OA for n=408,600,792,856,1368,2328,...

[ncohen]

Another construction from the paper used in #16662. It is recursive as always, and the find function is, this time, very easy to write :-)

Unfortunately there is only one updated value in the MOLS with v<600. The second is .... 600 :-P

Nathann

[git]

Branch pushed to git repo; I updated commit sha1. Last 10 new commits:

​ba6609d	trac #16604: OA for n=640, reviewer's remarks and silly mistake
​c218148	trac #16604: OA(15,896)
​6074f4b	trac #16604: OA(16,208)
​35cbadc	trac #16604: OA(16,176)
​bfcc6f5	trac #16604: Now without copy and paste
​326ece4	trac #16604: OA(20,352)
​a621220	trac #16604: OA(20,416)
​54fc1e1	trac #16604: OA(20,544)
​6e47439	trac #16662: OA for n=1046,1059,2164,3992,3994
​87dd885	trac #16665: New OA for n=408,600,792,856,1368,2328,...

[git]

Branch pushed to git repo; I updated commit sha1. This was a forced push. New commits:

​0232b73	trac #16604: OA(20,544)
​e5f428d	trac #16604: Merged with 6.3.beta6
​355ac2a	trac #16662: OA for n=1046,1059,2164,3992,3994
​15b449c	trac #16665: New OA for n=408,600,792,856,1368,2328,...

[vdelecroix]

Hello Nathann,

Few modifs at u/vdelecroix/16665

You prefer K.primitive_element instead of K.multiplicative_generator? I used to use the former... (not very important)

Checking that n is a prime power and then calling FiniteField(n) is redundant (not very important).

Don't you prefer raise AssertionError(msg) instead of assert False,msg?

How do you know that the following is satisfied

    # All blocks of PG should intersect 'AG_2_3' on !=2 AG_2_3.
    assert all(len(AG_2_3.intersection(B)) != 2 for B in PG)

Vincent

[ncohen]

Yo !

Few modifs at u/vdelecroix/16665

Look good !

You prefer K.primitive_element instead of K.multiplicative_generator? I used to use the former... (not very important)

I prefer the second, because it is clearer for me. But it's not my field, if I may say.

Checking that n is a prime power and then calling FiniteField(n) is redundant (not very important).

Don't you prefer raise AssertionError(msg) instead of assert False,msg?

I do.

How do you know that the following is satisfied

    # All blocks of PG should intersect 'AG_2_3' on !=2 AG_2_3.
    assert all(len(AG_2_3.intersection(B)) != 2 for B in PG)

That's what the paper claims, and the reason why 'points' is defined as it is.

Nathann

[ncohen]

New commits:

​afd77ce

trac #16665: cosmetic

[vdelecroix]

Replying to ncohen:

Yo !

Few modifs at u/vdelecroix/16665

Look good !

You prefer K.primitive_element instead of K.multiplicative_generator? I used to use the former... (not very important)

I prefer the second, because it is clearer for me. But it's not my field, if I may sa

Sorry: I prefer multiplicative_generator and you did use primitive_element (in orthogonal_arrays.py at line 1359 and orthogonal_arrays_recursive.py at line 1175). We seem to agree on using multiplicative_generator.

Don't you prefer raise AssertionError(msg) instead of assert False,msg?

I do.

But you wrote assert False,msg at line 1177 in orthogonal_arrays_recursive.py

How do you know that the following is satisfied

    # All blocks of PG should intersect 'AG_2_3' on !=2 AG_2_3.
    assert all(len(AG_2_3.intersection(B)) != 2 for B in PG)

That's what the paper claims, and the reason why 'points' is defined as it is.

I see.

Vincent

[ncohen]

Yo !

Sorry: I prefer multiplicative_generator and you did use primitive_element (in orthogonal_arrays.py at line 1359 and orthogonal_arrays_recursive.py at line 1175). We seem to agree on using multiplicative_generator.

I don't mind. Change it if you like.

But you wrote assert False,msg at line 1177 in orthogonal_arrays_recursive.py

It is true. Does that invalidate my taste ? :-P

I think I saw Volker use it somewhere, and I had not thought of it. Or maybe this "False" was something else before, I don't know.

Nathann

[git]

Branch pushed to git repo; I updated commit sha1. New commits:

​f7cc3c1

trac #16665: co-co-cosmetic

[ncohen]

raise ValueError("q(={}) must be congruent to 0 or 1 mod 3")

You forgot a .format. This calls for a co-co-co-cosmetic commit.

Nathann

[git]

Branch pushed to git repo; I updated commit sha1. New commits:

​2eaeb09

trac #16665: coooooosmetic

[vdelecroix]

Once #16662 is positively reviewed, there will be some rebase...

Vincent

[ncohen]

There is nothing I can do about that on this patch, #16662 and its dependencies must be reviewed first.

Nathann

[ncohen]

Rebased ! And I merged all your "cosmetic" commits into one.

Nathann

---

New commits:

​053945d	trac #16662: merge 6.3
​166b525	trac #16662: doc + test
​8fcaed0	trac #16662: New assertion
​2a64fa2	trac #16662: find_thwart_lemma_3_5
​68a2d99	trac #16662: Review
​7e0480e	trac #16662: Merged with 6.4.beta0
​a9b594f	trac #16662: A comment about n<N-2
​8615065	trac #16665: New OA for n=408,600,792,856,1368,2328,...
​64b21bc	trac #16665: cosmetic

[vdelecroix]

I thought we changed the name of _helper_function_when_n_is_prime_times_power_of_2 in #16604.

Vincent

[git]

Branch pushed to git repo; I updated commit sha1. This was a forced push. Last 10 new commits:

​bff470e	trac #16604: small check of dimensions
​6c21904	trac #16797: correct a row/column inversion
​48b6902	trac #16604: merge #16797
​579e75b	trac #16604: input check + doctest
​e1a83d0	trac #16604: Optional check flag
​05c6915	trac #16604: Variable rename and list->set
​24c4f7f	trac #16604: Merge with updated #16797
​7a73e74	trac #16662: Merge with updated #16604
​a71cf90	trac #16665: New OA for n=408,600,792,856,1368,2328,...
​539c365	trac #16665: cosmetic

[vdelecroix]

Doc builds. Test pass...

Vincent