updates to vo-dml spec

[CresitelloDittmar, Mark]

Gerard,

I have given the document a read, and have a few comments/questions which
I'll put here.
I will separately send you an annotated PDF (hope you can read it), which
highlights several type-os as well.

VO-UML
   + the document is very specific about the graphical display of each
item.. which is great.. but I suspect many of these are not reproducible by
the various modeling tools.
      For example:
      - the package name in Modelio is given within the 'tab', which is
also how it is described in the uml-diagrams.org link.. "The members of the
package may be shown within the boundaries of the package. In this case the
name of the package should be placed on the tab."
      - I doubt I can get a little triangle on 'model' packages, though I
can maybe add a 'model' stereotype.
      - for DataTypes, modelio puts a little "T" in a box in the upper-left
corner; enumerations get a "12.." box.
      - etc..

      So there are little differences.  How do propose these are dealt with?

Registry:
  + you mention in a couple places that the models, the docs/html and their
'vodml-id' are registered in an IVOA Registry.  This would be a new
requirement on the 'model' creation path?  So that "spec" is the 'official'
prefix of
the Spectrum Data model, with a registered location for its vodml
representation.

Subsets:
  + I haven't been able to figure out how to do this in Modelio.
  + possible use from my stuff:
     - At the moment, Dataset contains an abstract <Data> class which heads
the classes such as "NDPoint"
       There is currently no relation between them, because each 'type' of
Dataset would need to subset with
        a particular kind of Data and multiplicity thereof.
  + It isn't clear to me if the subsetting allows modification of the
multiplicity..
     ObsDataset allows 1:* CoordSys objects, can a QueryResponse restrict
that to 1:1? using subsetting?

Composition:
  + "an ObjectType can only be the target of at most one Composition
relation".
     Has big implications on model reuse.  For example, I cannot create an
object which includes Char:Accuracy in it's compositions, because it is
already in a composition relation with CharAxis.  (the example may be
wrong, but the flavor of the example is correct).  I'm not saying this is a
bad thing, and probably means that something in the models needs to be
given a good review.  But, it would be good to include in the doc how we
should deal with this.  In the above, would I have my model 'extend'
Accuracy (with no additional content), and make a composition relation to
that?

Reference:
  + does this allow a two-way M-N relation?

Multiplicity:
   + just to be clear.. this is restricting multiplicity for Attributes
(only) to 0..1, 1..1, n..n.
      other roles/relations (ie: composition) are unrestricted w.r.t.
multiplicity.

Section 4.1: Rules
  + this is one of the places it talks about registering the VO-DML/XML,
and HTML pages.
     Maybe this has all been covered in earlier talks.. but,
     what will we need to do to ensure persistence of this information when
a model becomes a recommendation?
     (This is all under 'googlecode' now, but would need to be in a static,
read-only and persistent location once
      the model is a recommendation.).


 Mark


On Sun, May 11, 2014 at 4:26 AM, Gerard Lemson
<lemson at mpa-garching.mpg.de>wrote:

> Dear All
> I have updated the docx and pdf versions of the VO-DML spec on volute
> (links
> below).
> I have fixed typos discovered by Florian and tried making at least some of
> the comments on the utypes spec less offensive following the email
> discussion with Francois. No doubt this is not the end of it. Please
> comment
> on the text and also watch the comments page at
> http://wiki.ivoa.net/twiki/bin/view/IVOA/VoDmlWGComments where I try to
> log
> individual comments.
>
> Cheers
> Gerard
>
> https://volute.googlecode.com/svn/trunk/projects/dm/vo-dml/doc/VO-DML-WD-v1
> .
> 0.docx
>
> https://volute.googlecode.com/svn/trunk/projects/dm/vo-dml/doc/VO-DML-WD-v1.
> 0.pdf
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.ivoa.net/pipermail/dm/attachments/20140514/5bbf498f/attachment.html>