On 4 Jan 2014 20:11, "David Joyner" <[email protected]> wrote:
>
>
>
> On Tuesday, December 24, 2013 6:22:27 AM UTC-5, Nathann Cohen wrote:
>>
>> Hellooooooo everybody !!!
>>
>> I send this email because I want to add new stuff to our
combinat/designs/ folder, and I really need some help to review patches.
These are constructions of combinatorial designs, i.e. SOoooooo very sexy
set systems.
>>
>
> I'm starting to get interested research-wise in this stuff.
> When trac was reasonably fast, I looked them over and the
formatting/documentation looked fine.
> It seemed like Dima has looked at them as well?

I did, briefly. I will have some time for this in the 2nd half of January,
after we are back to Oxford.

> Now trac is so slow, and since the review system has changed from hg to
git,
> I'd have to look up how to apply patches and test.
>
> I have code code of my own testing+constructing weighted PDSs
(essentially association schemes).
> It would be cool if Sage had Schur rings associated to association
schemes. Every time
> I have to deal with them, I have to construct them by hand...
>
>
>>
>> I don't think such code is available anywhere else. Those constructions
exist on paper but I don't think anybody implemented them, and it's a pity
for two reasons :
>>
>> 1) The constructions can be pretty tricky, but checking that they work
on a computer is very easy (and my code does it always, so no wrong result
can ever be returned, no fear of that)
>
>
> Nice idea, Nathann!
>
>>
>> 2) Sometimes it's even hard to just decipher the constructions from old
papers, and implementing [1] felt like saving them before it became
impossible to read it :-P
>>
>> What I try to implement these days are Balanced Incomplete Block Designs
[2], which you can see as an edge-decomposition of the complete graph K_v
into copies of K_k.
>>
>> - K_n can be decomposed into K_3 if and only if v = 1,3 [6] (
implemented there [3], examples on [4] )
>> - K_n can be decomposed into K_4 if and only if v = 1,4 [12] (several
patches waiting for a review)
>> - K_n can be decomposed into K_5 if and only if v = 1,5 [20] (on my hard
drive only, which can become a patch anytime)
>>
>> And then again I don't think that the decompositions are implemented
anywhere else for K_4 and K_5.
>>
>> I need help to review these patches. I understand the constructions are
very tricky, but then again *THEY ARE ALL CHECKED BEFORE SAGE RETURNS THEM*
so even if the whole construction code is gibberish there is nothing to
fear for as long as the code that checks it is correct. And this code is
dead simple.
>>
>> Sooooooo well... I need help for that, definitely. I tried to make it
*very well* documented, and if it isn't sufficient just ak me for more and
I will add them quick.
>>
>> (a shortcut, and a bugfix)
>> #15107 Projective Plane designs
>> #15285 Bug in AffineGeometryDesign
>>
>> (construction for k=4)
>> #15286 Latin squares
>> #15287 Orthogonal Arrays
>> #15288 Balanced Incomplete Block Designs with k=4
>>
>> (needed for k=5)
>> #15310 Wilson's construction of Transversal Designs/Orthogonal
Arrays/MOLS
>> #15431 Transversal Design TD(6,12)
>>
>> Sooo well... If you can help me with that, and of course if I can help
you with some of your patches tell me about it.
>>
>> Thaaaaaaaaaaaaaaaaaaaaaaaaank you very much !!!
>>
>> Nathann
>> http://www.steinertriples.fr/ncohen/
>>
>> P.S. : please don't only answer on sage-combinat
>>
>> [1]
http://www.sagemath.org/doc/reference/combinat/sage/combinat/designs/steiner_quadruple_systems.html
>> [2]
http://en.wikipedia.org/wiki/Block_design#Definition_of_a_BIBD_.28or_2-design.29
>> [3]
http://www.sagemath.org/doc/reference/combinat/sage/combinat/designs/block_design.html#sage.combinat.designs.block_design.steiner_triple_system
>> [4] http://www.steinertriples.fr/

-- 
You received this message because you are subscribed to the Google Groups 
"sage-devel" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To post to this group, send email to [email protected].
Visit this group at http://groups.google.com/group/sage-devel.
For more options, visit https://groups.google.com/groups/opt_out.

Reply via email to