If someone made a version while avoiding the specific implementation would they be alright? I think that where the author messed up is by reproducing the algori

If someone made a version while avoiding the specific implementation would they be alright? I think that where the author messed up is by reproducing the algorithm altogether. The original article highlighted deficiencies in the implementation. Maybe it would have been better to fix those issues with a better, unrelated algorithm and release that for free.