[Uml-devel] Re: Umbrello design (was UML XML tree viewer)

[Andrew Sutton]

> - Our projects are made of objects which are instances of UML object,
> which themselves are instances of MOF objects. All of which seem to
> have, at the (meta)*model level only getter/setter methods. Why not
> another level which has only 1 class, of which MOF objects are
> instances. Let's call it libOOTU (Lib Origin Of The Universe), which we
> could define for the purpose of our project. Sorry, that may be a dumb
> academic point of view, but I see now that I have been playing with only
> a few snow balls of the UML iceberg...

i mislead you in my previous answer.

for the domain of modeling, MOF is the root of all evil/origin of the 
universe. it ends there. so, we end up with 4 levels of "metaness".

M3. meta-metaclass (mof - instances are UML classes)
M2. metaclass (uml - instances are C++ classes)
M1. class (src code)
M0. object (binary code)

the reflective interfaces i mentioned are applied by the MOF standard to M1 
level to provide a "gateway" to the M2 level. Qt does something similar to 
this with QObject::metaObject() except its at the M0-M1 level.

anyways, i think you can stretch the definition to provide a gateway from any 
instance to any class. the instance must be derived from a reflective 
interface.

the standard defines a couple: RefObject, RefAssociation and RefPackage. 
There's an MOF to Java mapping specification that adds a bunch more, 
RefEnumeration, etc. the idea is that when you generate code from an MOF 
file, one of these Ref* objects becomes the first super class of the object 
being instantiated. e.g. for UML:

class UMLClass :
	public RefObject,
	public Classifier,
	public GeneralizableElement
{...};

anyway, in order to correctly generate C++ from MOF specifications, we need to 
adapt the MOF to Java specification for our purposes. that should define how 
the code is generated. then we extend it to include libXMI required features 
and possibly umbrello required feature and we should be done. yeah!

the only conceptual problem i'm having is this: if UML has reflective features 
that allow us to access their meta object definitions, aren't we required to 
implement MOF? i guess we could just return NULL if they try to get the meta 
object. i can't really even think why they might want to do that. oh well.

back to the design.

andy