Fwd: Re: Datalink vocabulary extension: sibling?

[François Bonnarel]

Dear all,

I don't really know if these belongs to apps or to semantics

cheers

François



-------- Message transféré --------
Sujet : 	Re: Datalink vocabulary extension: sibling?
Date : 	Mon, 20 Apr 2020 12:51:15 +0200
De : 	François Bonnarel <francois.bonnarel at astro.unistra.fr>
Pour : 	apps at ivoa.net



Dear all,

We have to go forward on this.

I agree that #see-also covers nearly everything we can imagine in the 
contexte of Datalink as was my initial proposal of "#associated_data" in 
VEP001.

I think it's true that something like what Markus defined as sharing the 
same "progenitor with #this" and that he called #sibling is needed.

But )

   -   Due to the   "non intuitive" terminology, I suggest a synonymous 
: #co-generated.  This is directly inspired by Provenance dataModel, a 
new IVOA recommendation dated  April the 11th this year. If you share a 
progenitor with somebody else, this means you have been *generated *by 
the same *activity* with this guy and that this activity *used *the 
same*entity,* you progenitor.

   - This doesn't cover everything we had in mind originally with 
"associated data" in the sense of VizieR, sepcially when #this is a 
source in a catalogue and not a dataset. Several proposals have been 
made. Pat proposed "contains" and "followup" and I Personnaly proposed 
""Observation_Result_of_source" and also "cross-correlated". none of 
these seemed to be generic enough to satisfy everybody. But I do think 
we need some term for linking something to #thing which is scientific 
data (that is not a process, not a preview, not a calibration file, no 
auxiliary data and not additional medata) and which is neither 
progenitor nor derived and nor sibling/co-generated.

       To illustrate this look at the VizieR example below. this is a 
catalog of Spitzer galaxies with link to their optical counterpart 
image: "Optical imaging for the Spitzer Survey of Stellar Structure in 
Galaxies. Data release and notes on interacting galaxies.".

As explained in the abstract 
(https://cdsarc.unistra.fr/viz-bin/cat/J/A%2bA/569/A91) a link is 
provided to opticals images of various origins for all galaxies 
extracted from Spitzer IR information. This kind of relation between the 
Spitzer galaxy and its optical counterpart image is not covered by 
sibling/co-generated, derived, or progenitor.

      I propose something like "#other" or "#alternate" ( the latter was 
already proposed by Markus ..... in 2015 !!! semantics  session during 
interop In Sydney)

      In addition i think #derived,#progenitor, #co-generated (#sibling 
? ) and #alternante (#other ?) should share the same "head term"  
"#science".

Cheers

François




Le 23/03/2020 à 13:33, Markus Demleitner a écrit :
> Dear Apps,
>
> Over in DAL, we've tried define a new concept in the datalink
> vocabulary -- see
> http://mail.ivoa.net/pipermail/dal/2020-February/008261.html.
>
> The use case that started this is that the Gaia astrometry satellite
> produces spectra and time series for a subset of the objects it
> measures, and the idea is to use datalink to link the files to
> catalogue rows using datalink.
>
> The question essentially is: What is "the" relationship between the
> astrometric record and the spectra (the quotes are there because
> there are of course many such relationships, but we'd like to find
> one that's at the same time characteristic and generalisable)?
>
> And more importantly: What are other cases like that, and what would
> applications like to see so they can display these extra links at the
> right moment (and not display them when inappropriate)?  That's why
> I'm here: I'd be grateful for application writers' thoughts about our
> (Semantics) problem.
>
>
> The thread above somewhat petered out when considering two
> alternatives:
>
> #sibling -- essentially, this marks artefacts that came up from the
> same basic measurement as #this; it nicely complements #progenitor
> and #derivation that we already have in datalink/core.  Against it it
> has been argued that it's not very intuitive to people who don't
> think in provenance terms too much.  And that it perhaps is a concept
> that's not very useful in practice (i.e., doesn't let clients do
> things users would like them to do).
>
> #see-also -- the idea is that it's "data people might be interested
> in when looking at #this".  The trouble is that once I try to have a
> clear definition for what this actually means I fail.  You see,
> essentially all datalink terms would fall under #see-also, and I
> couldn't find a good criterion to exclude them.
>
> So, if we do something along the lines see-also, it would have to be
> see-also-if-X, where X could be,  perhaps, "scientifically exploiting
> #this" or "you have found something interesting in #this and would
> like to learn more about it" or whatever.  I think you'll agree that
> these proposals are... uncovincing.
>
> After this, I'm essentially back to #sibling, which is well-defined
> and (IMHO) easy to explain, which, as concepts go, is great.  It's
> also been useful *for publishing* in several other data collection,
> such as when people did line maps for a spectral cube.
>
> But is it useful *when consuming*, i.e., when doing user interfaces?
> If you think it's not, what concept *would* be useful to you?  Do you
> perhaps have an idea how a see-also-if-X concept could usefully be
> built?
>
> Thanks,
>
>            Markus


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.ivoa.net/pipermail/semantics/attachments/20200420/b9569eb2/attachment-0001.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: opabbbnlpbennghi.png
Type: image/png
Size: 293348 bytes
Desc: not available
URL: <http://mail.ivoa.net/pipermail/semantics/attachments/20200420/b9569eb2/attachment-0001.png>