feature-wish: quicksave process based on some hash function?
hi,
I encountered often the same "unpleasantness" that whenever I change some minor detail, e.g try other normalization or just exchange the learner, I will have to save that whole process under some different name in a different folder to save my results proper with my process...
now my question is, wouldn't it be possible to just let save the process in the same folder, but with some different append-like structure , e.g some suffix in the name, that has the hash code of e.g the xml process, based on the count or distribution of the alphabet in the xml code (of the process)? or it can also be some other hash function...
I just mean it would be some automatic way to quickly save the same process with some minor change and also save the result space with that same suffix to be able to assign them to the same process.