John Sowa wrote:
> 1. Don't attempt to define any term that VIM doesn't define.
> If you want to use it, take it as an undefined primitive.
>
> 2. If VIM does define some term that is needed for UoM, then
> restate the VIM definition in whatever version of logic
> we use, but don't add more detail.
>
> Let the upper level gang do whatever they want to do at the
> upper levels, but don't attempt to define anything that they
> are already defining. (01)
Yes, except for one thing: it may well be that your detailed
analysis finds some inconsistencies and ambiguities with the
VIM that would be worthwhile to feed back and improve the
VIM. (02)
Martin S. Weber wrote:
> As has come up in the discussion itself, too, units are not meaningful
>without
> quantities. If you do not have a proper representation of "kind of quantity"
> (per VIM) then you'll never know if two things are comparable. In the light
>of
> UCUM this might be acceptable, but in the light of a units of measure
>standard
> *ontology* I don't see how this can ever be left out. (03)
Martin, (04)
I didn't say you shouldn't include kind-of-quantity in a general
sense, but you need to be careful not to want to define the
specific ones. Else you can't do it. (05)
/1 m/ is a unit of length. Units are quantities and belong to a class
of "kinds of quantities", but you should keep the kind-of-quantity
class general to include all quantities in that dimension L. (06)
Leave it to others to define what length is. Build in that 100 cm is
identical to 1 m, and that 1 cm is 2.54 international inch (which
you know based on detail knowledge relying on an authority telling
you this.) (07)
The reason why people want to drag specific kinds of quantity into
units is that they want to avoid two quantities to end up in the same
equivalence class, want to prevent them from being comparable. But
that is a futile attempt -- who says I can't compare the ratio of
two heights with the ratio of two masses? I can and it makes perfect
sense in many cases. I maintain that there is no problem with
"dimensionless units" that does not also exists with any other
dimension, because you can have some base dimension cancel out in
a term which then gives rise to the argument that two values of
that dimension are thought to be not comparable. (08)
On the other hand, you can have two lengths that are really lengths,
but still not comparable because the measurement conditions (e.g.,
temperature, atmospheric pressure) were different. The only way
to assure whether two measurement results are comparable is to make
sure that the respective specific kinds of quantity are comparable
for the given requirement to compare. The possibilities are endless. (09)
The only reason why we struggle with procedure defined (arbitrary)
units is because the biomedical industry that is adopting UCUM is
pushing us to do so. Most of the problem again comes from the fallacy
to want to differentiate specific kinds of quantity through the
unit rather than through stating the specific kind of quantity.
My current thinking on this is that essentially each procedure
defined unit that would be recognized must be treated as a dimension
onto itself (similar to candela in the SI). But I am watching ISO
TC12's 80003 project with interest. (010)
regards,
-Gunther (011)
--
Gunther Schadow, M.D., Ph.D. gschadow@xxxxxxxxxxxxxxx
Associate Professor Indiana University School of Informatics
Regenstrief Institute, Inc. Indiana University School of Medicine
tel:1(317)423-5521 http://aurora.regenstrief.org (012)
_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/uom-ontology-std/
Subscribe: mailto:uom-ontology-std-join@xxxxxxxxxxxxxxxx
Config/Unsubscribe: http://ontolog.cim3.net/mailman/listinfo/uom-ontology-std/
Shared Files: http://ontolog.cim3.net/file/work/UoM/
Wiki: http://ontolog.cim3.net/cgi-bin/wiki.pl?UoM_Ontology_Standard (013)
|