[SciPy-dev] Monte Carlo package
jonathan.taylor at utoronto.ca
Wed Jan 25 11:19:44 CST 2006
I agree... it is hard enough to write tests for scientific routines
already. Being able to make your routines deterministic is a very
desirable feature. On the other hand, I am sure it must be possible,
albeit with a little work, to make the routine accept a seed... no?
On 1/25/06, Robert Kern <robert.kern at gmail.com> wrote:
> Ed Schofield wrote:
> > Robert Kern wrote:
> >>So there's no way for the user to specify a seed? That's not acceptable for
> >>scientific use. Preferably, I should be able to pass in a
> >>numpy.random.RandomState object ...
> > If by 'scientific use' you mean cryptography, I'd agree with you. But
> > for most Monte Carlo applications (yes, scientific ones) a
> > user-definable seed is a luxury, not a necessity.
> No, it really is a necessity. I need to be able to do Monte Carlo runs
> repeatably. If nothing else, we need to be able to write unit tests that test
> the same thing every time.
> >>Actually, I won't accept an XORshift PRNG either unless if it
> >>addresses the issues discussed in this paper
> > Robert, your tone is arrogant. By saying that *you* won't accept it,
> > you are implying that you are in charge. Remember that SciPy is a
> > community project. This means we take joint decisions, and are polite
> > to each other.
> No, what I am trying to imply is that I can't accept such code for my own use
> just like I can't accept POSIX rand() or RANLIB ranf(). However, as I am also a
> Scipy developer with experience in this field, I will also vote against
> including bad XORshift PRNGs. I don't pretend that my acceptance is
> determinative, though.
> Robert Kern
> robert.kern at gmail.com
> "In the fields of hell where the grass grows high
> Are the graves of dreams allowed to die."
> -- Richard Harter
> Scipy-dev mailing list
> Scipy-dev at scipy.net
More information about the Scipy-dev