What is the actual (scientific) value of having units associated with
raw measurements?
If I record a temperature value in degreeC using instrument of type A,
and you record a temperature value in degreeC using instrument of type
B, we know formally what sorts of comparisons we can do between our two
values.
If, on the other hand, I record a temperature value in volts using
instrument of type A, and you record a temperature value in volts using
instrument of type B, then there is no way we can do any comparison
between the two values - the instruments are quite likely to have quite
different temperature-voltage characteristics. Unless, that is, we
know the relationship between volts and degreeC for both instruments, in
which case it is irrelevant that the raw data is in volts in both cases
- it could just as well have been mm mercury for instrument A, and
bloodles for instrument B.
Cheers,
Robert
John Graybeal wrote:
> With regard to adding a 'raw_units' attribute to the standard: I'm
> splitting this thread from the '_raw' standard name thread (as I think
> it is secondary to that thread).
>
> Prompted by Jonathan's observation below: I think most of the cases I
> have in hand can be expressed as some variation of 'count', 'bit', or
> 'volt', all of which are UDUnits. (counts? who would have thought?)
>
> If I can find some clean examples that require non-UDUnits units, then
> I might propose the creation of an alternate_units attribute via the
> TRAC list. (If anyone else has some, feel free to pass them along.)
>
> John
>
> On Mar 12, 2009, at 9:23 AM, Jonathan Gregory wrote:
>
>> I would suggest not deciding in advance that we have to treat them in
>> different ways, because it's best to minimise the complexity of the
>> convention. It's often tempting to regard something as a new case and
>> invent new conventions for it, but it may be easier in the end to
>> find a way to express it using structures we already have.
>
>
> _______________________________________________
> CF-metadata mailing list
> CF-metadata at cgd.ucar.edu
> http://mailman.cgd.ucar.edu/mailman/listinfo/cf-metadata
>
--
The University of Edinburgh is a charitable body, registered in
Scotland, with registration number SC005336.
Received on Thu Mar 12 2009 - 17:15:52 GMT
