STC2 Docs available but do you plan to provide xml schema ?

[Gerard Lemson]

HI Baptiste

On Fri, May 26, 2017 at 3:50 AM, Baptiste Cecconi <baptiste.cecconi at obspm.fr
> wrote:

> Dear Gerard,
>
> thank you for your answer.
> It is a bit unfortunate that there is no XSD instance of STC2 available.
>
> We would like to check the new model version by trying out simple XML
> instantiation for our VOEvent project and compare with the previous STC
> model version.
>
I'm afraid we won't have the time to do through XSLT set up and debugging
> any time soon in our project (nor Pierre nor other persons) and I don't
> have the skills to do so.
>
> The "entry fee" seems to be quite high just to test if the new version of
> the model solve the problems we have in the previous one.
>
> Might be very useful for the STC modeling effort as well as for a possible
XSD mapping if you could give an indication what those problems were.


> I hope that we can get a working solution soon!
>

That will depend on DM chairs and especially availability of people to do
the work.

I am interested to have a look at XML schema generation, but my priorities
are elsewhere so I won't have much time either.
And whatever the XSLT would be to generate XSD from VO-DML, for sure the
resulting schema would be very different from the previous one. (No
susbsitutionGroup-s! No element-refs!)

A custom XSD for STC could be created no doubt.
Seems to me that VOEvent does have its own STC-like definitions already
though. Are there plans for incorporating STC in VOEvent? Is what VOEvent
offers not sufficient?

It might be interesting to see if the VOEvent schema could somehow be
annotated with VO-DML elements similar to the VOTable mapping spec. To
indicate that some of its types have representations in STC/VO-DML. A
generic-XSD->VO-DML mapping (as opposed to VO-DML->XSD mapping) is a much
bigger topic, but could use that as an example.


Cheers

Gerard

>

>
> Cheers
> Baptiste
>
>
>
> Le 25 mai 2017 à 19:49, Gerard Lemson <gerard.lemson at gmail.com> a écrit :
>
> HI Pierre
>
>
> On Mon, May 22, 2017 at 11:12 PM, Pierre Le Sidaner <
> pierre.lesidaner at obspm.fr> wrote:
>
>> Hi
>>
>> I have seen the description of the STC2 DM at
>> http://volute.g-vo.org/svn/trunk/projects/dm/STC-2.0/doc
>>
>> I have seen also xmi files model directory
>>
>> My question is mainly for VO-DML group, is there a possibility to have
>> STC2 express in an xml schema (xsd) ?
>>
>> It would be very useful to make usage of it and validate.
>>
>> Does VO-DML have tool to transform the serialisation of schema into such
>> a format ?
>>
>>
> There is currently no script to translate from VO-DML to XSD.
> But nothing stops one from creating one, the mapping between VO-DMl and
> XSD is quite straightforward. See appendix B1 of the VO-DML spec for an
> example.
>
> In fact there is such an XSLT script for the related VO-URP "spec" at
> https://github.com/glemson/vo-urp/blob/master/res/intermediate2xsd.xsl
>
> This was used to create the XSD representation for the Simulation DM  in
> http://www.ivoa.net/documents/SimDM/20120503/xsd/
> It's source was what we then called the intermediate representation" of
> the data model in
> http://www.ivoa.net/documents/SimDM/20120503/uml/SimDM_INTERMEDIATE.xml
> See section 4.2 of the simulation data model (http://www.ivoa.net/
> documents/SimDM/20120503/index.html) for a short description.
>
> The intermediate XML is closely related to the VO-DML/XML. Hence it should
> not be hard to modify the XSLT to also work for VO-DML, and I have planned
> to do that once I found the time. Maybe someone else can use it as
> inspiration?
>
> Main difference between VO-DML and VO-URP is that the former can import
> models and in fact the primitive datatypes are defined in the base ivoa
> model in VO-DML, but as an enumeration in VO-URP.
>
> And also note that a particular mapping scheme was used in the XSLT.
> Laurent Bourges and I used that in a web application that translates
> between XSD, Java and RDB, but it may not be to everyone's liking. In
> particular the treatment of references was tricky. We also made the
> explicit decision to separate the type definitions from the definitions of
> the root elements. And those were limited to correspond to the non abstract
> root types only, i.e. those that are not the child in a composition
> relation.
>
> I am happy to assist in porting the XSLT from VO-URP to VO-DML, but hope
> others are interested as well.
>
> Cheers
>
> Gerard
>
>
>
>
>
>
>> regards
>>
>> Pierre
>>
>> --
>> -------------------------------------------------------------------------
>>                            Pierre Le Sidaner
>>                         Observatoire de Paris
>>
>> Direction Informatique de l'Observatoire
>> Observatoire Virtuel 01 40 51 20 82
>> 61, avenue de l'Observatoire 75014 Paris
>>
>> mailto:pierre.lesidaner at obspm.fr
>> http://padc.obspm.fr
>>
>> ------------------------------------------------------------
>> --------------
>>
>>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.ivoa.net/pipermail/dm/attachments/20170526/6d4225c3/attachment.html>