I've been tempted to write CW decoder algorithm for a some time. I've tried many available decoders but they all seem to work much or less pretty badly.
So I though "how come it can be so difficult?" - and by experience I know this kind of thoughts are dangerous. Like I am going to solve this out of all people?
So I did write one for iOS. And it seems to work decently well even on noisy signals.
So should I leave it as proof-of-concept? Job done? But it's raining and propagation is bad ..