Selective: move [branch] in core

[gr-im]

As mentioned by @snowleopard (feel free to say if you don't want be ping anymore) here: snowleopard/selective#30 select can be express via branch (and vice versa). So this PR move branch into Selective.CORE and allows to use select path or branch path for Selective incarnation.

[xvw]

Thanks!

[snowleopard]

Looks fine to me, although I have to admit that I'm getting lost in all of these conversion functors. Providing a lot of ways to instantiate selective functors brings convenience but also complexity.

[snowleopard]

As mentioned by @snowleopard (feel free to say if you don't want be ping anymore) here

@gr-im I don't mind to be pinged. Thanks for continuing to explore the design space of selective functors!

[gr-im]

@gr-im I don't mind to be pinged. Thanks for continuing to explore the design space of selective functors!

Thanks a lot. To be really honest, Selective as replaced my usage of ArrowChoice, no more point free and its lead to readable solution. So thanks to have pointed (and properly elaborated in a very enjoyable paper).

Looks fine to me, although I have to admit that I'm getting lost in all of these conversion functors. Providing a lot of ways to instantiate selective functors brings convenience but also complexity.

Yes, we differ from the Janestreet approach (which use [< Derived | Defined f]) and the lack of implicits and higher kinded types lead to some complexe functors gymnastic. When Preface will be public I guess that the "happy path" will always be choose instead of building a module block by block (using Via). So it introduce a lot of boilerplate for the library maintainer, but in my opinion, it introduce less boilerplate for the user.

Thanks a lot for your reviews.

[snowleopard]

Thanks @gr-im, all sounds good.