I just reread this message and want to apologize for its attitude. I was so excited to see another step in this useful technology. And men, am I disappointed. Sorry, again. Hope, you guys will gain something from this message.
Here it goes...
In theory AOP paradigm should give memory utilization gain. What I see from the description is memory consumption overhead introduced by interceptors. It's good for short length call chain. But, how do I use it for optimizing iterative logic? It was supposed to be able to recompose/optimize call chain not introduce a filter logic. I understand interceptors as one potential outcome of using AOP framework. But you seem to introducing them as a part of that concept.
Also, I don't quiet understand how do you see introductions could be used. I mean, does it bring us closer to AOP? You added an ability to extend object functionality. But it is still an object not an aspect. That looks like an object fusion not a logic fusion. If it did help to use AOP, it should be transparent to an aspect programmer. It should be just a detail of your implementation. Yet... It is another feature. Hm... Are you guys going easy way?
It must be due to lack of documentation or my ignorance to some details. But I don't see anything new except of new opportunity for a developer to go farther from real business aspects. Please, post a document, anything on how does this thing work not syntactic aspect of its features. To see a picture you have to stay back and not to examine pixels it's composed of. Thanks.