The advantage of stamp over stmp is that it is an anytime algorithm.
Currently the anytime nature of it is not exploitable. There are 2 ways that we might fix that.
1) Make the method asynchronous, and then provide a method to stop it in the middle of its execution and return an intermediate result.
2) Add an "accuracy" parameter so that only x% of the random steps are executed before terminating.
I believe that both can be supported and would be useful, but 2) is easier. I would like to start with 2).
The advantage of stamp over stmp is that it is an anytime algorithm. Currently the anytime nature of it is not exploitable. There are 2 ways that we might fix that. 1) Make the method asynchronous, and then provide a method to stop it in the middle of its execution and return an intermediate result. 2) Add an "accuracy" parameter so that only x% of the random steps are executed before terminating.
I believe that both can be supported and would be useful, but 2) is easier. I would like to start with 2).