Describing things for software to read

[Roy Williams]

> Or you might want to query on all filters that have greater than 90%
> transmission over some subrange of their .....

This kind of careful matching is possible in the UCD approach.

Consider these two:
    phot.flux.IR.100-200um
    phot.flux.radio.1500-3000GHz

a. If you use a matching function based on string comparison, these do not
match at all, although you might argue that the common substring "phot.flux"
indicates some non-zero matching.

b. However -- if I remebered the speed of light correctly -- these two UCDs
have significant overlap. Therefore a better matching function -- based on
numbers as well as strings -- will evaluate (correctly) to a strong match.

c. As the UCD specification for filters becomes more intricate, we see that
a really sophisticated matching function becomes possible -- as Doug and
Matthew would like to see.

d. NOTE here that nobody has said anything about proper, official,
VO-blessed names for filters or bandpasses. We just build structure that can
deal with the messy, fuzzy, human vocabulary that real astronomers use.