Re: [sage-devel] Re: 2 variable polynomial factorization over finite fields

2010-01-26 Thread John Cremona
2010/1/26 Bill Hart : > Indeed. I am putting in a URSS application here at Warwick for a > student to work over the summer on bivariate factoring in Sage. > Good! I still have not decided (and I know the deadline is soon). If I have a good idea I may do it, otherwise I'll leave the field for you

[sage-devel] Re: 2 variable polynomial factorization over finite fields

2010-01-26 Thread Bill Hart
Indeed. I am putting in a URSS application here at Warwick for a student to work over the summer on bivariate factoring in Sage. Bill. On Jan 13, 9:33 am, John Cremona wrote: > Thanks -- interesting discussion!  That would make a nice project for > someone to implement in Sage. > > John > > 2010

Re: [sage-devel] Re: 2 variable polynomial factorization over finite fields

2010-01-13 Thread John Cremona
Thanks -- interesting discussion! That would make a nice project for someone to implement in Sage. John 2010/1/13 William Stein : > On Tue, Jan 12, 2010 at 4:09 PM, Bill Hart > wrote: >> The algorithm QUICK FACTOR in the von zur Gather - Kaltofen paper > > PDF attached to this email. > >> look

[sage-devel] Re: 2 variable polynomial factorization over finite fields

2010-01-12 Thread Bill Hart
The algorithm QUICK FACTOR in the von zur Gather - Kaltofen paper looks very easy to implement and only returns failure if you use a probabilistic univariate factoring algorithm. You could implement that in Sage probably with a very small amount of work. I don't suggest you'll get it down to 0.1s

Re: [sage-devel] Re: 2 variable polynomial factorization over finite fields

2010-01-12 Thread John Cremona
2010/1/12 YannLC : > > > On Jan 12, 3:44 pm, John Cremona wrote: >> No, the van Hoeij / Belabas algorithms are for univariate polynomials, >> over Q (and then over number fields).  Pari does not have multivariate >> polynomial factorization > > It might not be implemented in Pari, but the algorith

[sage-devel] Re: 2 variable polynomial factorization over finite fields

2010-01-12 Thread YannLC
On Jan 12, 3:44 pm, John Cremona wrote: > No, the van Hoeij / Belabas algorithms are for univariate polynomials, > over Q (and then over number fields).  Pari does not have multivariate > polynomial factorization It might not be implemented in Pari, but the algorithm has been further extended a

Re: [sage-devel] Re: 2 variable polynomial factorization over finite fields

2010-01-12 Thread William Stein
On Tue, Jan 12, 2010 at 6:13 AM, John Cremona wrote: > thanks for both replies!  OK, so there are various fast algorithms and > Magma has implemented at least one of them.  Sage is using Singular. > I wonder why Singular has not got a fast algorithm for this? My impression is that historically th

Re: [sage-devel] Re: 2 variable polynomial factorization over finite fields

2010-01-12 Thread John Cremona
No, the van Hoeij / Belabas algorithms are for univariate polynomials, over Q (and then over number fields). Pari does not have multivariate polynomial factorization: j...@selmer%sage -gp Reading GPRC: /etc/gprc ...Done. GP/PARI CALCULATOR Version 2.3.3 (released)

[sage-devel] Re: 2 variable polynomial factorization over finite fields

2010-01-12 Thread javier
Looking at Mark van Hoeij's website, he has a (maple) implementation of his algorithm: http://www.math.fsu.edu/~hoeij/knapsack.html he also mentions "My implementation is not tuned in the best possible way. A much better way (more efficient, more robust and simpler) to tune the algorithm is give

Re: [sage-devel] Re: 2 variable polynomial factorization over finite fields

2010-01-12 Thread John Cremona
thanks for both replies! OK, so there are various fast algorithms and Magma has implemented at least one of them. Sage is using Singular. I wonder why Singular has not got a fast algorithm for this? For this application of mine, I need to do a number of one-off factorizations like this in order

[sage-devel] Re: 2 variable polynomial factorization over finite fields

2010-01-12 Thread YannLC
On Jan 12, 2:46 pm, javier wrote: > There are indeed well known (sort of) fast algorithms for > factorization of multivariable polynomials over finite > fields:http://portal.acm.org/citation.cfm?id=808748http://www.jstor.org/stable/2008063?seq=1 > > In the second paper there is a particular (pr

[sage-devel] Re: 2 variable polynomial factorization over finite fields

2010-01-12 Thread javier
There are indeed well known (sort of) fast algorithms for factorization of multivariable polynomials over finite fields: http://portal.acm.org/citation.cfm?id=808748 http://www.jstor.org/stable/2008063?seq=1 In the second paper there is a particular (probabilistic) algorithm for bivariate polynomi