[Uml-devel] news on 2.crap

[Andrew Sutton]

> i've stumbed across another sticky problem that i can't quite resolve - the
> refMetaObject() method. i'm not sure how to tie an object to its meta
> object. it would be really, really convenient for a lot of things if i
> could figure it out. i don't know. i've only been thinking about it for a
> day or two... oh well...

update on that problem. i've found a convenient way around it. i read this 
paper on a formal MOF proposal that seems to think that an explicitly defined 
MOF would eliminate the need for a reflective layer. interesting. anyway, i 
can build the rest of this stuff by pretending that the RefClass related to a 
RefObject is the metaobject. of course, i'd have to modify all the RefClass 
implementations to contain, shall we say... certain structural features of 
the class? fine be me. any complaints?

hmm... the more i implement this crap, the more deviant i get from the 
standards. not too far afield though. just enough to make it interesting.

andy