On Tue, 12 Jun 2007, Frederic V. Hessman wrote:
> On 12 Jun 2007, at 3:49 pm, Steve Allen wrote:
>> Please don't use adjectives "heliocentric" or HJD anymore.
>> Everything IAU-approved is currently "barycentric".
>
> It may be "IAU-approved", but "barycentric" is not unambiguous: a
> "barycenter" is simply w.r.t. SOME (but not any particular) center-of-mass,
> so if you want the Solar System we'll have to come up with something like
>
> ucd="time.julian;time.barycentric;SolarSystem"
>
> (note that NONE of these are UCD1+). The IAU may not used "HJD" anymore,
> but I know zillions of astronomers who do ;-)
>
> Rick
Yes, we have to be careful not to get confused between what we might be trying to enforce as journal editors, supervisors/students, etc. etc., and what the VO should do. The VO should describe data as accurately as necessary - just about everything possible is in STC, so it isn't hard to figure out what to use. If significant data use a heliocentric time frame then we need to know what it is. I don't know what is worse, to mistakenly thnk that there are not data for something because the VO would not publish data in the 'wrong' coordinate frame, or if the VO makes an assumption that everything is in our favourite frame when in fact it's in something different. I am a scrupulous J2000, SI user as approved by the IAU, but I guarantee that if I talk about a magnetic field in Tesla or a number density in m-3, but did not specify the units, most astronomers over 30 think I mean Gauss and cm-3.
We also run into possible conversion problems - I don't know if this applies to time, so the VOEvent people can probably ignor the rest of this para, but I am often faced with converting Heliocentric velocities into VLSR, or B1950 positions into J2000. But in both cases you need to know the epoch of observation which is rarely available for old data, if you want to be fully accurate. A rule of thumb conversion is good enough for registry searches but the actual data returned should be in the original units unless the user specifies otherwise - and only converted if the VO can do it accurately to within the precisioin of the data (e.g. for IRAS, the epoch is less important than for VLBI...)
cheers
a