javier wrote:
> On Jul 31, 4:47 pm, David Joyner <[email protected]> wrote:
>> Maybe I don't understand your question. It seems you are claiming
>> thatif G is a permutation group and H is a normal subgroup then
>> the quotient G/H embeds into G. Are you sure that is true?
>
> In general, no it isn't. We have the short exact sequence
> 1 --> H --> G --> G/H -->1
> to embed G/H into G we would need it to split, which would mean that
> the extension is trivial, or that G factorizes as a product of H and
> G/H.

OK, my fault, too much wishful thinking here.

kcrisman wrote:
> 
> Look at
> 
> sage: G.quotient_group??
> 
> It turns out that Sage asks GAP to create the image of the morphism G -
>> G/H, as far as I can tell, and in so doing creates that image as a
> separate (sub)permutation group.  In particular, it using
> RegularActionHomomorphism to do this, and at
> http://www.gap-system.org/Manuals/doc/htm/ref/CHAP039.htm#SSEC007.2 it
> says "returns an isomorphism from G onto the regular permutation
> representation of G" and certainly in this case G/H (the relevant
> group) has six elements!
> 
> Though I agree that this could be confusing, the good part is that
> this creates (an isomorphic) group without having to talk about which
> element of the coset you pick each time.  It would be misleading to
> say that (1234) was an element of G/H (which I think is what David was
> getting at). There are ways to get cosets in GAP, of course (maybe
> wrapped in Sage?) but I don't know much about them.
> 
> I hope this helps!
> 

Yes, the cosets are, what I really want (although generators would be
very nice, too, if they exist, e.g. if the quotient is normal).

The application was the following: Suppose you have a linear equation
with multi-indexed variables, like x_1,2 + x_2,3 + x_3,1 == 0, which
also holds for all permutations of {1, 2, 3}, and I want to enumerate
all possible equations, but without duplicates. I hoped it was possible
to first compute the permutations under whose operation the exact same
equation result, then take the subgroup H generated by those and use
representants from the cosets of S_n/H to get all unique equations.
Looks like it's not that simple, since H doesn't even have to be normal,
in general.

Thanks a lot , though.

-- 
Robert Schwarz <[email protected]>

Get my public key at http://rschwarz.net/key.asc

--~--~---------~--~----~------------~-------~--~----~
To post to this group, send email to [email protected]
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://www.sagemath.org
-~----------~----~----~----~------~----~------~--~---

Reply via email to