Spectrum model: Github issues

[Markus Demleitner]

Hi Mark,

(tangential: I've taken the liberty to fix git to github in the
subject; I've read "git issues" as "problems with git", and I was
worried that there's ivoatex trouble brewing.  And really, github's
bug tracker has very little to do with git as such)

On Thu, Feb 06, 2025 at 03:36:30PM -0500, CresitelloDittmar, Mark via dm wrote:
> Spectrum Model: recent topics
>
>    - Spectrum.Char.SpatialAxis.Coverage.Location (20240705)
>    - https://github.com/ivoa-std/SpectrumDM/issues/13
>       - Is there an action here to look into making this OPTIONAL in the
>       model?

I'd very much support that to the extent that I'd write an Erratum,
in particular because I've handed out position-less SDM files for
more than a decade by now.  Gently prod me to get me started.

>       - Do we need to add something to identify Theoretical?

Good question.  I'd say we need a usage model for that.

Clearly, the distinction between observed and "something else" is
very important in discovery, and for that, we've had
SimpleDALRegExt's DataSource
<http://docs.g-vo.org/schemadoc/schemas/SSA-v1_2_xsd/simpleTypes/DataSource.html>
for a long time.  Splat's use of that, for one, is fairly popular.

Do we have indications that we need this machine-readably in the
dataset itself?

>       - Do we want to recommend how they should identify the dataset as
>       "GAIA DR4"?

In discovery, that would be covered in Obscore's obs_collection and
its SSAP counterpart.  Having it in the dataset certainly isn't
wrong.  As an informal convention and until we've finally gotten rid
of more or less ad-hoc utype usage, I'd say a PARAM with a utype of
obscore:dataid.collection would be the least surprising measure.

But again: is there a usage scenario?


>       - VOCLASS keyword value in spec does not match example
>    - https://github.com/ivoa-std/SpectrumDM/issues/12

Let's fix this to Spectrum-1.0, because that's in line with the
syntax of similar things in the VO.  While we're doing this, let's
also tell people that to determine if the file is something they can
process they should in general ignore the minor version, i.e., the
test would be hdr["VOCLASS"].startswith("Spectrum-1."), and that it's
case-sensitive.

Let's do an erratum.  This is clearly an editorial oversight.
Volunteers?

>       - DATE-OBS keyword
>    - https://github.com/ivoa-std/SpectrumDM/issues/8

You have my vote.  Not so sure about the "Start of observation",
though; in SIAP, the corresponding metadata field is for the
midpoint, because that's in some sense more robust.  So, before
issuing an erratum, I'd say we should first try to research whether
all the various FITS conventions agree DATE-OBS is the *start* of the
observation.  Oh, and let's be modern and insist on DALI-style
timestamps in there.

Thanks,

           Markus