RE: SUO: Non-gibberish Version of Ontological Additions
Yang,
Thanks for your comments. Please see my replies below.
-Ian
> -----Original Message-----
> From: Yang Yun [mailto:yangyun@metacrawler.com]
> Sent: Sunday, March 11, 2001 3:00 PM
> To: iniles@teknowledge.com; standard-upper-ontology (e-mail)
> Subject: Re: SUO: Non-gibberish Version of Ontological Additions
>
>
> Four points:
>
> 1. Set theory
> add Symmetric Difference for convenience?
The "Set Theory" section of the merged ontology contains an operator for set
theoretic difference, viz. 'DifferenceFn'. Are you proposing another
function?
>
> 2. Quantities, Calendar Time
>
> Base engineering quantities should be easy
> to find in the upper ontology, say as
> 'Measurable Thing'.
>
> Spatial dimension, time dimension, gravitational force,
> temperature, frequency, EMF, and chemical molecules are
> all basic .
>
> Measurables are things that can be quantified by an observer,
> resulting in either a (value, unit) pair
> value : number X unit : ReferenceUnit
> or a triple
> value : number X uncertainty : number X unit : ReferenceUnit.
>
>
> ReferenceUnits are part of ReferenceSystems such as
> Universal Coordinated Time, SI Length, SI Mass etc.
> The upper ontology may simply state ReferenceSystems exist,
> and allow specific domains such as Software Metrics or
> Global Positioning Systems to choose relevant units.
>
>
> ReferenceSystems are often combined in the one measurement,
> eg height position from one datum, lat-long position from
> another datum , time , date, and uncertainty from other systems.
>
> Naive notions of geospatial position could be
> accommodated as local referencesystems with simple
> orderings such as 'the computer monitor is on the desk which
> is on the floor'.
>
> Temporal orderings such as names for days of the week
> could be placed into a specific domain also.
Sofia Pinto of the Instituto Superior Tecnico is working on revising the
"Quantities and Units of Measure" section of the ontology. Once she's
finished with her revisions, we'll distibute it to the group for comments.
>
> Advantages
>
> The advantage of such an approach is that
> ISO metrology terms are used where possible,
> rather than inventing a new vocabulary in
> place of terms every engineer has grown up with.
>
>
> Neologisms like Abstract, Quantity, PhysicalQuantity,
> ConstantQuantity, ScalarQuantity, FunctionQuantity
> UnaryScalarFunctionQuantity, TimeDependentQuantity,
> Unit-Of-Measure, TimeMeasure-Position, birthTime,
> deathTime, YearFn, MonthFn, DayFn, HourFn, MinuteFn
> , SecondFn, and RelationExtendedToQuantities Relation
> are just unnecessary and will have no effect.
Could you provide some reasoning here? I understand that you don't like the
listed constants, but, in the absence of any argument against their
inclusion in the merged ontology, we can't make progress.
> Simplicity is usually a good criterion in science.
>
> Both coordinate-based and other reference systems
> can be accommodated. Specific engineering areas like
> precise spatial location or precise time can draw on
> more complicated domain ReferenceSystems as needed.
>
>
> 3. Positions
>
> If a single ReferenceSystem hierarchy is used,
> it will be possible to define comparisions in
> a single hierarchy based on geometric dimension
> of the items being compared. In some ways time
> comparisons are like linear geometry for example.
>
> ISO standard geometry terms and comparisons could
> be used in the core, based on 0-dimension, 1-dimension,
> etc. The set-theoretic core terms are :
> disjoint, relates, boundary, exterior, interior, equals,
> contains, intersects, overlaps, touches, within,
> and crosses.
>
> The up-down, left-right reference system referred
> to with words like left-of, right-of, on, under , below,
> above, behind etc could be constructed more clearly
> from the base terminology, no?
I guess you're saying here that the positional predicates could be defined
in terms of geometrical concepts. I'm sceptical but interested. Could you
show concretely how this might be done for a couple of the positional
predicates?
>
>
> 4. Artifacts, Social Hierarchy
>
> For the engineering ontology, what traceability is there
> to terminology that working engineers actually want here?
OK, I think we need to clarify something here. We have all agreed that the
SUO should be regarded as an engineering artifact, but that doesn't imply
that every concept expressed in the SUO must be a term from an engineering
discipline. The goal of the SUO is to facilitate interoperability between
different information systems by providing a common semantics. Since many
systems created by engineers interface with people who don't have an
engineering background, many systems will inevitably make use of
non-engineering concepts. For example, the Amazon.com database(s) must have
notions of book and customer. Accordingly, we should, I think, offer
support for notions like these in the SUO.
>
> EngineeringElement seems indistinguishable from
> EngineeringComponent. Then I became completely lost
> around about 'engineeringSubcomponent IrreflexiveRelation'.
> huh?
OK, 'EngineeringElement' is the superclass for both 'Artifact', a
self-contained, human-made object, and 'EngineeringComponent', a component
of an 'Artifact'. Thus, the chair I'm sitting on is an 'Artifact', while
the legs of the chair are 'EngineeringComponents'.
>
>
> The miscellaneous Social groups are from what theory?
> Whatever..
I'm not sure what you're asking here.
> yy
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
> -------------------------------------------------------
> Searching for the best free email? Try MetaCrawler Mail,
> from the #1 metasearch service on the Web, http://www.metacrawler.com
>