I agree that varying approaches leading to different methods and software are of obvious interest, but so are careful scrutiny and tweakings on an existing software and at a much lower cost if the original research source code is available.

The original source code being not available, there is a high cost in doing the above, and so it's not done, all for bad reasons.