AMP started at the same time as the BBP, both were started by threads of roughly the same size, with the same level of active participation. The major difference is no warnings were added for the Foldable/Traversable proposal as at the time, generalizing types was seen as pretty simple.

e.g. foldMap comes into scope without a warning in 7.10 due to BBP, but then (<*>) also comes into scope in the AMP and we didn't warn for that, either.

I do fully agree that the BBP could have gone in a bit cleaner if we'd realized the full scope of the issues we would face during the 7.8 timeline. Most of the hiccups only came clear after Herbert had spent months refactoring the internals of base to even make the change possible.

Just an aside, though, it does appear objectively that those warnings actually did very little good in terms of getting folks to fix their code. By far the vast majority of the 7.10 related breakage we've seen has come from folks who still don't have Applicative instances for their Monad.

-Edward

On Wed, Jan 28, 2015 at 4:20 PM, Joachim Breitner <mail@joachim-breitner.de> wrote:
Hi,

Am Mittwoch, den 28.01.2015, 15:45 -0500 schrieb Edward Kmett:


> We absolutely do need to work out more effective ways to communicate
> what is going on.

I’d like to point out that the AMP proposal and process can serve as a
positive example. It is a change of comparable impact, but has sired
much less noise. My impression was that there were more people aware of
it, and sooner in the process, and that has helped.


Maybe every possibly far-reaching change should be advertised with a
comic from now on? https://ro-che.info/ccc/21 :-)

Greetings,
Joachim
>

--
Joachim “nomeata” Breitner
  mail@joachim-breitner.dehttp://www.joachim-breitner.de/
  Jabber: nomeata@joachim-breitner.de  • GPG-Key: 0xF0FBF51F
  Debian Developer: nomeata@debian.org


_______________________________________________
Libraries mailing list
Libraries@haskell.org
http://www.haskell.org/mailman/listinfo/libraries