Suggestion to replace use of generics with inheritence in future RM versions

Technical (archive)
21

Hi Heath,
Just to clarify: my original issue is not about auto generation of classes from XSD, whether or not one choses to do that is irrelevant if the XSD itself is not supporting generics. The problem is with the XSD (not having support for generics) and it may or may not cascade to programming language space.

Separately, I agree with you on not taking XSD and code generation as the basis of RM implementation, and I am neither following nor suggesting this route.

At the moment, I am connecting Rong’s implementation to JAXB, which is my choice of design to handle XML related functionality in Java RM.

I find these discussions very helpful, even though they have a tendency to fork, but hey, this is why have the groups :slight_smile:

Kind regards
Seref

(attachments)

OceanInformaticsl.JPG

22

Ok, thanks for the clarification.

Heath

(attachments)

OceanInformaticsl.JPG

23

David,

This is what I would realistically propose, for the CLUSTER/ELEMENT part of the model. I will also post a version with integrated changes - this change, plus the simplification of ITEM_STRUCTURE etc.

  • thomas
24

Two further variants are now up, with the second containing the more radical changes.

Comments welcome.

  • thomas
25

How true. That statement should be on the wall of every UML tool vendor's office!

- thomas

26

If any developer would do better than a tool, why would anybody write tools in the first place?

27

Developers are expensive, tools are reusable :smiley: