Whoah... Ok.. so. Basically. PubSub was first published in 2002, which was, ergh, not now. Personally I would prefer anyone which comes out with a 'use pubsub in real situations' protocol, and implemented platform elements. Basically you just invented an RSS-ircbot for MUC I think (although without MUC), but I understand your intentions - which were most possibly the intentions why pubsub was built, but with another (a protocol designer's) viewpoint. Still I think if every major client would implement a general PubSubStanzaReceiver (for pubsub or a well-defined sub-protocol like personal eventing), your work was definitely useful, even if it wouldn't be the original format and place where those things would be built - which isn't the main goal in such situations anyway I think.

So: Idea for a generic receiver isn't bad, still I think it's a client, or clientlib -side problem that needs to be solved (even if it wasn't for 5 years).

I say this because I still see pubsub a much more flexible protocol for publish-subscribe situations than anything that could be easily handled.

But the other answer could be - and there's a possibility for it unfortunately - that pubsub is a wrong path, because, for example, it makes client implementations more complex.