Revised ImageDM-ObsCore architecture

[Laurino, Omar]

So, if I can try to summarize the two approaches we are discussing, we have:

1. Build on top of ObsCore and integrate the new data models (Spectral 2.0,
major release change, and Image DM, new standard) with ObsCore to support
the Data Cubes.

2. Change Obscore, with changes that might break backward compatibility
and thus require a new major release. Then integrate new models  with it.

Both approaches have the same goal and, as far as we can tell, they would
both achieve it.

The definition of backward compatibility I am using: a file compliant with
the new version of the model is also compliant with the old version, so
that software designed to support the old version does not choke on new
files.

I would like to hear people's opinion on this.

Cheers,

Omar.
On Nov 13, 2013 12:13 AM, "Douglas Tody" <dtody at nrao.edu> wrote:

> On Tue, 12 Nov 2013, CresitelloDittmar, Mark wrote:
>
>  Just a note on this point.. side bar though it is..
>>
>> On Tue, Nov 12, 2013 at 8:38 PM, Douglas Tody <dtody at nrao.edu> wrote:
>>
>>  year and we urgently need to get on with development for real projects,
>>> with some expectation that the standards will not completely change out
>>> from under us while this is going on.  Some sort of interim
>>> update/solution is not an option at this point.
>>>
>>>
>>>  My view on the document to be turned over Friday, is that this is the
>> next
>> DRAFT
>> of the ImageDM which addresses the ObsCore link and other feedback from
>> the
>> previous draft.  I expect the utypes list will reflect the work you have
>> been doing
>> with Francois and Mireille.  This drop is to provide the basis to support
>> the work
>> you mentioned.
>>
>
> I agree with all this - it is just an updated draft, and will of course
> continue to evolve.  However we need something stable, hopefully close
> to the final architecture, integrated with SIAV2, to support Cube
> project development for at least the next 4-6 months while the standards
> evolve.  The outside groups involved in implementation cannot deal with
> our continually evolving internal standards development process.  An
> occasional, e.g., yearly update is probably ok.
>
>  From that point, I will take lead on it's further evolution.. wherever
>>> that
>>>
>> goes.
>> I plan to make a list of topics to discuss, from my own review, and the
>> comments
>> by others (Jose, Arnold, Mireille).  I cannot guarantee that there will
>> not
>> be
>> significant changes to the overall model after this drop.
>>
>
> Of course there can be no such guarantee that things won't change
> significantly in response to an open discussion, I just ask that we try
> to minimize the impact on outside groups once we start to have real
> take-up outside the VO projects.  I would hope that NRAO and ALMA (not
> mentioned above) continue to have some input into further evolution as
> well; we are possibly the largest producer of cube data directly
> involved in this project, have a critical and immediate need for these
> standards, and will be doing a lot of relevant development over the next
> year.
>
>         - Doug
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.ivoa.net/pipermail/dm/attachments/20131113/d8816408/attachment.html>