Group cohomology now even on t2 and with full doctest coverage
Simon King
we proudly present the new version 2.1 of the modular group cohomology
spkg.
Now, all classes, methods and functions in the Cython and Python code
are covered by tests. Moreover, it builds and tests not only on
sage.math and bsd.math (64 and 32 bit mode), but also on t2. It is
able to take advantage of new features and bugfixes of Singular 3-1-1,
but also works with Singular 3-1-0. Documentation is at
http://sage.math.washington.edu/home/SimonKing/Cohomology/
New features include the computation of essential and depth essential
ideals. We found a new (in fact: only the second known) example of a
group whose modular cohomology ring has an essential ideal whose
square does not vanish.
In order to make it work on t2, it was needed to change several
function names in the underlying C-code, since apparently the Sun
linker was confusing them with names from pari. Actually the same
happened before, when a name from Singular was mistaken for a name of
pari (ticket #1396). So, I think the problem of pari infiltrating
other packages' name spaces should be studied in more detail.
The ticket for the new package version is #9894 -- and of course I'd
appreciate if you could review it. I think it would be a good idea to
test on some more platforms, such as PPC (but I have no access to
one).
Best regards,
Simon
Minh Nguyen
On Sat, Sep 11, 2010 at 9:28 AM, Simon King <simon...@nuigalway.ie> wrote:
> Dear group theorists and friends of group theorists,
>
> we proudly present the new version 2.1 of the modular group cohomology
> spkg.
Thanks, Simon. I've spread your announcement on Facebook [1] and
Twitter [2] so the package can get wider attention.
[1] http://www.facebook.com/pages/Sage-Math/26593144945
[2] http://twitter.com/sagemath
--
Regards
Minh Van Nguyen