2 ms·
This is a standard which few kernels will ever meet. I'd say requiring a numerical proof is the same as requiring no proof at all - because it won't ever happen
by krapht 10mo ago
This is a standard which few kernels will ever meet. I'd say requiring a numerical proof is the same as requiring no proof at all - because it won't ever happen unless you're validating silicon or something equally expensive.
- Q6T46nT668w6i3m 10mo agoI guess it depends on your definition of proof but I’d say the reasoning and justifications sections of a TOMS article qualifies and that’s a standard nearly every popular library meets.