DM Running Meeting: 20241009 - Notes

CresitelloDittmar, Mark mdittmar at cfa.harvard.edu
Thu Oct 17 17:46:06 CEST 2024


All,
Thanks to all who participated in the running meeting last week!  There was
good coverage and conversation over the agenda items.  We have a few
actions coming from that.. so please follow-up there and keep the momentum
building.

Below are some notes from the meeting.. I'll put them in-line with the
agenda *(in bold)*, and a summary of actions.  My apologies for any errors
in interpretation.. feel free to make corrections.

I've also created a Twiki Page
<https://wiki.ivoa.net/twiki/bin/view/IVOA/IvoaDMRunningMeetings>(following
the DAL example), to coordinate the Running Meeting information, agenda,
and notes.
(I'll add a link to these notes on that table when they appear in the email
archive.)

*Action Summary:*

   - MS: schedule and setup regular DM Running Meetings going forward
   - MCD: follow-up on DM session contribution request by YX
      - DONE - this contribution will be in the Plenary panel session.
   - PH: pull some/all, hopefully non-controversial, items from the VO-DML
   1.1 milestone as candidates to move forward.. announce these to the DM
   mailing list for consideration/discussion.   Goal to have Working group
   consensus items identified, defining the scope of 1.1 update for Malta.
   - MCD: bring the 'Shape' discussion to the DM mail list, with goal to
   gain consensus on where this model element should live in the DM ecosystem
   - ??: bring vo-dml 'version' attribute to DM mail list for discussion..
   at least to form a plan IF the attribute is, in fact, needed for upcoming
   vo-dml standard updates.
   - LM:  Announce to the DM mail list when the MANGO model document is
   ready for general review.
   -


*Attendees:*
    Mark CD, Mathieu Servillat, Pat Dowler, Marco Molinaro, Gerard Lemson,
Mireille Louys, Laurent Michel, Paul Harrison, Grégory Mantelet, Janet
Evans, Erin Brassfield Bourke

*Agenda:*

   - Running meetings going forward:
      - I’d like to set a fixed day/time each month to pre-schedule the
      running meetings going forward.  *[ACTION: MS to setup/schedule]*
   - Interop Prep:
      - Programme page:
      https://wiki.ivoa.net/twiki/bin/view/IVOA/InterOpNov2024
         - DM session scheduled for Saturday Nov 16; 11:00 local time.
         - DM Session Page:
            - Responses to “Call for presentations” so far
               - Paul H:  VO-DML tools/Std update
               - Laurent M: MANGO updates/status
               - Mark C-D: Follow-up on Joint session topic
            - HEIG
               - Ian Evans:  possible contribution coming
            - ??
               - Yunfei Xu: a talk on Einstein Probe (EP) scientific data
               management
                  - ?? DM the right place for this?

*                  NOTES:*
*                      o ACTION: MCD to follow-up with Yunfei on where to
best fit his contribution.*

   - CAOM Integration
      - Document uploaded to IVOA D&S page
         - quick summary of what that contains (any changes made toward
         integration goal)
         - what is needed from DM group now?
         - where do we go from here?

*                  NOTES:*
*                      o PD gave a good overview of the current document
status*
*                            o Moving toward an IVOA model for CAOM.. has
community buy-in*
*                            o Changes to CAOM-2.4 are posted on D&S page,
includes*
*                                 o support for radio*
*                                 o reconciles some items with IVOA
practices*
*                                 o includes use cases from MAST, IRSA,
IPAC*
*                            o The document is rough; *
*                                 o **has a lot of editorial (LaTex)
updates to make (current doc. target format is HTML)*
*                                 o needs introduction and abstract...*
*                                 o work to do on vocabularies modeling and
how to handle new vocabularies*
*                            o Going Forward:*
*                                 o doing the cleanup (PD); and
broadcasting to the community for comment*
*                                 o At Malta: Looking to have a brutal
discussion on the details.*
*                      o It is intended that the CAOM model (DM) and
associated TAP interface (DAL) are kept separate*
*                      o CAOM - ties into the 'Shape as DataType'
discussion (see VO-DML)*


   - VO-DML: https://github.com/ivoa/vo-dml
      - Schema change:
      http://mail.ivoa.net/pipermail/dm/2024-August/006502.html
         - History review:
               - 2018: email discussion about having a ‘version’ attribute;
               the attribute appears to have been added to volute
document at that time
               - 4/23/2018: attribute added to Git schema (commit 4933)
               - Unfortunately, the version that went to REC did not
               include the document snippet or the schema version
which included it.
               - The attribute was removed from the Git schema in 2021
               during a sweep by Paul H. to match the published schema.
            - Do we add this to the 1.1 list?
      - 1.1 candidates:
         - ROADMAP item:  Reconcile STC
         <https://wiki.ivoa.net/twiki/bin/view/IVOA/STC> (Region), FoV
         <https://wiki.ivoa.net/twiki/bin/edit/IVOA/FoV?topicparent=IVOA.2024ARoadmap;nowysiwyg=0>
         (Shape), DALI <https://wiki.ivoa.net/twiki/bin/view/IVOA/DALI>
         (dtypes), CAOM (dtypes)
            - this came up multiple times during the interop and shouldn't
            be too difficult to make a plan
            - STC <https://wiki.ivoa.net/twiki/bin/view/IVOA/STC> (Region)
            -- separate shape from Frame => STC
            <https://wiki.ivoa.net/twiki/bin/view/IVOA/STC>-S serialization
               - Note: we don't necessarily have to update STC
               <https://wiki.ivoa.net/twiki/bin/view/IVOA/STC>, just recast
               STC <https://wiki.ivoa.net/twiki/bin/view/IVOA/STC>-S to
               refer to the replacement models.
            - Git Issues: (Tagged with V1.1 Milestone)
            - These have been on the table for some time, and reported by
            PH at various interops.  They have not gotten any attention
            within the working group beyond that.  I think we should get
            the short-list to the working group for review/comment
with dates/goals to
            push a V1.1 update through.
               - #44: Grouping of attribute composition and reference in
               ObjectType and DataType
               - #43: Update Appendix B Mapping to serialization meta-models
               - #41: Document rules for ‘contained’ references
               - #37: specify that datetime string serialization to conform
               to ISO8601
               - #20: update VO-DML Standard document to 1.1
               - #18: make the form of vodml-ids mandatory
               - #14: Schema Generation (XML and JSON)
               - #5: primitive type for ivorn
               - #4: way to indicate ‘natural’ primary key

*                  NOTES:*
*                      o MCD: suggested the VO-DML presentation at Malta be
more directed as a plan for moving V1.1 forward with change-list, status
and roadmap.*
*                            o ACTION: PH - pull some/all, hopefully
non-controversial, items from the VO-DML 1.1 milestone as candidates to
move forward.. announce these to the DM mailing list for
consideration/discussion.   Goal to have Working group consensus items
identified, defining the scope of 1.1 update for Malta.*

*                      o MCD: opened discussion on adding the 'shape'
datatypes in the "ivoa" model.*
*                            o PD: in DALI/CAOM these are expressed as
xtypes.. and items that can be expressed that way are good candidates as
base types.*
*                            o PH: adding this for 1.1 seems rather
rushed.. the other items on the milestone have been implemented for some
time.*
*                                      - reminder that there is a
'playground' in the vo-dml toolkit where things like this could be tried
out.*
*                            o GL: suggested this is really more of a
serialization issue.  Not sure VO-DML needs to care about that.*
*                                       - serialization (as xtypes) should
be outside of the ivoa model.*
*                            o PD: pointed out that the Shape datatype is
already defined and implemented in the DALI standard, *
*                                      so this is a relocation of
implemented content and not starting from scratch.*
*                            o **General discussion that this concept is
one that is currently appearing in multiple models and models in progress.*
*                                If it is not added to the ivoa model base
types, we should consider separating it as a core model just above the base
types level which *
*                                can be reused in various contexts.*
*                                       - eg: this model would describe the
shape datatype, the various models (FOV, DALI, ...) would use that type and
add context such as *
*                                                an associated coordinate
system, creating a spatial Region of a particular Shape.*

*                            ACTION: MCD - bring the discussion to the DM
mail list, with goal to gain consensus on where this model element should
be *
*                      o MCD: opened discussion on the 'version' attribute
topic*
*                            o version attribute is the 'version of vo-dml'
which is different from the version element in vo-dml which is the 'version
of the model being described'.*
*                            o PH: initial impression is that this isn't a
necessary addition*
*                            o PD: is not sure this is needed, it depends
on when model changes produce a change in namespace.*
*                            o GL: what if the schema doesn't change, but
the document does?  Would we want a new namespace for the new content?*
*                            I feel like there was an action here, for this
to be discussed on the DM mail list, but I don't have any written down.*
*                               - I recall expressing concern that we get
well into the V1.1 process and discover that the attribute is needed..
causing a delay.*


   - Mango:  Status update by Laurent

*                  NOTES:*
*                      o LM:  Working draft is published, refinements are
on-going.*
*                      o ACTION - LM:  Announce to the DM mail list when
the document is ready for general review.*


   - Formally Expressing Model B as a “view” of Model A
      - MCD - summarize where we are and what the plan is going toward
      Malta and beyond
         - This would probably benefit from a separate/dedicated meeting
         prior to the interop.
            - We had discussed a meeting of some sort in Sydney which I did
            not follow-up on to make happen.. but a short working “jam
session” would
            be helpful here.

*                  NOTES:*
*                      o MCD: gave a brief overview of current work..
mostly around mapping the MANGO:EpochPosition element to the 'parent'
Coordinates model*
*                      o GL: this is a very hard problem to solve, asks
what is the goal of this mapping?  What are we going to do with it?*
*                             MCD response: My concern is that IF we allow
models to define objects which overlap in concepts with other models, *
*                                  but catered to particular clients/cases,
we want to ensure that the content is consistent.  Not doing so will put us
in the same position we *
*                                  have had with ObsCore, STC and
Spectrum.. all with inconsistent/overlapping content.*
*                                  For example: I'd like to be able to
automatically **verify that *
*                                        o mango:EpochPosition is, in fact,
a valid consolidation of Coordinates model elements.*

*                                        o obscore is a consistent 'view'
of CAOM and/or DataSet*
*                                        o caom content which overlaps
DataSet, Cube etc concepts is consistent.*
*                      o MM: reminder that Baptiste is a good resource for
plugging into the Ontology community*


   - AOB
   - IVOA Note for FITS serializations:
      http://mail.ivoa.net/pipermail/dm/2024-June/006476.html
         - discussion started as question about lightcurve data model, but
         was concluding with a proposal to develop an IVOA Note with
‘preferred’
         serializations for various products in FITS format.
         - This project would require review/assessment of the ‘big
         players’ currently in practice and define a recommended
serialization.
         This would probably include a dictionary for FITS keywords for
         metadata at the Dataset/ObsCore level.
         - Kelle Cruz was very keen on this project and, I think, would be
         willing (eager to contribute)

*                  NOTES:*
*                      o This topic was not covered in this meeting.. ran
out of time.*
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.ivoa.net/pipermail/dm/attachments/20241017/9ee32213/attachment-0001.htm>


More information about the dm mailing list