DM Running Meeting Notes: 20251031
CresitelloDittmar, Mark
mdittmar at cfa.harvard.edu
Thu Nov 6 19:13:50 CET 2025
All,
Notes from the recent DM Running Meeting (20251031)
Attendees: Mark C-D, Mathieu S., Pat D., Janet E., Mark K., Mireille L.,
Tom D., Renaud S., Francois B.
Actions:
- MCD: Follow-up with LM about APPS session presentation on new
implementation for MANGO
- MCD: Send email to 'interop' advertising the DM Session and request
for attendees to come prepared to talk about the details.
- MCD/MS: put together a DM Session page for the programme.
- MCD: work the Spectrum erratum; and finalize the RFE issue in Git.
Agenda: (Notes are in PURPLE)
- In-person attendance
- will not be in-person as planned (or at the ADASS component),
travel authorizations denied.
- Session Content
- Laurent offered to do a MANGO update, but not sure this is useful..
- MCD: agreed, not sure if he has an Apps presentation lined up.
- TD: an Apps presentation on the new implementation would be
good, and maybe helpful in raising awareness/interest in the workflow.
- ACTION - MCD : Follow-up with LM about APPS session presentation
on new implementation.
- Paul H. offered VO-DML 1.1 status - document and toolkit
- Good; ties-in with discussion regarding base model updates
requested by CAOM/ObsCore projects
- 15 minutes ??
- Pat D. CAOM-2.5 - new revision ready before interop; probable
migration to ivoa-std,
- PD: status update, document complete.. polishing; migrated to
ivoa-std yesterday
- Severin did analysis of holdings to assess usage of different
items of the model.. possible deprecation candidates.
This can be a
discussion topic, but probably premature, and detailed for
the interop
session. Will probably just note that it has been done
and point people to
it in the presentation.
- MCD: maybe good for a special meeting, or Running meeting
post interop
- PD: yeah
- TD: this is a bulky topic, so a separate discussion may be
best. There are some people on his end that would be
interested in that.
- MCD: is Running meeting going to have the right scope? Maybe
we need to focus the invitation to be sure to get the
right people in the
room.
- PD: sounds good, would also want to record the meeting.
- TD: serves as dis-usecases
- includes vodml updates; currently based on v1.0 with
secondary model covering requested changes. (if I
understood that correctly)
- MCD: How much time would you like?
- PD: 15-20 minutes should be fine.. can do 15 if needed.
- ObsCore-1.2 + extensions
- A working meeting to discuss the data model changes needed to
support the Radio+HE+TD notes
- ObsCore - assessing recommendations for scope and conformity
with domain-of-interest.. alternative solutions?
- e.g. FB recent post regarding response functions in ObsCore
vs being available via datalink
- Extension Models - Single standard for various ObsCore
extensions? or individual? or other?
- VO-DML: datatypes model extensions requested by ObsCore-1.2
and CAOM-2.5
- DataProduct: modeling of single file product (Spectrum) vs
multi-file product ( EventList + IRF )
-
- MCD: general description of what the goals of this portion
would be.. outlined above.
- For this to work, people need to come prepared and informed
about the content of the Notes.
- MK: suggests sending a message to the interop list with the
Notes attached, advertising the session topic.
- MCD: is that too wide a distribution? This is a very
DM-specific meeting.
- TD: casting wide net is OK
- JE: As an FYI - HEIG will be doing prototypes (2) CXC and
CTA, which should follow shortly after the Note (targeting
end of the year).
- MCD: Do we want to hold off the session email until after the
next pass on the Note.. which is in progress?
- JE: Yes. Should be ready within the week. Could go out with
the advertisement and follow-up with the notes.
- JE: Need to make sure the people responsible for the models are
there..
- PD/ML: ported the current ObsCore to ivoa-tex.. haven’t
started 1.2 content
- ML will be attending the meeting remotely
- PD will be there in person..
- MCD: We should have about 1hr in the session for this topic..
IF people come prepared, this could be very productive.
- We will want to be sure not to get bogged down in
re-discussing things like real vs double/float
- Topics like hashing out the Dataset/DataProduct
distinction/modeling ( 1 file vs multiple files in 1
'bundle' ) would be
very useful/productive
- The comment re-raised by FB (noted above) about Response
Functions in ObsCore vs DataLink can be a time-consuming debate.
- JE: Ian has really studied this and the proposal agrees
with the current ObsCore description of 'Advanced data
products', and
shouldn't need a lot of discussion again.
- MCD: When we get to the concepts in current models which
aren't well exercised, the descriptions may not be very
detailed and/or
reflect the scope and goals of the model. The
1-file/n-file Datasets in
the Dataset model is another example.
- This may simply be a matter of improving the
description of Advanced data products in ObsCore,
but now is the time to
look at these cases which push the boundaries, and
decide on the proper
solution, and not just use ObsCore because it is there.
-
- Spectrum Model: DATE-OBS keyword RFE -
https://github.com/ivoa-std/SpectrumDM/issues/8
- MCD: this popped up again after submitting the notes from the last
running meeting.
- The preferred solution was pretty well agreed-upon in College Park,
but hasn't progressed from there.
- Need to make an effort to close this out.
-
- AOB
- PD: to touch back on an earlier topic. He has a different perspective
about IRF in ObsCore or DataLink..
- MCD interpretation of his description
- ObsCore == designed for discoverable product (Plane) , which
may have multiple files (Artifact) associated which are
provided via
DataLink.
- Plane/Artifact are CAOM concepts and relate to the DataSet
model Dataset vs Dataproduct ?? ( 1 vs multiple ‘files’ )
- It kind-of comes down to whether we're describing Response
Functions as a Plane or an Artifact.
- MCD: it seems like that would depend on how tightly
coupled they are?
- In general, we don't dictate the organization/representation
of user data, but establishing best practices may be in order.
- MCD: so this seems to boil down to how well are we enabling
interoperability.. and how much a user may need to know
about the data they
are looking for.
- JE: This part of the HEIG note is driven mostly by CTA data,
Mathieu will be there in-person so you can talk with him during
interop - then bring to HEIG
- JE: what is the state of the other models? Transform, etc..
Spectrum was to be a view of the 'cube'..
- MCD:
- Transform needs better documentation of the implementations..
then could go forward (matter of taking/making the time to do that)
- Dataset metadata: small dependency on Measurements to make
sure is accurate; and last review of the model called for
re-consolidating
the content to be less spread out .. not connecting to a
non-existent
Observation model. This might be worth doing before the
interop since it
may be affected by that discussion.
- Cube is going to need implementations; Given the experience
with Mango.. I'm not really seeing a clear path forward
with this yet.
- there are discussions in the TCG regarding what the
'clients' (Apps, DAL, etc) are looking for from the models.
-
- Next Meeting will be at the interop; will schedule follow-up Running
meeting.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.ivoa.net/pipermail/dm/attachments/20251106/21a4f7e3/attachment-0001.htm>
More information about the dm
mailing list