4675 duplicates (down from ~5200). If you had a checkbox in every one of those (many of which are false positives, probably at least a couple thousand), it would take me forever. Using ctrl/shift/click to highlight more than one is good in this instance.
Some people want autoselect. Some people (like me) like it how it is. Perhaps make it configurable so it can be either way.
What I would REALLY like to see is the ability for Similarity to 'remember' your selections - it would be nice to NOT have to scan 15000 files and compare them every time I run the program. Then I can choose to scan any new files I add to my machine and Similarity could bounce the results of scanning those files to the stored results from previous scans and wouldn't have to run the algorithms in real time, but could still tell me if any of the files in previous scans are matches for the new files.
Also, if I "ignore" a similarity match, it should remember it, but I should be able to go into a config file and "unignore" a match.
Great program, by the way!
