Hello, Thank you for the comments.
I will open a JIRA issue as soon as possible. I would like to tame the code a little bit more. I haven't mentioned this but for configuration we actually used a properties file wrapped by a class. So, it's pretty compatible with anything. But Isabel is right about backward compatibility. I will check whether we can provide the same functionality by using a similar argument system to the existing algorithms. For unit testing, I was talking about strict rules. For example, using 'new's in the code instead of dependency injection may result in non-unit-testible code. I will improve the test quality before submitting the patch. But, code is already available, so any reviews are very welcome at any time. Currently, there are two different algorithms that address two different problems: One of them (Dist-Eclat) is faster if the data fits into memory, and the other one (BigFim) is slower but scales better. But you are absolutely right about adding new code. What I meant is to actively maintain and swap the algorithms with faster ones when they become available. Thank you again for your time. Cheers, -- Emin Aksehirli
signature.asc
Description: OpenPGP digital signature
