Re: String representations of numeric values

From: Rob Seaman <seaman-at-noao.edu>
Date: Fri, 28 Apr 2006 06:27:44 -0700


On Apr 28, 2006, at 2:04 AM, Mark Taylor wrote:

> Although this suggestion has met with considerable resistance on the
> list, I would still like to discuss it at Victoria.

That wasn't resistance, that was enthusiasm for the discussion. (Don't get me started on leap seconds!) Sorry I'll miss the fun.

> Even if one takes the position that the correct answer in the case of
> angles is to require data providers to use decimal degrees and simply
> to ignore tables which don't do so, I think (having heard
> representations
> from solar physicists at least) that this doesn't solve the problem of
> times, since unlike unlike angles there is no obvious zero point.

I wouldn't count on zero points for angles, either - the IAU has recently deprecated the Vernal Equinox.

> MJD is one convention, but there are others (e.g. JD), and there is
> currently no way within VOTable to specify which of these conventions
> you're using. The correct answer to this may be STC, but we don't
> currently have that within VOTable either. People know what is meant
> by ISO-8601 time strings. I'm not a time expert so it may be that
> this format carries subtle ambiguities as well, but from a pragmatic
> point of view it's clear that it would satisfy a need of a large
> number of potential users of the format without a major change.

"Pragmatic" is a slippery concept. You may not end up at STC before you're done - but a complete solution for VOTable would be something just a complex. Suggest an attempt to focus exclusively on representation may escape you. Note that none of MJD, JD or ISO-8601 convey a specific system of time, somehow a table needs to provide this whether UTC, UT1, TAI, TT (or too often in astronomy, some poorly constrained mixture). We've been having a rousing discussion of topocentric, geocentric and barycentric coordinates over in VOEvent.

Sexigesimality is not the most harrowing issue. If the VOTable team were to adopt STC and write a reference implementation of an STC class library along with some Swiss Army Knife style coordinate handling tools, VOEvent would be delighted to stea...er...borrow your handiwork.

Rob
seaman-at-noao.edu Received on 2006-04-28Z15:27:47