Re: [sage-combinat-devel] implementing covering arrays

[David Joyner]

On Mon, May 30, 2022 at 2:35 PM brett stevens <bret...@gmail.com> wrote:
>
> Myself and my M.Sc. student Aaron Dwyer are interested in adding covering arrays to sagemath in the design theory code. We have been reviewing the orthogonal array code as guidance and have some questions for sage-combinat-devel community.
>
> - We note that orthogonal arrays are not implemented as a class. Our initial thoughts were to implement covering arrays as a class. What do you all think about that?
>

This sounds okay to me, however, I'm ccing sage-devel in case some
experts aren't subscribed to this list.

> - Would it additionally be useful to re-implement orthogonal arrays as a child class of a covering arrays class and move the orthogonal array methods into that class? What would be the least disruptive way to do this?
>
> - We know that in sagemath development it is often a good idea to make small patches that are easy to review and edit. Would the right initial things to implement be the necessary class structure, documentation, self-checker (analogous to ```is_t_design``` in ```block_design```), formatted print and output to equivalent objects like a group divisible covering design?
>
> - One of our more substantive goals is to implement the perfect hash family and covering perfect hash family constructions. We are happy to hear what other people think would be important methods to implement for in a covering array class.
>
> brett stevens
> Aaron Dwyer
>
> --
> You received this message because you are subscribed to the Google Groups "sage-combinat-devel" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to sage-combinat-d...@googlegroups.com.
> To view this discussion on the web visit https://groups.google.com/d/msgid/sage-combinat-devel/8926605c-d908-4d97-bdd7-2bd9bb552040n%40googlegroups.com.

[David Joyner]

On Thu, Mar 2, 2023 at 12:11 PM brett stevens <bret...@gmail.com> wrote:

We have completed our initial implementation of covering arrays with basic methods.  It is ready for review and we are looking for reviewers.  The pull request is here:

https://github.com/sagemath/sage/pull/35008

Againi, I'm cc'ing sage-devel in case there is someone competent to review your code there. (FYI, If Nathann Cohen was the author of the OA code, AFAIK he is no longer with SageMath.)

 

brett

On Monday, June 6, 2022 at 1:44:33 PM UTC-4 brett stevens wrote:

Thanks David.  I wanted to additionally flag our second question:

> - Would it additionally be useful to re-implement orthogonal arrays as a child class of a covering arrays class and move the orthogonal array methods into that class? What would be the least disruptive way to do this?

because we know that this would involve modifying code that other people have written and invested time and effort into.  I think there are advantages to having OAs be a class but I would be happier with the change in approach being approved by some of the original implementers of the OA source code.

thanks

brett

To view this discussion on the web visit https://groups.google.com/d/msgid/sage-combinat-devel/4d18c48c-b87f-40cd-af10-8094570519a7n%40googlegroups.com.

[brettpim]

Thanks David.  Our initial patch does not change the OA code at all.  I think once we get a fully featured CA code, we can propose a transition of the OA code to a method at that time.