Admits his solution is bad, might be buggy. Finds a better replacement, then ships his anyway.
What is it about audio that attracts this curious level of "engineering"?
2. The module provides infrastructure for other beamformer implementations to be plugged in, my implementation is intended to be a trivial test
3. module-beamformer did not ship (the link points to a branch in my git repository)
What is it about HN that attracts haters of open source POCs? Why don't you contribute and implement an FIR filter? You probably won't, because it's easier to bash someone's work rather than do the work yourself.
My dang-compliant version of the same sentiment would go something like, "If you're going to write posts for public consumption, please consider limiting the topic to things you actually know about. You never know who might read and/or learn from what you write."