Multiplicity typo in SimDM

[Gerard Lemson]

Hi Kristin

On Tue, May 23, 2017 at 5:21 AM, Kristin Riebe <kriebe at aip.de> wrote:

> Hi all,
>
> when looking at SimDM at implementing it (well, parts of it) for our
> simulations, I realized that one of the multiplicities in the UML
> diagram (and its description) at
> http://www.ivoa.net/documents/SimDM/20120503/html/SimDM.html
> is not correct.
>
> The multiplicity for the composition relation from InputDataset to
> InputDataObject is given as 0..1 (at the end), but it should be 0..*. 1
>
> Each InputDataset can consist of 0 to many InputDataObjects (and each of
> them has a reference to 1 DataObject).
> That's also what the annotation to InputDataset.object says:
> "Collection of object associations that identify explicitly which
> objects from a collection of objects are used as input data in an
> experiment."
> I had discussed this already with Gerard, and he agrees that this is a
> typo.
>
I still do.

>
> Since these classes and their multiplicities are not explicitly
> mentioned in the standard document, could someone please just correct
> the vodml-version and re-generate the html-description and diagram? Or
> is a minor version update/erratum needed for this?
>
>
My proposal would be to actually create a proper VO-DML version of SimDM
and do the fixes there.
The existing representations of SimDM are MagicDraw and an "intermediate"
XML representation. The latter was produced by and used in VO-URP, which
was also used to generate HTML and XSD representations of the model.
As VO-DML is VERY close to the VO-URP representation, a port will be
straightforward (i have one already, though need updating to latest
VO-DML). This update will contain:
- (VO-)UML diagram in magicdraw, conforming to the VO-DML profile. This
will replace the UML conforming to the VO-URP propfile in the
- VO-DML/XML representaiton generated form the UML.
- HTML documentaiton generated from the VO-DML/XML.
At the moment we do not have an automated mapping from UML to XML schema
(XSD), but that is not a requirement.

What do theory and dm chairs think about this proposal? An extra benefit of
a Vo-DML representation could be its usage in SimDAL or so, if that wants
annotated VOTables. With the mapping spec in progress that would be a nice
use case.

I can update the vo-dml/models/simdm/ folder on volute with these artifacts.

Cheers
Gerard

Cheers,
> Kristin
>
>
> P.S.:
> In the diagram at
> http://www.ivoa.net/documents/SimDM/20120503/uml/SimDM_DM.png (I guess
> this is a direct export from MagicDraw?), there is no multiplicity given
> and thus it probably defaults to 0..1 and thus the auto-generated
> diagram and description at
> http://www.ivoa.net/documents/SimDM/20120503/html/SimDM.html contain the
> 0..1.
>
> indeed that is the case.

>
> P.P.S: There are also still some "TODO" flags in annotations to some
> classes, maybe these could be filled with some meaningful text as well.
> (e.g. ObjectCollection, OutputDataset, all attributes of DataObject)
>
>
> --
> -------------------------------------------------------
> Dr. Kristin Riebe
> E-Science & GAVO
>
> Email: kriebe at aip.de
> Phone: +49 331 7499-377
> Room:  B6/25
> -------------------------------------------------------
> Leibniz-Institut für Astrophysik Potsdam (AIP)
> An der Sternwarte 16, D-14482 Potsdam
> Vorstand: Prof. Dr. Matthias Steinmetz, Matthias Winker
> Stiftung bürgerlichen Rechts
> Stiftungsverzeichnis Brandenburg: 26 742-00/7026
> -------------------------------------------------------
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.ivoa.net/pipermail/theory/attachments/20170524/e4bfaf3d/attachment.html>