Exploring alternative formulations is great, but I think it's (mostly?) orthogonal to this thread's original email: Jan found the RebindableSyntax support for Arrow to be disappointing hamstrung. I've had a similar experience in the past; the occurrences of the combinators seem to have overly restrictive type ascriptions in the desugared terms.
I don't think resolving that necessarily involves changing the Arrow class. Just the desugaring algorithm would have to change (hopefully).
And Opaleye (a successor to haskellDB, for safe interaction with SQL databases) also uses arrow notation last I checked. As I recall do-notation is too powerful, whereas proc-notation provides exactly the right expressive power (no illegal SQL queries can be expressed). But that's not to say Tom (author of Opaleye) couldn't be content with a profunctor-based desugaring rather than an Arrow-based one?_______________________________________________On 21 December 2016 at 16:31, Brandon Allbery <allbery.b@gmail.com> wrote:On Wed, Dec 21, 2016 at 12:15 AM, Edward Kmett <ekmett@gmail.com> wrote:Given that little new code seems to be being written with Arrows in mind, while some older code makes heavy use of it (hxt, etc.), refactoring the arrow hierarchy is kind of a hard sell. It is by no means impossible, just something that would require a fair bit of community wrangling and a lot of work showing clear advantages to a new status quo at a time when its very hard to get anybody to care about arrow notation at all.
The arrowized-FRP folks seem to care a fair bit.--brandon s allbery kf8nh sine nomine associatesunix, openafs, kerberos, infrastructure, xmonad http://sinenomine.net
_______________________________________________
ghc-devs mailing list
ghc-devs@haskell.org
http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs
ghc-devs mailing list
ghc-devs@haskell.org
http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs