Re:1 in 4 Statisticians Say They Were Asked to Commit Scient...
by raverbashing on Saturday November 03, 2018 @03:28AM
(#18369142)
Oh, scientific/research code is hard
First of all, a lot of people writing it are inexperienced in SW best practices. Then, turning formulas into code is hard, not to mention some article "handwavings" like showing pseudocode that's almost ok except for "do hard thing in this line" and that line expands into a lot of code.
I've also had some weird bugs in code like that (but nothing that would invalidate 18mo of results - btw did you know RAND_MAX is as low as 64k in different compilers, like some older versions of VS?)
(And computation times, though the cloud helps a lot with this)
First of all, a lot of people writing it are inexperienced in SW best practices. Then, turning formulas into code is hard, not to mention some article "handwavings" like showing pseudocode that's almost ok except for "do hard thing in this line" and that line expands into a lot of code.
I've also had some weird bugs in code like that (but nothing that would invalidate 18mo of results - btw did you know RAND_MAX is as low as 64k in different compilers, like some older versions of VS?)
(And computation times, though the cloud helps a lot with this)