Originally posted by Ali Pope:
Hehe. Unfortunately i still do not agree. I do not consider that a piece of data that is used at runtime verification can be seen as related to reflection.
A good deal of your discussion in the forum has been about bytecode generation. I agree that it's a topic that's relevant to reflection. However, it's not necessarily more relevant than usage of these dynabeans. Specifically:
- Bytecode generation works around metaobject protocol (reflection API) limitations to changing and constructing classes inside the runtime.
- What you posted seems work around MOP limitations to modifying fields using the runtime.
True, the structure you showed in the posting I'm responding to does not look like something related. However, a bytecode generator suffers the same dissimilarities. In this case--it's the intent that matters. Both are trying to get around limitations in the MOP.
Best Regards,
Nate