and tool between sub-units, those implicit contextual assumptions collide; that is! a manifestation of the differing semantics along activity lines. So, often, the "mapping nightmare" is repeated within the organization, as well as across enterprises. On the plus side, there may be shared semantics between similar activity sub-units of different organizations. [09:58] ToddSchneider: Doug, yes EDI systems work. But participants that use them (implicitly) agree a priori to the (static and non-explicit) semantics/interpretations. This example brings out the temporal aspects of system design and development: How flexible must the system be? As Leo suggested to be interoperable is not a static relation (in a net-centric environment). [09:59] LeoObrst: @Doug: yes, but even the EDI community knows EDI's many limitations, hence the rise of UBL and electronic commerce ontologies (such as we had at VerticalNet) (and my wife was an EDI analyst). Controlled vocabularies are a step up, but require everyone involved to have the common semantics in their heads or readable in natural language, and is intensely human-in-the-loop. [09:59] AmandaVizedom: Are we on slide 6 now? [09:59] AliHashemi: yes [09:59] AmandaVizedom: @Ali thanks. [10:01] MikeBennett: @Amanda that raises an interesting question. For any one context (whoch for instance is the context of one application) there is no real need to segregate second order 'thing in a role' from first order things. There is only one context, so the first and second order concepts are the same. As soon as one needs to do an ontology that covers several contexts, each context is that third order construct in which a given set of second order concepts have their meaning, and within which the first order things fulfil those roles. So there is a need for partitioning for complex ontoloigies, which cannot be understood by (or justified to) the people who only need to ontologise something in one single context. [10:01] ToddSchneider: Leo, a question I asked Pat Hayes (and didn't get an answer), for the domain of 'engineering' is ontological dependence (operationally/effectively) equivalent to modal necessity? [10:02] JackRing: I note seeming interchange class and type. SE's need to discrimminate between endogenous attributes vs. exogenous attributes of a system, e.g. class and type. What am I missing? [10:02] LeoObrst: @Victor: on your slide 7: how about Process Specification Language (for resources across manufacturers, etc.), the old FIPA agent ontologies, etc.? [10:04] DougFoxvog: I certainly do not claim that EDI systems are anywhere near perfect. However, enterprises have been communicating with commercial EDI messages for decades. They have certainly sped up business communications and avoided much of the ambiguity of natural language. [10:05] LeoObrst: @Todd: I don't think so necessarily ;) I think you can define dependency in FOL. But more expressivity as in modal logic can help, including of course deontic logic with obligations, prohibitions, etc. [10:06] Anatoly Levenchuk: to slide 9 of Victor Agroskin -- my slides about ontology engineering for PLM systems: http://www.slideshare.net/ailev/ontology-engineering-for-systems-engineering-low-hanging-fruits [10:07] TimWilson: Slide 9 is where I am with my workplace. Management doesn't want to talk about ontology, but we are gearing up to port a proprietary PDM database into a PLM (Teamcenter, to be exact). Whether or not Siemens uses ontology under the hood is irrelevant to getting the capabilities to trace requirements to design to verification and validation to maintenance. [10:07] JackRing: How do you know that the valve installed was actually a Siemens.... (vs. some bogus copy?) Where does the ontology describe identification confirmation data? [10:07] PeterYim: == ChrisPartridge presenting ... [10:07] AmandaVizedom: @JackR and @MikeB: Actually, I think those points are related. One interesting phenomenon as you get into the complex, multiple-perspective (multiple-activity) exchanges is that the foreground and background move. That is, in one community of practice (or use context or application), certain things may be thought of and treated as primary objects and others as variable properties or roles, while in a different community (context, application), within or outside the same enterprise, the focus is reversed. IME, modelers within any such community tend to treat whichever is primary in terms of classes, and whatever is secondary in terms of roles, properties, or attributes. [10:08] PeterYim: @Anatoly - your slides for last week and this week's session are accessible at: http://ontolog.cim3.net/cgi-bin/wiki.pl?ConferenceCall_2012_02_02#nid345M [10:08] GiancarloGuizzardi: @ToddSchneider: ontological dependence is sometimes used as a synonym for existential dependence, sometimes as a catch all name for historical depencence, existential dependence, generic dependence, etc... [10:09] MikeBennett: @Amanda that's interesting, I'd like to know more about this (not now of course). [10:10] GiancarloGuizzardi: @ToddSchneider: in any case, these are modal notions in my view (and yes, we can characterize them in classical FOL as usual with reification of worlds) [10:11] MikeRiben: how do you pull of the slides out of the zip..which file do you start with [10:11] Anatoly Levenchuk: @Peter. Thank you! [10:12] AmandaVizedom: @JackR and @MikeB: This can be handled and mapped quite easily in an expressive representation language, but modeling personnel often don't know that. This is OT, but deals with knowing your requirements and the corresponding requirements of personnel. I have worked with a shocking number of "ontologists" initially trained in domains that regard specific alignments of classes to specific kinds of kinds, etc., and who have no logic training, in consequence of which they waste a lot of time trying to adjudicate between class and property representations that are completely logically equivalent and can be appropriately mapped! [10:12] SteveRay: @MikeRiben: Are you trying to download the slides? You can download each slide presentation separately, without using the zip. [10:13] MikeRiben: when I click on his presetnation, I get a zip with xml..might be I am using an old browser..let me try with chrome [10:13] SteveRay: Try going to http://ontolog.cim3.net/cgi-bin/wiki.pl?ConferenceCall_2012_02_02#nid32YC to get the individual slides [10:13] MikeRiben: that was it... [10:13] LeoObrst: @Amanda, etc.: yes, think of a supply chain. Chemical manufactures focus more on physical properties, purity of chemicals. However, their downstream supply chain paint manufacturers bring in new properties such as drying time, light reflectivity, maintenance half-life, etc. [10:14] AliHashemi: MikeRiben who's slides are you trying to access? [10:14] MikeRiben: I was using IE 8 ( my enterprise sanctioned browser!)and it was taking the pptx file and downloading as a zip so I saw all the xml files that new pptx has..ugh! [10:18] BobbinTeegarden: @Amanda @JackR Classes of classes, systems of systems, etc are fractal, i.e. tree structures; but the complexity we're trying to address is graph shaped. Could this be some of what we're struggling with? [10:19] anonymous morphed into Ilya Zaslavsky [10:19] DavidLeal: have typed *7 [10:20] AliHashemi: on the dial pad? your skype microphone itself may also be muted [10:20] Victor Agroskin: @Leo We are starting from workflow engines of PLM/PDM vendors, they are quite exotic usually, and always influenced by some theoretic paradigm - unfortunately these ranges from IDEF to BPMN and beyond. It is not yet clear for us whether an ontology of any particular standard will be enough to achieve interopeability in one case we have -- with one Owner, 2 Designers and 2 procurement and construction companies. We are taking elements from DEMO, BPMN, SEMDM - to map aspects of their processes, practices, legacy docflow rules. [10:20] ChristopherSpottiswoode: Mike, rename the .zip files to .pptx [10:20] SimonSpero: [legacy systems] - this a job for "semantic shims" [10:20] AmandaVizedom: @Leo, etc: Exactly. And more: suppose they have a unit that manages batch manufacture. Now, new properties related to handling and roles in manufacturing processes may come into focus. Moreover, that process-management group's own systems may well treat the *processes* as primary and the inputs and outputs as roles. When integrating information from local systems, chemicals may be treated as classes in the order data system, for example, and property values in the process data system. [10:21] LeoObrst: @ChrisP: your slide 7, concerning different kinds of decomposition: partially dependent on perspectives, i.e., roles? [10:21] LarryLefkowitz: @Leo: To the extent that these foci are truly disjoint, then integration - other than some shared concepts to ensure each is talking about the same paint/product (for instance) -- may be straight-forward. However, it is often the case that there will be interactions across them -- such as the light reflectivity depending on the chemical composition -- that it would a shame to inadvertently duplicate. It is also common for some third perspective -- e.g., a market analysis that might include both customer demand for a certain paint type as well as availability or cost of certain chemicals -- to leverage each of these nominally separate areas. [10:22] PeterYim: == DavidLeal presenting ... [10:23] JackRing: Eng. Analysis is properly part of V&V, particularly if V&V properly focuses on Likelihood of System Effectiveness. [10:23] BruceBray: @MikeRiben I had the same problem with IE 8 downloading pptx files as xml zipped - you can just change the file extension to .pptx when saving the download and that works. [10:25] JackRing: It isn't about right or correct. It is about Fit For Purpose. [10:25] DougFoxvog: I have another meeting now, but will stay logged in so that i can read the chat log when i return. I'm hanging up now. [10:27] TrishWhetzel: @ChrisP WRT to semantic legacy modernization, do you have examples of groups/projects that have moved from UML-based software infrastructure to an ontology-based infrastructure? [10:27] AmandaVizedom: @Bobbin: That is often a problem. And in fact, I have at times tried to fight the ontology version of creeping features by arguing that indeed, what people often think of (at first) as a single "domain" is in fact fractal. The further down you go, the more distinctions you will find, and it is easy to get stuck in a cycle of endless refinement. Fortunately, there are ways to address this. One of them is to have a particular purpose to your modeling, and to take time to go through a requirements specification stage before starting to ontologize! This is underdeveloped in the ontology field, but very important to success, IME. It's also important to notice that a major benefit of ontologies is that the declarative, logic-based nature of them makes the addition and extension of models *as needed* technically easy. [10:27] JackRing: standardization or at least harmonization of analysis data is ripe for ontology. [10:29] TerryLongstreth: It strikes me that one metaphor for engineering analysis is forensic medicine. Are autopsies better documented? [10:30] JackRing: System Engineering makes sure that the gradients on selected relationships throughout the sysem model are 'filled in' only by design analysts, not the designer. [10:30] JackRing: Autopsies are past tense. Analysis predicts exogenous behavior. [10:30] SimonSpero: [Archival recording of analysis] - Where does e.g. HDF5 fit in as a base? [10:31] SimonSpero: see - http://www.hdfgroup.org/HDF5/ [10:32] Anatoly Levenchuk: Ontology for engineering (dynamic) simulations used in https://www.simantics.org/ [10:32] TerryLongstreth: Perhaps autopsy was disingenuous. How about medical diagnosis? I note that the medical community is trying very hard to standardize their ontologies. [10:33] JackRing: Yes, and prognosis is the objective. [10:33] JackRing: If you don't know the metrics needed how do you decide which analyses to do? [10:35] RexBrooks: I'm wondering whether the images David is explaining are just bitmaps as opposed to vector drawings that can carry the metrics as well as scaling up and down. I would have thought the manufacturing, materials industry is using vector drawings as graphics. [10:35] LeoObrst: @David: how much does analysis use simulations? [10:37] AmandaVizedom: RE: DavidLeal's slides 9 & 10 point to the importance of provenance in information analysis activities. I think that many kinds of analysis match what is described here, in that there is benefit to representing the objects of analysis, the content of analysis, and the process of analysis, and to include or provide a path to artifacts that were used (e.g., the mentioned pictures). The ability to reason and query through all of these levels has benefits for analysis itself, for evaluation and quality of analysis, and for end-user trust of the analysis. [10:41] JackRing: @Amanda: Quite so. And the ontology for documenting all SE activities and decisions. After all, system engineering, the activity, amounts to a system in operation (that generates a pattern for the realization of another system). [10:41] DavidLeal: @SimonSpero: HDF5 fits very much with this paradigm. The data about a field consists of 1) what the field is, which needs to be expressed with semantic precision and in a way that is easy to query; 2) a definition of how data is interpolated between discretization points within the field - i.e. the mesh, which can be expressed in many different ways, and can possibly be encoded as a table in HDF5; 3) values at the discretization points for sequences of states, which is definitely where HDF5 fits. [10:42] ToddSchneider: Network operations/defense is an example of (continuous) large data that needs semantic interoperability across vendors. [10:22] PeterYim: == ElisaKendall presenting ... [10:42] LeoObrst: @Elisa: Would you include Siri in this area? I don't know if Apple is aggregating the data that Siri sends back to the mother ship, and analyzing it, etc. [10:42] JackRing: @Elisa, the advent of personalized medicine envisioions the better part of 7 billion people auto-reporting approximately 2K bytes per day. [10:43] NicolaGuarino: @Amanda: does your approach mean that we need to give systems the capability of being aware of themselves (and their customers)? A kind of introspection capability? [10:44] JackRing: @Nicola, If you want cybersecurity then system self-awareness is mandatory. [10:44] NicolaGuarino: @Amanda: I guess what you are talking (more or less) of autonomic systems... [10:45] PeterYim: == HensonGaves presenting; and moving onto the panel and open discussion ... [10:46] LeoObrst: @Nicola: that's very interesting, a kind of reflective capability, i.e., using meta-level reasoning? [10:46] DavidLeal: RexBrooks: I used the term image "deliberately", because the first thing that happens is that an analyst looks at an image of the field on a screen and says "OK - that looks reasonable". The analyst then extracts a handful of numeric values out of the many gigabytes, often on the basis of hot spots shown on the screen, as the values which are used for further calculations about a design. [10:47] NicolaGuarino: @Jack (&Amanda): Yes, cybersecurity is just one possible motivations for self-awareness. Anyway, self-awareness requires some ontology, I guess, so the need for self-awareness seems to me a clear reason for having (good) ontologies of these systems [10:49] SimonSpero: @Nicola: NASA's Deep Space One had a lot of introspection, but I don't know what kind of KR they used. [10:49] DavidLeal: LeoObrst: I do not distinguish between analysis and simulation at all. If you want to make a distinction, then perhaps a simulation can be regarded an analysis that looks at change with respect to time. [10:49] JackRing: describing an engagement is needed first, then rules of engagement. [10:50] AmandaVizedom: @Nicola: I'm cautious of using the language, but my answer is basically yes, often. Caveat: I'm strongly of the view that the requirements of the particular system should be considered, and we need not build skyscrapers when a hut will do. Given that, many systems I've worked on, from command & control to web search to vehicle maintenance management to semantic SOA for enterprise interoperability have the kinds of requirements that Elisa and David both discussed. At minimum, process and provenance behind system "conclusions" needs to be traceable and human evaluatable. Often, there is concrete benefit to going further, so that these can be reasoned over, backward and forward, for a range of purposes. [10:50] MikeBennett: Natural v artificial systems: there's two aspects of that: an ontology /of/ a natural system has additional considerations about ontological commitment, granularity and so on compared to an ontology of an artificial system - whether or not the ontology is subsequently used within an artificial) system or as part of the design lifecycle. [10:51] SimonSpero: or the vote button [10:51] LeoObrst: @Henson: I think Cory will be addressing federated systems (as a cross-track), and I'm sure he will be addressing semantic interoperability, etc. [10:52] anonymous morphed into ChrisPartridge [10:53] JackRing: List already available [10:53] SteveRay: Henson, you are suggesting we hold off further discussions of Natural vs. Artificial systems, yes? [10:54] MikeBennett: I have to drop off now. Thanks for a great session [10:55] AmandaVizedom: @Simon: re: "semantic shims" for legacy systems: I like this phrase. I'm not sure whether this is what you meant, but indeed you might so describe something I've seen used very successfully: ontological representations that of the semantics of specific information systems that are kind of lousy ontologies in and of themselves, but are used as stand-ins for legacy systems in mapping their data to ontologies that are engineered to support discovery, reasoning, and interoperation. [10:56] ToddSchneider: Have to go. Thank you. [10:56] PeterYim: bye, Todd! [10:58] SimonSpero: @Amanda: yes - though I also include in the shim actionable descriptions of how to perform actions on the data [10:58] JackRing: SysML, Modelica and OPM need formal ontologies Need modeling for context sensitive systems, e.g., OpEMCSS [10:59] LeoObrst: @Amanda, Simon: yes, I think these are often called "wrappers", since they intervene between a very messy system and (perhaps) a better ontology. You wrap the system that allows some declarative communication. [11:02] NicolaGuarino: @Peter: I am lost, are we discussing/voting on which topics we have to address? [11:02] GiancarloGuizzardi: My votes on topics to discuss go to: (i) components, functional roles, functional dependence, functional parthood, social roles, normative descriptions , perspectiles, qua entities, etc... I believe these are very important notions for engineered systems in general and for socio-technical systems, in particular; (ii) semantic interoperability...but yes, I agree with Leo that perhaps this could also be addressed on Cory's federated sytems track [11:03] Anatoly Levenchuk: my remark: we need professional methodologists (discipline of situational method engineering, SME) to describe systems engineering as a method and not only systems engineers and ontologists. We need to build community of methodologists, ontologists and systems engineers to have success. We started to build such a community here in Russia on a base of INCOSE Russian Chapter and ISO 15926 community and intensive usage of SME thinking. [11:04] ElisaKendall: @LeoObrst -- oops. Following up on your earlier remark, Siri is one of many applications on the iPhone, and I believe many phone home. One reason they do this is to provide more targeted advertising to their users - which helps them make $$$$. That's of course not the only reason. And they are by far not the only one - we have several clients and potential clients who want to do something similar. [11:04] PeterYim: @Nicola - we are accessing the refined list of worthwhile threads outlined in Henson's slide#3~7 - http://ontolog.cim3.net/file/work/OntologySummit2012/2012-02-02_OntologySummit2012-SolutionOpportunities/OntologySummit2012_BigSystems-SE-Track_Triage--HensonGraves_20120202.pptx - also temporarily available as http://ontolog.cim3.net/file/work/OntologySummit2012/2012-02-02_OntologySummit2012-SolutionOpportunities/HensonGraves_BigSystems-SE-Track_Triage_20120202a.pdf [11:04] SimonSpero: @leo: wrappers have connotations of elegance; shims sound cheaper [11:05] SimonSpero: @Elise: Siri does all processing on servers at base camp [11:05] HensonGraves: Please, lets decide which topics we want to pursue in this Summit, rather than just talking about interesting things. [11:05] LeoObrst: @Jack: I think human-centered design tries to address some aspects you mention. [11:06] ElisaKendall: @SimonSpero: other manufacturers are starting to do some on-board processing, including a client. [11:06] NicolaGuarino: @Jack: I agree very much on the importance of modelling human role in engineering systems, that's why I have been insisting on the ontology of socio-technical systems as a topic. [11:06] Anatoly Levenchuk: My vote to the list: please, define function and construction aspects of system decomposition. E.g. even in 4D ontology family there are different opinions about FunctionalObjects. And define something with modalities (for requirements and design variants). [11:07] RickSteiner: Does "semantic interoperability" address the analysis data disconnect issue that David L discussed? [11:07] SteveRay: Peter: Let's choose a spot on the wiki for the evolving list. [11:07] HensonGraves: Is any one willing to take the lead a thread.The leader is perfectly free to focus it as needed [11:08] RexBrooks: There has been a fair amount of interest in simulation, and I have a use for the results of a discussion, but no time to ride herd on a discussion. [11:08] JackRing: @Leo, Yea, but we need more than tries. Tempus is fugittin' c.f., www.starkermann.com [11:08] LeoObrst: I think we should probably exclude natural systems, simply to focus, and hope that teleological projection from artificial systems will help for understanding natural systems. [11:09] GiancarloGuizzardi: @Nicola: Yes. Ontology of Socio-Technical systems would be an excellent topic. Those things I mentioned would need to be considered in such an ontology. [11:10] DavidLeal: Recording the human role in engineering design and analysis processes is important. Design and analysis choices are made by people, and this needs to be recorded as part of the record of the process. It's part of the audit trail. [11:10] ElisaKendall: @Jack, to your point on medical systems, I believe you're right, and that the military is already funding wearable technology, at least at DARPA, to collect health information/medical information regarding "the warfighter", which is intended to be semantically enabled and processed. I'm not sure how much would be done on=board, likely little if any at first, but certainly the back-end requirements will be enormous. [11:12] Anatoly Levenchuk: @Leo. Distinction of natural systems and artifact is not nessesary for systems thinking. There are "stakeholder" that define function (and boundary) of a system. If you interested in rain of Solar system I simply ask "What you whant to do with that?". After this reasoning will be the same as with artifacts. [11:13] JackRing: My estimate of priorities: 1) Semantic I/O among people and machines. 2) Modeling human activity system, also called socio-technical systems, 3) Making SysML, Modelica, OPM and the one in Norway (mental blank) ontology-based but also modeling languages and tools for context-sensitive and autocatalytic systems. [11:14] LinePouchard: @Elisa: I worked on such a project for DARPA in 2004-5, it was called Virtual Soldier, collecting medical information and trying to do predictions for triage at field hospital. I don't know if this kind of research is still pursued. [11:14] ChrisPartridge: Need to go. Bye. [11:14] anonymous morphed into Julio [11:15] DavidLeal: @RickSteiner: Yes, but we need not do too much. Merely recognising that a loading path, which an analyst has just spent the last two months analysing, is the same as a particular state transition defined by the system model is enough. There is lots of other semantics which we need not worry about integrating. [11:15] Julio morphed into JulioNardi [11:16] LeoObrst: Ravi Sharma is on the phone now, but I don't see him in the chat room. [11:16] PeterYim: @Henson - ref SteveRay's suggestion - I have posted your list at: http://ontolog.cim3.net/cgi-bin/wiki.pl?OntologySummit2012_BigSystemsEngineering_CommunityInput#nid345U ... we can all review and edit that [11:16] NicolaGuarino: I have to go folks, bye... [11:16] SteveRay: For the record, Ravi Sharma has volunteered to champion a thread on semantic mapping among languages. [11:17] AmandaVizedom: @Anatoly: +1 regarding the use of stakeholders to define function and boundary of system. This also serves to define function and boundary of supporting ontology: represent the breadth and depth and distinctions relevant to what you want to do, taking some care to do so in a way that leaves open the possibility of unanticipated extension and mapping. [11:17] ElisaKendall: @Henson, there was a DARPA project that dealt with various aspects of systems engineering and design, in the late 1990s, that may be a source of relevant papers. It was called RASSP -- rapid prototyping of application specific signal processors, and my role on one of two teams was to develop the requirements for and prototype a library of design information across 26 cad vendors whereby one could deal with vocabulary clashes as well as semantic mismatches. See http://www.atl.external.lmco.com/projects/rassp/RASSP_legacy/ and look for applications including the reuse library for more on that. [11:17] AliHashemi: My vote for list prioritie: 1)Composite Systems 2) System Descriptions 2) Success and Relevance of Semantic Issues in Eng [11:18] LeoObrst: @Steve, Ravi: I think that is semantic interoperability. [11:18] GiancarloGuizzardi: Semantic Mapping across languages is an aspect of the language federation problem we have discussed last week and which will probably be addressedin Cory's track [11:19] GiancarloGuizzardi: but yes, this is an aspect of semantic interoperability and one which is fundamental for Enterprise Modeling [11:19] LarryLefkowitz: Dropping off audio.. still on chat for the reaminder of the call [11:20] LinePouchard: @Ali: My vote follows your list exactly - but I have to go. [11:22] SimonSpero: [how to tell when an ontology is complete enough] - This fits in to the quality and metrics cross-track [11:23] AliHashemi: i've got to run. thanks. [11:23] DavidLeal: @SimonSpero: I agree - important. [11:24] JackRing: We must be prepared to highlight and discard current conventions which are too incoherent with systems speak. c.f., DODAF [11:24] AmandaVizedom: I am unsure how much of the thread described on Henson's slide 5 intersects the Quality cross-track. I think some but not all. I'd be happy to take suggestions, comments, thoughts regarding issues under the slide 5 topic that are at least significantly issues of ontology quality, metrics, and evaluation. We can use those suggestions to prioritize issues to cover in that cross-track. [11:25] GiancarloGuizzardi: Folks, unfortunately I have to leave now. thanks for the interesting discussions. bye [11:25] PeterYim: @Amanda - I captured that - http://ontolog.cim3.net/cgi-bin/wiki.pl?OntologySummit2012_BigSystemsEngineering_CommunityInput#nid345Z [11:26] AmandaVizedom: @Simon and @David: I agree, and will make "how to tell when an ontology is complete enough" into the quality cross-track focus. [11:26] PeterYim: bye, Giancarlo ... thank you! [11:27] AmandaVizedom: @Peter: Super, thanks. [11:27] SimonSpero: [federated semantic systems engineering] - There are aspects of engineering federated systems that are not currently being addressed in some of the federated semantic systems that have been shown. For example, if one system is not reachable, or has low bandwidth, you need TMS [11:29] FrankOlken: I think that adding cybersecurity to the tracks will expand the scope too far. [11:30] BobbinTeegarden1: @Jack Cybersecurity certainly is an immediate issue for federation, probably on Cory's plate [11:30] SimonSpero: Cybersecurity has come up in a few places; it is key to handle federation [11:30] SimonSpero: @bobbin: jinx [11:32] LeoObrst: @Jack: I agree cyber-security is important, but could be under Track 4, e.g. [11:32] LeoObrst: Folks, I must leave now. Great session! [11:33] FrankOlken: I must leave for another meeting. Bye! [11:36] JackRing: Elisa, I am not addressing health monitoring. More on the quarterly collection of DNA and blood protien and deciphering these using a reference model to anticipated organic malfunction then to therapy selection model then to dosage intensity model. [11:36] RexBrooks: Unfortunately, I must leave, too. I look forward to the results of this discussion. [11:37] ElisaKendall: @Henson -- I can assist with a couple of these areas, but cannot take the lead due to family obligations [11:37] JackRing: @Henson, TKU for this. I think you have helped us get traction. [11:38] DavidLeal: I think that responding to an e-mail and volunteering for different tasks off-line will be good. [11:39] JoelBender: Thank you! [11:39] SimonSpero: Thanks to all - bye [11:39] DavidLeal: Bye [11:39] PeterYim: -- session ended: 11:39am PST -- -------------------