Time Series Cube DM - IVOA Note
mdittmar at cfa.harvard.edu
Tue Feb 21 19:18:07 CET 2017
No worries about overstraining the time budget.. my primary goal is to get
of models out there that are useful and usable.
The thing is, that when I read through your message, just about everything
is how the instance looks. Dataset stuff is annotated as ds:blah, STC
stuff is annotated
as stc:blah. Since the model prefix does not change with minor version,
the addition of elements from a minor version update should be a no-op...
there would be
annotated content that an implementation of earlier sub-versions would not
But the data model itself (NDCube), if it uses/extends from other models
identify which models it uses.. by URL. This URL is used in the vo-dml
of the model among other things.
I think this version connection is the only point to be concerned about.
+ if a Datatset-2.0 came out, you could annotate your file with both
Dataset-1.x and 2.0 annotation,
(I think... that is a question for the Mapping folks)
+ But, in validating the instance, the version which is associated with
the model would
be the one used (Dataset-1.x). The 2.0 content would not be
'official/valid' until the model migrated
to be based off the new version.
I am generating full VOTable instances of chandra event lists using ds,
You've gotten some pointers to this work on-the-side.
This week I'm formalizing the vo-dml annotation to the Mapping spec. After
we're going to put the package out to the group for formal review. The
goal on this is the end of Feb.
On Mon, Feb 20, 2017 at 4:09 AM, Markus Demleitner <
msdemlei at ari.uni-heidelberg.de> wrote:
> Sorry for this somewhat verbose marketing gig, but since I believe
> this is the centerpiece of what I've been hoping for as the VO-DML
> dividend I thought I might beg forgiveness for overstraining your time
> -- Markus
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the dm