Hey,
I've been looking into changing RR.random_element and RDF.random_element to
do something more sensible than create a random integer in the range -2 to 2
and coerce it into the reals. The reasonable thing to do for RDF seems to be
to use GSL, create a RealDistribution and then return a random element from
that. Two issues arise:
1) This doesn't work for RR, since GSL's distributions (as far as I can
tell) only support doubles. Does anyone know how to get various
distributions in mpfr? Also, where can I find documentation on the
interface to mpfr? Their website suggests there are multiple ones....
2) If you create a lot of RealDistributions quickly, their initial values
seem to agree fairly often. For example,
sage: L = [RealDistribution('gaussian', RDF(1)).get_random_element() for _
in range(50)]
sage: Ldict = {}
sage: for m in L:
...: if Ldict.has_key(m):
...: Ldict[m] = Ldict[m] + 1
...: else:
...: Ldict[m] = 1
...:
sage: Ldict
{-1.97183627646 : 1,
-1.38176904179: 4,
-1.1574992194: 2,
-0.885770378312: 2,
-0.771164851336: 2,
-0.757752144086: 1,
-0.746099959575: 4,
-0.693091109446: 1,
-0.685217028642: 5,
-0.38901172123: 1,
- 0.315558170546: 2,
-0.264326023715: 2,
-0.183537311896: 2,
-0.153138950133: 3,
0.0996438605326: 1,
0.30639951233: 1,
0.338001249663: 1,
0.393989493557: 3,
0.51919088944: 2,
0.6674486868 : 2,
0.818610064197: 1,
0.822495637678: 1,
0.996828351594: 2,
2.06664604023: 4}
Naively, this seems like the wrong behavior: they should all be distinct.
Now, maybe I just need to set the seed manually. It seems like we could set
the seed by default with some other source of randomness though...
David
--~--~---------~--~----~------------~-------~--~----~
To post to this group, send email to sage-devel@googlegroups.com
To unsubscribe from this group, send email to [EMAIL PROTECTED]
For more options, visit this group at http://groups.google.com/group/sage-devel
URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/
-~----------~----~----~----~------~----~------~--~---
