miniblog.
← Back to all posts
Wilfred Hughes
Apr 1, 2016 at 07:21
It still seems odd to me that there are many different NaN values in floating point arithmetic. Those bits could have improved precision!