mlhim-owners team mailing list archive

[Tim Cook <timothywayne.cook@xxxxxxxxx>]

Hi Marcos,

Thanks for the prompt response. 

On Sun, 2011-02-27 at 13:50 -0300, Marcos Didonet Del Fabro wrote:

> The solution is to manually edit the Ecore adding these packages.
> However, we lost the synchronization between the Ecore and the XML
> Schema, so new modifications in the XSD will need to be manually
> handled in the Ecore.

I am not concerned about synchronization as long as we can regenerate
new schemas from the Ecore. Which as I understand it, that is possible.
So making the Ecore the "source" for any changes is a good thing to do
if we can generate new XSDs based on those changes. 

> I didn't understand this very well. You mean they are not shown in the
> diagram? Attributes that are containment or association relations are
> not shown in the class, but are shown as connections between the
> different types. Is that what you mean?

Okay, yes I see them now.  However the state of the diagram makes that
information impossible to analyze. 

> It won't be different if we use UML modeling. I think I am missing
> some point here..

Probably my misunderstanding then. 

> The EDataTypes are used only for primitive types, such as String,
> Integer, etc. The other classes can inherit from any other class.
> A more elegant solution would be to create separate types, e.g.,
> String, Integer, Double (without the initial "E"), and set up the
> property "implementation class"; This will do the correspondence with
> the Java implementation. Still, we loose the automatic synchronization
> between the XSD and Ecore. I would need to study a bit more to be sure
> if this synchronization could still be assured.

As long as we can generate valid XML Schemas from the Ecore model that
do not depend on datatypes not listed in
http://www.w3.org/TR/xmlschema11-2/ it is good.  The concept is that all
other programming languages can generate bindings to these schemas via
existing tools to create their own implementation of the reference
model. For example, we would use PyXB http://pyxb.sourceforge.net/ to
generate a compliant reference model from the schemas for Python.  Then
use Ruby LibXml http://libxml.rubyforge.org to generate Ruby bindings
from the same schemas. 


So in order of authoring/precedence we would have:
All changes are modified in the Ecore model (diagram?) -> generate new
schemas -> generate new reference models for he various programming
languages.  

Maybe the Java code gets generated directly from the model.  But I need
to know that this works for the other languages. 

=======================================================================

So to cleanup the existing model do we need to create the packages and
then the classes from scratch?  Is there some way to create the packages
and then import just that XSD for that package?  I tried creating a
package and then moving a class.  But it seems to make a copy of the
class and all of the associations are lost.  

Thanks again for your help.

Regards,
Tim



-- 
***************************************************************
Timothy Cook, MSc
Project Lead - Multi-Level Healthcare Information Modeling
http://www.mlhim.org 

LinkedIn Profile:http://www.linkedin.com/in/timothywaynecook 
Skype ID == timothy.cook
Academic.Edu Profile: http://uff.academia.edu/TimothyCook

You may get my Public GPG key from  popular keyservers or    
from this link http://timothywayne.cook.googlepages.com/home

Attachment: signature.asc
Description: This is a digitally signed message part