I got some impressively lengthy, detailed and mostly positive feedback in a double-blind peer review for a paper. Is it ok (or even possible), or is it against best practices (and considered taboo), to try to find out who the reviewer was - just in case you can get more feedback and share ideas related to the topic at hand?
As @SE318 explains (+1), it is unethical to try and find out who a reviewer is. However it is not unethical for a reviewer to identify themselves if the so choose (some journals specifically give this option in their review form). I can't see much problem with sending the journal a polite request for the editor to pass on a message explaining what you had in mind so that if the reviewer was interested in collaborating they could contact you.
It is not just taboo, it is unethical to try to find out who the reviewer is, and would defeat the purpose of the double-blind. Part of the idea of double-blind is that one is able to be completely honest since they know their comments cannot be tracked back to them. I strongly advise against trying to find out who the commenter is.
Theoretically it may be possible to find out who the reviewer is, but this is a problem with the system, and you should not attempt to do this.