[ssig] refframe vocabulary

[Markus Demleitner]

Hi all,

I'm sure followups to this should be restricted to ssig at ivoa.net.
For those on ssig: this is about a proposed vocabulary at
http://docs.g-vo.org/rdf/refframe/.

On Thu, Apr 11, 2019 at 02:04:30PM +0200, Baptiste Cecconi wrote:
> First, here we talk about space reference frames, right ? The STC
> model defines reference frames for space (SpaceRefFrame), time
> (TimeFrame), spectral (SpectralFrame), velocity (RedshiftFrame). 

This is a list of spatial reference frame names, yes.

> I see that in your proposal, all solar system reference frames
> should fall into the BODY coordinates bag. I guess that is
> possible, but the definition of the the BODY generic frame will not
> work in this case. Some planetary reference frames are
> planetocentric, others are planetographic, planetodetic (don't ask
> the difference, I don't really know :-), or even barycentric

Well, I'm open to new top-level terms as long as they help machines
make sense of what they're seeing, because that's the purpose of
these terms and their hierarchy.  I'm not enough of a cartographer to
figure out if it helps to know if a frame is planetographic or
-centric, though.

> (planetary system (planet+moons) barycenter). On Martian rovers,
> for instance, AZ_EL may be used (but on a different planet than
> Earth), etc. Moreover, for Earth magnetospheric frames, would

On AZ_EL, I suspect the nature of the ground shouldn't matter.  To
make sense of these automatically, we'd need to define the observer
location, and when we do this, I don't think we'll much favour Earth
over anything else.

> On the VESPA wiki, there is a page where we have tried to list all
> reference frames we know of for planetary and solar sciences:
> https://voparis-confluence.obspm.fr/display/VES/Planetary+Coordinate+Systems
> 
> There is also coordinate reference system list for planetary bodies
> maintained for GIS applications on Mars and the Moon:
> https://github.com/planetserver/planetary-crs
> 
> We have also started a github repo for SolarSystemRefFrames, with a
> list of bibliographic references and sources for Ref Frames:
> https://github.com/epn-vespa/SolarSystemRefFrames/blob/master/list-of-sources.md

Hm -- that's... quite a bit.

Although STC has never been explicit about that and thus the question
might really be up for discussion, I would suggest that the purpose
of this vocabulary is operational rather than encyclopaedic.  This
means: Before we add a term in here, there should be a VOTable in
which the term needs to be used, and there should be a consumer of
that VOTable that can actually make use of the term when it's there.

We (i.e., the Semantics WG) plan to make it reasonably easy to add
terms to the lists as needed, so there's no reason to put in terms
"because we might need it later".

So: is there a top 10 or top 20 of spatial reference frames you'd
actually need and can use in VESPA right now?  Something of that
order would be great for prototyping.

You see, the whole big list *is* a bit overwhelming (bordering on
discouraging), and if we get it wrong with 500 terms, cleaning up is
a lot harder than when we've got just a dozen.

Thanks,

        Markus