[fitsbits] Start of the WCS Paper III Public Comment Period

[Doug Mink]

William Thompson wrote:
> One of the keywords introduced in this paper is MJD-AVG, representing the
> average time of the observation.  If MJD-AVG is absent, then DATE-OBS is used in
> its place, but the latter refers to the start of the observation.  Why not also
> introduce a keyword DATE-AVG, with the same ISO 8601 format as DATE-OBS?

I second this.  All of my code which deals with timing already has to deal with both
Julian Dates and ISO 8601 dates, so implementation of the more human readable
DATE-AVG would not be difficult.  If both are allowed, would the standard have to
require that one take precedence over the other if both are present?

> I say this mainly because I'm concerned that there is developing a split between
> two different time-handling methodologies, one based on ISO 8601, and a
> different methodology used within WCS, based on MJD.

It would be optimum to support both types of time specification, which is what
anyone who currently deals with time in real FITS files has to do anyway.

-Doug Mink