PeterYim: Welcome to the Ontolog Panel Discussion: Advancing the UoM_Ontology_Standard work to OASIS - Thu 19-Nov-2009 * Session Chair: Dr. FrankOlken (NSF) * Panelists: o Mr. HowardMason (BAE; ISO) -- UoM_Ontology_Standard_OASIS_TC_Charter_draft Adoption and Next Steps towards the OASIS QUOMOS TC Formation o Mr. EdwardBarkmeyer (NIST) -- UoM_Ontology_Standard_English_draft o Mr. DavidLeal (CAESAR) -- UoM_Ontology_Standard_UML_Model_draft o Dr. RobRaskin (NASA/JPL) [in absentia], Dr. JamesMasters (TopQuadrant) & Mr. HansPeter_de_Koning (ESA/ESTEC) -- UoM_Ontology_Standard_OWL_draft o Dr. PatHayes (IHMC) -- UoM_Ontology_Standard_CLIF_draft Please refer to details on the session page at: http://ontolog.cim3.net/cgi-bin/wiki.pl?UoM_Ontology_Standard/ConferenceCall_2009_11_19 anonymous morphed into DavidLeal anonymous morphed into ChipMasters HajoRijgersberg: Hi, my telephone connection is very bad. I try to understand what you are saying. FrankOlken: This is Frank Olken. I am on the line. I am having a hard time hearing folks - but I can hear Peter Yim okay. FrankOlken: I spoke yesterday at the CENDI/NFAIS/FLICC workshop on Semantic Web: Fact or Myth? in Washington, DC. FrankOlken: Oops, the workshop web page is: http://cendievents.infointl.com/cfn1109/ FrankOlken: Eventually, there will copies of the speakers' slides and perhaps video at the CENDI web site. PatHayes: sorry im late. EdBarkmeyer: Schedule: the first day after 15 December is 2 January PeterYim: if Jan-14 is too tight, Jan-21 or Jan-28 are both open too (if we still want to do it on an Ontolog Thursday event slot) MatthewWest: Who can be an "eligible person" to register as a participant? HowardMason: See the Charter page for a link to the OASIS procedures that define the terms explicitly FrankOlken: I have received authorization from Mary Ann Piette (LBNL) to act on behalf of Lawrence Berkeley National Laboratory FrankOlken: w.r.t the QUOMOS effort and TC charter. HajoRijgersberg: Do I understand correctly we can now make comments on the charter? FrankOlken: You will recall that I (Frank Olken) am employed by LBNL - but am detailed here to NSF. FrankOlken: Hajo, Yes. Raise your hand on the chat room, or send message to the chat room. EdBarkmeyer: I am pleased with the revised draft charter. All of our concerns are addressed. Thanks. EdBarkmeyer: NIST has formally approved our participation, that is, both the NIST OASIS principal and my management. FrankOlken: Thus the list on the charter is tentative list, until we get email confirming participation (proper confirmation). anonymous morphed into PavithraKenjige EdBarkmeyer: You can actually extend the supporting membership during the participation call period HansPeter_de_Koning: ESA is not an OASIS member and unlikely to become one any time soon. I can provide inputs as an individual but cannot act on behalf of an OASIS member organisation. FrankOlken: Perhaps we can get NASA to participate - they are members. ChipMasters: I am discussing membership options with our NASA sponsor for QUDT. Do you need a firm commitment for a specific individual (e.g me) or is a commitment from a member organization to participate sufficient? FrankOlken: EdBarkmeyers suggests: Paragraph 2d should be "the TC will seek to coordinate development with ...." FrankOlken: JamieClark: in paragraph 2a, "we will talk ... and attempt to coordinate" PeterYim: @Howard and Jamie: we have our five already - HowardMason (BAE), EdBarkmeyer (NIST), FrankOlken (LBNL), JerrySmith (DoD) and PeterYim (individual) HajoRijgersberg: Thank you, Frank. Yes, I think the draft is fine. I see many concepts mentioned that I think are important in the domain of units of measure. However, to get to the point directly, I argue measurement scales should be more prominent in the scope; converting temperatures (absolute or differenc) has a strong relation with scale refered to or unit. EdBarkmeyer: I have to leave now for the "dragoon" function. I am saddling up and grabbing my musket... FrankOlken: James Clark: "contributions" should read something else "contributions means that the TC may modify these items" PatHayes: hajo: your point will emerge in the work no mater what the charter says, so relax PatHayes: mater//matter JamieClark: thanks - Frank captured my second point; my first was to use something aspirational -- The TC will liaise with and consider coordination of its specification(s) with XYZ ... rather than the presceptive "development WILL be coordinated with ...". Thanks, apologizes that I must depart early. HajoRijgersberg: Yes, I'm relaxed, thank you. Just a little bit unused to a teleworkshop... HajoRijgersberg: OK, I just get to another point. In 1b it says that existing formal models for quantities and units are not quite comparable. I'm not sure that is true. More evidence should be provided. FrankOlken: Howard will edit the draft charter on the wiki. Peter will submit to OASIS. PeterYim: I'll be happy to do that FrankOlken: David Leal will now speak on the UML model of QUOMOS based on the VIM from his slides. HajoRijgersberg: Maybe it's more important that a new standard is shared widely. And that we try to accomplish that by the Ontolog discussion group. The different existing approaches may, in the future, be comparable, so that they can be used next to each other. FrankOlken: Summer discussion of quantity and kind of quantity overlap somewhat. PeterYim: @DavidLeal - ref. the mangled up slides, if you can supply a set in pdf later, I will swap that in (for the archives) HajoRijgersberg: Reminds me of a question I have (maybe a stupid one): are we going to continue the summer discussions? HajoRijgersberg: Systems of units. They should be in the scope too. JoeCollins: Quantity Kind, such as when two Derived Quantities with the same Quantity Dimension have different Quantity Kind, is rather ill defined in VIM. Making the definition clear is an endless, case by case task. HajoRijgersberg: Being able to set the value of a quantity, may be a requirement (i.e., part of the scope) too. It will have an effect on how quantities are modeled: as independent concepts or properties (referring to out summer discussions too). FrankOlken: We are now on slide 6 (aka 66) of David Leal's slides. HajoRijgersberg: No, it does not necessarily have to be difficult: different quantity kinds (such as breadth and width) can have the same dimension (length). FrankOlken: We are now on slide 7 (aka 77). PatHayes: I am finding this completely confusing (confused?) ArturoSanchez: %} JoeCollins: I believe that "mutually comparable" quantities, i.e., having the same Quantity Kind, means that they are referred to in some "Law of Physics", a meaningful equality or inequality within a scientific theory. ArturoSanchez: I wonder how you will be able to express concepts such as "equivalence relations" in UML ... JoeCollins: To H.R.: I didn't mean difficult, just endless. HajoRijgersberg: OK, but why exactly endless? PatHayes: i don't knowe what the difference is and i have no idea how to separate them. Why are we doing this? What is a particular quantity HajoRijgersberg: We're back in the summer discussions. JoeCollins: For example, there's no conceptual limit on the number of dimensionless quantities, many of which may be of the same kind. The Kind equivalence classes are resultant of scientific theories. PatHayes: i seem to be muted ArturoSanchez: :X PatHayes: yes im trying HajoRijgersberg: Yes, there's no conceptual limit on the number of dim'less quantities. We have to model them and assign their dimension "dimension one". HajoRijgersberg: I think it is very important to continue our summer discussions. HajoRijgersberg: My statement at the time was and still is that particular quantities, such as the length of my table, is an instance (or member) of length. HajoRijgersberg: And length is a quantity class. MarkLinehan: VIM 1.1 for "quantity" has "Note 1: The generic concept 'quantity' can be divided into several levels of specific concepts, as shown in the following table. The left hand side of the table shows specific concepts under 'quantity'. These are generic concepts for the individual quantities in the right hand column." MarkLinehan: (and a table follows, where the first line has length, radius, and radius of a circle on one row) HajoRijgersberg: So a generic level of quantities is "length", "mass", etc. and a specific length has representatives like "length of my table"? MatthewWest: Treating classes as instances, and then adding classification and specialisation at the instance level is something ISO 15926 did around properties MarkLinehan: yes -- but to be parallel to the note 1 that I cited, generic would be "length", an example mid-level quantity would be "width" (versus perhaps "depth") and a specific quantity would be "width of my table" HajoRijgersberg: Yes, Mark, that's what I mean. In my eyes this is a hierarchy of quantities. MarkLinehan: yes, I agree HajoRijgersberg: But it can of course also be modeled as Matthew says. HajoRijgersberg: But I expect that people are more familiar with hierarchies? HansPeter_de_Koning: To H.R: I agree with the need explicit "dimension one". That's how we do it in SysML QUDV. MarkLinehan: I suppose, but Matt's chart 4 defines 3 specific levels. I think Pat's point is that may be the wrong way to go. HajoRijgersberg: Yes, there may be more levels in some cases. There's no official limit to the number of levels. PatHayes: like sublasses, right? MarkLinehan: yes HajoRijgersberg: We have to model this for the most important quantities, starting with length, mass, time, etc. HajoRijgersberg: Yes, like subclasses, I would say. MarkLinehan: in VIM, those are "kind of quantity" and are standards HajoRijgersberg: Yes, but we should not define "kind of quantity" as a class. (I stated that also in our summer discussions.) FrankOlken: We are looking at http://ontolog.cim3.net/cgi-bin/wiki.pl?UoM_Ontology_Standard_OWL_Draft MarkLinehan: yes DavidLeal: For length as a class of particular quantity it is reasonable to have waist size as a subclass, because some particular lengths are waist sizes. For length as a class of magnitude of quantity it is not reasonable, because there is nothing special about 37 inches which makes it a waist size. HajoRijgersberg: Indeed, measures/magnitudes have no hierarchy. FrankOlken: HansPeter - if understand this, simple units and base units are differentiated in order to deal with the fact that HajoRijgersberg: As to waste size: we can have a class "Quantity", with subclass "Length", with subclass "Size" on its turn, with subclass "Waste size" on its turn. This is a nice example. FrankOlken: base unit for SI mass dimension is the kilogram? HajoRijgersberg: Base units are a property of a system of units. HajoRijgersberg: They are simple units, but derived units may also be simple units, such as pascal and newton. HajoRijgersberg: In the NIST standardadization document of Taylor, simple units are called "units that have a special name". Are we talking about the same thing here? FrankOlken: HansPeter - your hand is still up in the chat room ... FrankOlken: Howard Mason has updated the TC charter, send comments to him via email. HowardMason: *** ACTION item - the draft of the Charter has been completed, apart from the explicit statement from NIST. Please review and send any comments direct to me for inclusion. Signing off for tonight. HajoRijgersberg: OK, I see you are defining simple units as base units. Why? PeterYim: for our editors of the OWL ontology: if you are planning to do some comparison (and possibly mapping) between the various owl ontologies, may I suggest you upload them to the Open Ontology Repository instance (OOR-sandbox at http://oor-01.cim3.net ... also ref. http://OpenOntologyRepository.org which another community working in the Ontolog-CWE) FrankOlken: Pat Hayes is now speaking. HajoRijgersberg: Maybe a simple unit is a unit that is not defined in terms of other units but in terms of standard quantities? Why distinguish these kinds of units? They can just have a different definition. HansPeter_de_Koning: To HR: We define a SimpleUnit as a class. It is instantiated independent from any SystemOfUnits. We do not preempt SI. Then any SimpleUnit or DerivedUnit or ConversionBasedUnit (all subclasses of Unit) can be selected as a baseUnit in a SystemOfUnits. In QUDV baseUnit is an OWL object property with range=SystemOfUnits and domain=Unit. Similar for simple and base QuantityKind. See http://www.omgwiki.org/OMGSysML/doku.php?id=sysml-qudv:qudv_owl for details. HansPeter_de_Koning: Sorry, it should read "In QUDV baseUnit is an OWL object property with domain=SystemOfUnits and range=Unit". HajoRijgersberg: Your baseUnit property is great. Right domain and right range. HajoRijgersberg: But derived unit should also be such a property. Derived unit is system of units related. MatthewWest: I agree to rise above and give the capability to explain how different systems interrelate. DavidLeal: An issue is how a unit relates to a quantity. For length as a class of particular quantity the metre is a subclass. For length as a class of magnitude of quantity the metre is a member. Both views of length are useful. The first supports hierarchies, and the second is mathematically tractable being a 1D vector space for which the metre can be selected as a basis. PeterYim: @Pat: if it's doable, your "rise above them" (ref. your slide #1) approach would make a lot of sense, especially in light of the goal that we are not *just* trying to produce a good ontology standard, we also *really* want to be adopted and embraced by stewards of existing standard, as the overarching objective of having ontology-based standard as a first class citizen that will sit alongside (natural language) standards HajoRijgersberg: David, the metre can't be a subclass of quantity. It's a unit. DavidLeal: Hajo - why not? MatthewWest: I think that it is ok to differ from the ISO standards, as long as we can support the intent. But then there is an obligation to engage with the owners of those standards and try to achieve consensus with them. HajoRijgersberg: David, because we have the model the concepts in a clear way to all people. A metre is not a quantity, it's definition (or reference if you wish) is in terms of a standard quantity. And the metre isn't a measure either. HajoRijgersberg: Hans Peter, why distinguish SimpleUnit and ConversionBasedUnit? Do they have different properties? MatthewWest: I need to leave now. HansPeter_de_Koning: To HR: Yes. A ConversionBasedUnit has a referenceUnit object property with domain:ConversionBasedUnit range:Unit. HansPeter_de_Koning: I have to sign off now. FrankOlken: New tentative date is Jan. 21, 2010 for the first meeting of the QUOMOS TC. This is to accommodate OASIS scheduling constraints. HajoRijgersberg: Hans Peter, you'll maybe read this later: I would propose a property "definition" or "reference" with domain Unit and range Unit and Quantity (the latter for specific standard quantities). FrankOlken: Jan 7 Thursday willbe our next working teleconference. Expecting new drafts of texts, and models. PeterYim: good session ... thanks! ... bye everyone. HajoRijgersberg: Thank you for the organization! PeterYim: session ended 12:19pm PST