Re: [sage-devel] Re: bug with FinitelyGeneratedMatrixGroup (worked with previous version)

2014-04-21 Thread Nicolas M. Thiery
> Volker wrote: > But having an iterator would be nice even for infinite > finitely-generated groups. +1 On Mon, Apr 21, 2014 at 12:10:13PM +, Dima Pasechnik wrote: > it does not look obvious how you can ensure that the stream will not > have repeated elements (without keeping the previous s

[sage-devel] Re: bug with FinitelyGeneratedMatrixGroup (worked with previous version)

2014-04-21 Thread Volker Braun
I don't think you can do that in general, this sounds like it would be awfully close to deciding whether a fp group is finite or not. So if the group is really infinite then iterating all the way needs infinite memory in addition to infinite time. I'm fine with that. Sometimes you just need a c

[sage-devel] Re: bug with FinitelyGeneratedMatrixGroup (worked with previous version)

2014-04-21 Thread Dima Pasechnik
On 2014-04-21, Volker Braun wrote: > On Monday, April 21, 2014 10:37:11 AM UTC+1, Nicolas M. Thiéry wrote: >> >> We should probably add a "category" optional argument to MatrixGroup, >> so that we can specify this when we build the group. > > > But having an iterator would be nice even for infin

Re: [sage-devel] Re: bug with FinitelyGeneratedMatrixGroup (worked with previous version)

2014-04-21 Thread Volker Braun
On Monday, April 21, 2014 10:37:11 AM UTC+1, Nicolas M. Thiéry wrote: > > We should probably add a "category" optional argument to MatrixGroup, > so that we can specify this when we build the group. But having an iterator would be nice even for infinite finitely-generated groups. -- You rec

Re: [sage-devel] Re: bug with FinitelyGeneratedMatrixGroup (worked with previous version)

2014-04-21 Thread Nicolas M. Thiery
On Fri, Apr 18, 2014 at 09:17:16AM -0700, Volker Braun wrote: >The problem is that we don't have an algorithm to find out that the group >is finite (short of generating all elements). So it is placed in the >"Groups" category instead of "FiniteGroups", and that doesn't have an >iter

[sage-devel] Re: bug with FinitelyGeneratedMatrixGroup (worked with previous version)

2014-04-18 Thread Ben Cote
Volker, Thanks for the workaround. What gets me is that it works in an older version of sage. On Friday, April 18, 2014 9:17:16 AM UTC-7, Volker Braun wrote: > > The problem is that we don't have an algorithm to find out that the group > is finite (short of generating all elements). So it is

[sage-devel] Re: bug with FinitelyGeneratedMatrixGroup (worked with previous version)

2014-04-18 Thread Volker Braun
The problem is that we don't have an algorithm to find out that the group is finite (short of generating all elements). So it is placed in the "Groups" category instead of "FiniteGroups", and that doesn't have an iterator. There should be an in-between category of finitely generated groups whic