Hi,
First, it's good to see Bob's document (sorry Bob, you have to take most of the credit, the rest of us should get the blame for muddying the waters). I kept out of the debate because I had more than my go last year and the present version seems just right for the present time.
There have been suggestions made which will be important in the long term (i.e. years) and I have a comment for the next round (which everyone will have forgotten by then).
I think that how data are collected _is_ a factor, as is the number of catalogues and other data resources involved including publications, per spectral division. The purpose of the Registry is not to be a model of symmetry and logic (although it helps human maintainers if the structure is not too arbitrary). The purpose is to help users find data. So it makes sense to divide approximately along the lines of where data are provided from (which archives to search, which catalogues) and what researchers think of in the appropriate labels. We also don't want to have to search too many resources in one group, or too many groups for one query. Of course, some archives and many queries are multi-wavelength and many categories overlap, but the present division seems to be reasonably close to an even spread across the current science cases I know of from AVO and AstroGrid, apart from the UV and gamma rays - but that is a deficiency of the science cases or maybe of archive availability.
In the longer term:
When I proposed dropping EUV last year, there were objections from the solar community. However we are not catering to them much anyway - but we shouldn't forget them. So perhaps we will need to reinstate that category - and create a separate decametre category - when we do include solar/STP work. But for now there are too little data available to astronomers to be worthwhile.
Similarly, there will be a flood of data from ALMA in a few years time, and even before that, I hope that the involvment of the JACH in RadioNet will lead to more JCMT and SCUBA data being available from VOs - but at present millimetre covering the whole range is underpopulated.
Finally, Nic is right about Astroparticles - how are these usually described? Are these usually in the gamma ray energy range? Do we serve these data? Could they be served by the gamma ray energy range until a fuller definition is obtained?
In short, the bandpass definitions should be science-case-driven; the present divisions meet the present science case needs as far as I am aware, and we should come up with the science cases before we attempt re-definition - in several years' time!
cheers
a