chat-transcript_20090619f_edited.txt VNC2: Welcome to the Ontolog Panel Discussion: Towards A Quantities and Units of Measure Ontology-based Standard - Fri 19-Jun-2009 * Session Chair: Dr. FrankOlken (NSF) * Panelists: o Dr. StanHuff (Univ. of Utah) -- "Experience and Requirements for Units of Measure in Patient Data Exchange" o Mr. DavidPrice (Eurostep) -- "OASIS PLCS Committee Requirements for a Units Ontology" o Mr. DaveMcComb (Semantic Arts) -- "The gist Unit of Measure Ontology" o Dr. RobertDragoset (NIST) -- "Units Markup Language" o Dr. RobRaskin (NASA/JPL) -- "SWEET 2.0 Scientific Units Ontology" o Dr. PatCassidy (MICRA) -- "Aligning NASA SWEET sciUnits Ontology with COSMO" o Mr. HowardMason (ISO, BEA) -- "Units of Measure - How many standards?" o Dr. PatHayes (IHMC) -- "Making Distinctions" VNC2: Please point your browser to the session page at: http://ontolog.cim3.net/cgi-bin/wiki.pl?ConferenceCall_2009_06_19 anonymous1 morphed into DaveMcComb anonymous1 morphed into DavidPrice FrankOlken: I am on the chat room now, will join the teleconference momentarily. DavidPrice: On call and chat DaveMcComb: I'm on the call (I think) anonymous1 morphed into Joe Collins anonymous2 morphed into Mark Linehan anonymous1 morphed into KurtConrad anonymous morphed into StanHuff, Yim, Mark Stan Huff, Yim, Mark morphed into StanHuff, Yan, Mark Chip Masters: ChipMasters My email is cmasters[at]topquadrant.com. James Davenport (OpenMath): James Davenport (OpenMath) here Mark Linehan: This is Mark Linehan from IBM (mlinehan@us.ibm.com). I'm leading a team at OMG that is attempting to come up with a model for Date and Time. It turned out that we needed a model for quantities and units of measure before doing date and time. So we have something to contribute for quantities & units of measure. anonymous1 morphed into Mitch Kokar FrankOlken: This is Frank Olken at the National Science Foundation. I will be chairing the teleconference once we are set up. FrankOlken: I suggest that everyone on the teleconference introduce themselves on the chat room discussion. anonymous1 morphed into SusanTurnbull Roger Burkhart: I'm involved in OMG SysML (UML for Systems Engineering) including a model for quantities and units FrankOlken: EvanWallace you are entirely silent .....??? anonymous1 morphed into Douglas Mann FrankOlken: We have two anonymous participants in the chat room. Please go to settings and change your login to your actual name. Laurent morphed into LaurentLiscia anonymous morphed into EvanWallace EvanWallace: Thanks Mitch FrankOlken: Note that due to the number of speakers for this teleconference we will ask each speaker to limit their remarks to 10 minutes. LaurentLiscia: On behalf of OASIS: DavidPrice and DaveMcComb: thanks for being on the call! JoelBender: Joel here, and in spite of being late, I'm glad I havn't missed it! FrankOlken: The first speaker will be StanHuff of Intermountain Health Care from Salt Lake City in Utah. FrankOlken: The second speaker is DavidPrice from OASIS Product Lifecycle Committee. FrankOlken: Dave McComb is the 3rd speaker of the teleconference, will speak on the gist units of measure ontology. SteveRay: Typhoid Mary must have been at the Semantic Technologies Conference. I picked up a bad cold there also! SteveRay: In DaveMcComb's work, there seems to be a blending of units and dimensions. Seems to me these are distinct concepts. FrankOlken: Dave, You need to wrap up your talk now. PeterYim: I doubt if Dave will be reading the chat screen while talking ... DaveMcComb: Sorry, you're right I have a tough time reading and speaking at the same time FrankOlken: Is BobDragoset on the call now? FrankOlken: We are passing over BobDragoset, who is apparently not on the call. FrankOlken: Our speaker is now RobRaskin, speaking about the NASA SWEET units ontology. RaviSharma: Welcome Dr. RobRaskin, Ravi here - great presentation. Thanks. FrankOlken: DaveMcComb, DaveMcComb: I did have a dimension class in the ontology a couple of years ago, but I found there wasn't much of a need for it. FrankOlken: @DaveMcComb, Please post the URL of the gist units ontology to the chat room discussion. DaveMcComb: The gist unit of emeasure ontologogy is at http://ontologies.semanticarts.com/gist/gistUOM.owl DaveMcComb: There is some documentation on gist at http://www.gist-ont.com FrankChum: Minimalist!!! DaveMcComb: As I look at it and think about it, I guess the dimensions just sort of folded into the subtypes of UnitsOfMeasure (I didn't really intend that but it sort of worked out that way)_ RaviSharma: Some confusion between Power as raised to Power in Math vs Power and energy flow measure. Can you kindly provide some other example as well next time? just a suggestion. but some notes may clarify. Glad to learn that JPL and ESIP are collaborating on these. FrankOlken: PatCassidy is now speaking on the alignment of the NASA SWEET Units ontology to COSMO .... SteveRay: @DaveMcComb: The reason that looks problematic to me is that there can be several different units to measure a given dimension. MikeBennett: @SteveRay - I think the concept of dimension is critical, and then common concepts like amount and quantity, before we even get down into specific domains. As Dave says, currency exchange rate has a time dimension. DaveMcComb: @SteveRay Yeah, but if they each have the same base (so if fortnights and hours each have "second" as their base unit, they will be inferred to be "duration" units) which I think gets the dimension idea, without having to have another class or property anonymous morphed into BobDragoset FrankOlken: @BobDragoset, We will have you speak after PatCassidy, who is speaking now. FrankOlken: Pat, We have a problem in that the base unit for mass in SI is the kilogram. RaviSharma: PatCassidy - If we could agree to represent the Units (generally) in form of tuples then perhaps time and calendar typpe conversions and ontological meanings would be clearer? JoelBender: An extension/implementation of the unix 'units' command to support this effort would be excellent. FrankOlken: HowardMason, from ISO will now speak. Joe Collins1: Actually, all units have time dependence, not just currencies WRT rates of exchange. It's just that rates of exchange change on a much shorter time-scale. anonymous morphed into BobDragoset Joe Collins1: ISO documents are not cheap. I suspect this is why people generally refer to abstracts of them. RaviSharma: HowardMason - often units are used across-communities, standards are followed cross communities. Ontology would allow reasoning to compare different ways of expressing physical entity such as energy and also different automation mechanisms based on language type etc. hence there is a cse for XML or ontological representation of units. FrankOlken: BobDragoset is now speaking on the UnitsML Units Markup Language. FrankChum: @HowardMason: Scope specific standards can prevent ontology from being unneccessarily bloated. HowardMason: I am merely keen to ensure that we avoid duplication and conflict FrankChum: @HowardMason, I concur! FrankOlken: What sort of units need non-integer exponents of the "root" units? RaviSharma: Frank - any measures of geometry such as PI! JoelBender: ah, that's how I get my warp factor, v = w^3 * c FrankOlken: @BobDragoset, You need to start wrapping up. anonymous morphed into Martin S Weber (NIST) MikeBennett: It seems to me there is a real art to not designing something. That's the difficult part of ontology. Joe Collins1: Rational powers of units & dimensions may be encountered in manipulating model equations. I would not want to put a finite bound on what is expressible. FrankChum: @BobDragoset, seems to me that the UnitML you presented is for the scope of physical science. Martin S Weber (NIST): @JoelBender (timestamps) : speaking about UnitsML: there will be something like the unix units command, and UnitsML does include capabilities for timestamps so this should well be possible. FrankOlken: PatHayes is now speaking on "Making Distinctions". Joe Collins1: Q@RobRaskin: What will be maintained by the ESIP Federation, just units and dimensions, or all of SWEET? Joe Collins1: Q@ BobDragoset: What tool did you use to represent your schema diagrams? Martin S Weber (NIST): @FrankChum: The initial scope was/is 'encoding scientific units of measure'. I myself will (ab)use it for CS "units" too. I don't see much in the schema limiting UnitsML from being used to model other units Martin S Weber (NIST): @Joe Collins1: this was done with XMLSpy* (yeah you asked Bob, I'm on the group, too, though.) [ *Certain commercial software was identified as being used by the UnitsML Group at the National Institute Of Technology. Such identification is not intended to imply recommendation or endorsement by the National Institute of Standards and Technology, nor is it intended to imply that the software identified is necessarily the best available for the purpose. ] FrankOlken: From BobDragoset's slides: For the most recent UnitsML schema and documentation with images, go to: http://unitsml.nist.gov FrankOlken: From BobDragoset's slides: For information about SI units and non-SI units for the U.S., go to: http://physics.nist.gov/sp811 RaviSharma: Pat- Money is a measure but its units are measured by currency type, but gold is ounces or gms? Joe Collins1: Gold - troy ounces!! MikeBennett: @Ravi: the world moved from measuring money in weights of metals, to recognising the property ofr money as an information construct. If it has a dimension of a basic "stuff" it's information. Joe Collins1: Yes, Money is a social construct and not measurable in the same way as physical quantities. anonymous morphed into LinePouchard Martin S Weber (NIST): @Ravi We were jokingly talking about money last UnitsML TC Meeting. You -could- model that as dimension & quantity (counted item) money.. and model 1 unit per currency involved. UnitsML wasn't designed for that, you -could- model it .. Martin S Weber (NIST): several problems with the extreme time-dependance with "unit" conversions arise though DavidLeal: A distinction that nobody has yet made is between unit and scale (oops Pat just has). Decibel is a scale, but not a unit. A very important scale is ITS90 - the practical temperature scale, which is an appoximation to the linear scale derived from the unit Kelvin. MikeBennett: @Martin - then you would (jokingly) fall in to the trap of "designing" some solution to the problem. We have seen a lot of clever stuff in ontology, when what we should be doing is teasing out the simplcity of reality itself. James Davenport (OpenMath): @Martin S Weber (NIST): time AND space - consider pount/guinea arbitrage in 17th century England Joe Collins1: Units and Quantities ARE a physical theory, mathematically constructed. Math constructs are inseparable from meaningful discussion. Martin S Weber (NIST): @JD: true. (We talked about using external WSDL described services to do these conversions. You could model time in unitsml. not space (directly) though) MikeBennett: @PatHayes - some real clarity there. PatHayes: @MikeBennett: reality isnt simple, and it doesn't have units RaviSharma: Thanks Martin and Mike. units are required for common understanding such as measuring a given string or reproducing it by production, etc. and so is wealth or money by measuring the amount required for exchange conversion or barter. Douglas Mann: ISO 31 has quantity, dimension, and unit. (see: ISO 31-0 "section 2.2.6 Dimension of a quantity") Douglas Mann: ISO 31-0 says velocity is a Quantity and L/T is the Dimension of velocity. Douglas Mann: Are people aware of this website? http://www.unc.edu/~rowlett/units/ ... Just a few units. anonymous morphed into RobRaskin EvanWallace: VIM James Davenport (OpenMath): Not sure if I was heard, but fuel efficiency can be in miles/gallon or litres/100km. Martin S Weber (NIST): @just speaking: dimension analysis is dangerous here: Think energy / torque. You really have to talk about quantities, not dimensions when trying to compare unit'ized values MikeBennett: @PatHayes - good point. I guess Einstein's dictum applies. I just get worried when I see some clever "solution" to the "Problem". That's what the next level of system design is about whereas the ontology should be accurately depicting the problem. You said it cleare than I could tho. James Davenport (OpenMath): And the two CAN be directly compared, even though they are reciprocal. Mark Linehan: The VIM standard that I mentioned is called "International vocabulary of metrology Basic and general concepts and associated terms (VIM)" JoshLieberman: Many indirect but few direct references to the fact that units / dimensions / quantities are only meaningful together within a coordinate reference system. How and should this work into a units ontology? JoelBender: We don't need MathML? EvanWallace: Josh: either we have to adopt one reference quantity (dimension) system or we need to Mark Linehan: it is standard number JCGM 200:2008 issued by BIPM - the Bureau International des Poids et Measure -- the people who put out SI EvanWallace: allow (as VIM does) the definition and reference of different quantity systems. Martin S Weber (NIST): Joel: at some point math comes in and when it comes to marking this up, will you invent something new or use something existing like OM or MathML? Joe Collins1: The distinction between energy and torque can be made using the SI concept of Kind of Quantity. JoshLieberman: This is unavoidable for spatial units, but I wonder to what extent it is behind other multiple unit conventions in other domains. RaviSharma: Speakers: Object of measure is either a physical object or geotemporal - geometric object. However processes can alkso be measured thru these. any thoughts? Martin S Weber (NIST): @Joe Collins1: that was what I was pointing at. Unit & dimension is not enough alone. PatHayes: Do dimensions have 'reasonable' ranges of scale? Eg mm and parsecs are both lengths, but maybe should be treated as different dimensions (?) MikeBennett: I think the fundamental ontological problem is "Amount" v "Quantity", extended out into what they are a numeric number or a measured amount of, and in the case of some continuous quantity, what unit that is measured in. Each thing so measured is measured along some dimension. JoshLieberman: @PatHayes - coordinate reference systems have different scales, which might indicate why mm and parsec are not convertible in most practices. Joe Collins1: The amount of a quantity is definable using the SI concepts unit(Q) and num(Q). Martin S Weber (NIST): @MikeBennett to me you have a reading, a measurement of something. All possible readings make up the quantity. You reason about this reading with reference to a "1" thing, which is the unit. And the dimension is there to describe the class of all possible readings. Doesn't that describe amount vs. quantity? James Davenport (OpenMath): @PatHayes: that's a consequence, not a property. months are days are botrh time, bt aren't interconvertible EvanWallace: As mentioned depending on units alone is ambiguous since derived units can represent different quantities. PatHayes: @James Davenport: months are a notorious special case, because they aren't even units. BUt my point would be, should we treat say milliseconds and millennia as belonging in different dimensions? DaveMcComb: to PatHayes: I agree, but from a pragmatic point, if I have a dimension and unit, to communicate I need to get people to agree on two things (ie the concept of length and an agreement on "meter") where if I derive length from convertTo meter, once they've agreed on "meter" I've got both. MikeBennett: @Douglas Mann: dimensionality versus Dimension. Dimension is a feature of reality so if we ignore it we will be stuck forever in some technical design "workaound" that is not an ontology at all. MikeBennett: sorry I have to jump off at 11 so I won't have time to get to the top of the list. James Davenport (OpenMath): @PatHayes I disagree that months aren't units, but they are certainly notorious. For your second, no: they ARE the same dimension, since there's a contunuous spectrum between them. Martin S Weber (NIST): @PatHayes: milliseconds and millenia, there's just a tiny constant between them (well or function depending on which calender you choose to use). So I'd throw them in the same pot... RaviSharma: Speakers: can NOTHING have quantity or dimension? Obviously it is something physical or concept of reality in space-time --- that only requires measurability and therefore units - the subject of today's ontology dialog! PatHayes: @DaveMcComb: True. You have efficiency on your side, I have robustness. I guess my point could be phrased: robustness is more important than efficiency when we are waning to keep a wide scope. PatHayes: waning/wanting DaveMcComb: I also find that people find it easier to agree to concrete things than abstract things. so getting them to agree to a couple of scratches on a piece of platinum (a meter) is actually easier than getting them to agree on something more abstract, derived from that (length) Joe Collins1: The vacuum (nothing?) has physical properties. MikeBennett: We should end up with the same view of dimensions and units and things as we learnt about in school. Length is a dimension. Time is orthoginal to length and so is another dimension. As it charge and mass. Mark Linehan: PatHayes: In our OMG date-time effort, we ended up distinguishing "precise units" (e.g. second) from "nominal units" (e.g. months). Common language has both but they have different properties for reasoning purposes. JoshLieberman: Re: millisecond and millenium - there may be a nominal conversion, but over any real length of time, the number of seconds in a year will vary, so a particular millenium will have different numbers - coordinate reference system! PatHayes: @Joe Collins1: If it has properties, the it's not nothing (spoken as a logician). Martin S Weber (NIST): @Joe Collins: The vacuum is a certain reading of amount of o2 (and pressure?), isn't it? So it's not nothing JoshLieberman: Is a unit ontology of "inherent properties" or "observed phenomena"? DaveMcComb: Who was just speaking? That was good Joe Collins1: @ Martin S Weber (NIST) - Not just that. there are quantum fluctuations regardless of gas pressure. James Davenport (OpenMath): @Mark, so both a year and a month are nominal, but nevertheless a year is precisely 12 months (PS: did we meet at YKT?) EvanWallace: +1 to speaker comment ChipMasters: ChipMasters (Me) Mark Linehan: @James -- yes, and yes. Martin S Weber (NIST): @Joe C: what I meant to say, if you can 'measure' it being there, it's not nothing ... you can't measure nothing But then again, 0 has properties.. DaveMcComb: Chip, your points, I think was that the dimension is baked into the system (not independent)? Interesting SusanTurnbull: Perhaps use the term dimensions (plural) vs. dimentionality (as in higher orders of dimension - i.e. high energy physics, string theory, n-dimentionality, etc.) RaviSharma: Speakers: it is one object but does not get described in all aspercts by dimensions that imply origin and geometric concepts - but there are properties that do not belong to common dimension such as spin and quantum levels etc? Joe Collins1: The effects of quantum fluctuations in a vacuum are measurable. Whether you want to refer to the vacuum as "nothing" is a matter of definitions. RaviSharma: Susan - yes similar to what i wanted to express. DaveMcComb: On Frank's point of scope: should we have a scientific scope and a general use scope (I'd suggest the general including the scientific so it's not inconsistent when they use the same units) RaviSharma: zero has no properties really it is approximating the object with zero that has property. RaviSharma: Speakers: Electron - what are its units/ dimension (geo) or energy or charge, etc. therefore every object can have different units for different purposes? Martin S Weber (NIST): @Ravi: is the neutral element to e.g. integer addition, that by which you cannot divide, to you zero (which thus would have properties), or the object trying to represent itself? RaviSharma: Martin - not clear on Neutral element but zero and infinity are more powerful concepts than any enumerables - this is the basis of indian philosphy - at least one of them! PatHayes: I am wondering what kind of dimension mg-per-gram can be? Does that make sense in CGI? EvanWallace: NIST is not an SDO. RaviSharma: Pat - measure in a mixture such as alloy, imprity, etc for example, very useful. SteveRay: NIST would still need to identify an SDO to work with. PeterYim: @PatHayes - another example of imprecise units is, maybe, "lunch time" (this was an example brought up when we were trying to convince the EHR people that they need to seriously consider ontologies back in 2004) Martin S Weber (NIST): @PeterYim: It's not imprecise, it's just not constant. and I wouldn't call it a unit but a measurement to which you can relate. PeterYim: @Martin S Weber (NIST): indeed ... thanks PatHayes: @ravi: I understand, but if you do dimensional analysis, its mass/mass= dimensionless. Which of course it is, since it "measures" a ratio. IMO this is one illustration of why different communities will need different ontologies. Some thing is useful for one community but incoherent to another. DavidPrice: So W3C, OMG or OASIS are the options for open stds. JoshLieberman: concentration is a quantity / measurand / phenomenon which, having dimensions of gg-1, is dimensionless PatHayes: I doubt very much if the W3C would consider this within its purview. Just a personal opinion. Martin S Weber (NIST): @ Josh, Pat : to be pedantic: It is dimensionless but it's measuring the ratio of two different quantities RaviSharma: Pat your example of mg per gm is pertinent. I agree with you that we will have different ontologies but also a need to converge or interoperate so as to not have knowledge sioes? RaviSharma: siloes? DavidLeal: I support the use of MathML too. We don't want to reinvent things, and MathML has useful capabilities like lambda calculus. SteveRay: I concur with HowardMason's strategy. PatHayes: MathML is too general, and does not support the needed ontological expressiveness. RaviSharma: Pat - dimensionless analysis is usual in high energy and gravitational work and is a way to normalize or scale. Martin S Weber (NIST): @PatHayes don't use it to express everything, use it to express the parts of math you need and embed it. Better than trying to come up with something to describe formulae if you need to imho PeterYim: @HowardMason - the recommendation you made (verbally) just now is important ... would you document it here on the chat, please (so it can go into the proceedings, besides just being captured on the recording) ... thanks JoshLieberman: That does bring up the question of what are the initial applications and therefore requirements (logical, expressive, computable) of this ontology? Are "weird" units indeed an initial target, or a later problem for example? James Davenport (OpenMath): @PatHayes: I disagree, but we should have this discussion offline. J.H.Davenport@bath.ac.uk PatHayes: @Ravi: engineering and bioinformatics can likely be in different siloes for the forseeable future without harm. Trying to find one standard/ontology to cover both is likely to be too large a stretch, IMO. Put StanHuff and BobDragoset in a room and wait for the color of the smoke to change? PatHayes: @James: OK, lets. Note I wasnt suggesting ignoring MathML altogether. Pat Cassidy: On the issue of ratios of quantities of substances as a 'unit', it may be necessary to have a complex 'unit' (or attribute value) consisting of the ratio of two substances. That 'unit' would have pointers to each substance, and the attribute value would include the numerical ratio. Joe Collins1: I think that the definitions of the units and definitions within MathML must follow from their definition by a metrological standards body. PatHayes: @Joe Collins1: metrological? RaviSharma: Pat - you have a point but MEMS - nano mechanical delivery of medicines is a real case for understanding both together? James Davenport (OpenMath): @PatHayes, JC: what I said was that the MATHEMATICS should be in MathML, e.g. NIST's x:=a+b*(y+d)/c Joe Collins1: Metrology, the science of measurement. PatHayes: @Ravi: good point. I hadnt thought of that particular nexus. EvanWallace: Please add me to the list Peter. Martin S Weber (NIST): @James Davenport: I agree and MathML and/or OpenMath will be supported in a near-future version (not the dreaded "1.0") for that reason BobDragoset: @JamesD: there is a strong possibility that UnitsML will adopt MathML to represent mathematical expressions. DavidPrice: Can NIST organize, facilitate the work with the aim of eventually standardizing in BIPM? PatHayes: @Stan O, speaking: There seemed also to be some consensus on the need for extendability. DaveMcComb: Agree completely with PatHayes's last statement Joe Collins1: BobDragoset, I'd like to forward my paper "OpenMath Content Dictionaries for SI Quantities and Units". DaveMcComb: We do agree on a lot, and should start there EvanWallace: Good idea Frank. PeterYim: as discussed: (1) I will be creating a new mailing list to cover this "Towards A Quantities and Units of Measure Ontology-based Standard" discussion, (2) I will automatically subscribe everyone who is here today to that list (if you want to opt out, please email me peter.yim@cim3.com), (3) I will announce the creation of this mailing list both on [ontolog-forum] and on the session page we are using today, i.e. http://ontolog.cim3.net/cgi-bin/wiki.pl?ConferenceCall_2009_06_19 BobDragoset: @DavidPrice: I think that's a possibility, but I'm not an ontology expert. Mark Linehan: Pat, there may be a core set of concepts that can have relatively broad concensus, and then extension sets for specific disciplines PatHayes: Can anyone gather together all the readings people think we should all have read? PatHayes: We can't say goodbye as we are all muted. DaveMcComb: bye DavidPrice: BobP / BobD! FrankOlken: We will continue the discussion on an email list that Peter Yim will construct and put a link to on the EvanWallace: Bye. DougHolmes: Adios FrankOlken: conference call wiki page at Ontolog Forum FrankOlken: Signing off PeterYim: also, there are people who I don't know (and don't have a email address for) ... therefore if you attended this session, but did not pre-register by email, please drop me a note to to make sure I have your email addresses and your affiliations PeterYim: great session ... thanks you everyone! -- session ended: 2009.06.19-12.33pm PDT --