[Haskell-cafe] Monad of no `return` Proposal (MRP): Moving `return` out of `Monad`

David Feuer david.feuer at gmail.com
Tue Oct 6 18:12:46 UTC 2015

On Oct 6, 2015 7:32 AM, "Henrik Nilsson" <Henrik.Nilsson at nottingham.ac.uk>
> Executive Summary: Please let us defer further discussion
> and ultimate decision on MRP to the resurrected HaskellPrime
> committee

Many more people are on this mailing list than will be chosen for the
committee. Those who are not chosen have useful perspectives as well.

>   1. Is the Haskell Libraries list and informal voting process
>      really an appropriate, or even acceptable, way to adopt
>      such far-reaching changes to what effectively amounts to
>      Haskell itself?

As others have said, no one wants that.

> But, as has been pointed out in a
> number of postings, a lot of people with very valuable
> perspectives are also very busy, and thus likely to
> miss a short discussion period (as has happened in the
> past in relation to the Burning the Bridges proposal)

The Foldable/Traversable BBP indeed was not as well discussed as it should
have been. AMP, on the other hand, was discussed extensively and publicly
for months. I understand that some people need months of notice to prepare
to participate in a discussion. Unfortunately, I don't think those people
can always be included. Life moves too quickly for that. I do think it
might be valuable to set up a moderated, extremely low volume mailing list
for discussion of only the most important changes, with its messages
forwarded to the general list.

> The need to "field test" MRP prior to discussing
> it in HaskellPrime has been mentioned. Graham and I
> are very sceptical. In the past, at least in the
> past leading up to Haskell 2010 or so, the community
> at large was not roped in as involuntary field testers.

No, and Haskell 2010 was, by most measures, a failure. It introduced no new
language features (as far as I can recall) and only a few of the most
conservative library changes imaginable. Standard Haskell has stagnated
since 1998, 17 years ago. Haskell 2010 did not reflect the Haskell people
used in their research our practical work then, and I think people are
justified in their concern that the next standard may be similarly
disappointing. One of the major problems is the (understandable, and in
many ways productive) concentration of development effort in a single
compiler. When there is only one modern Haskell implementation that is
commonly used, it's hard to know how changes to the standard will affect
other important implementation techniques, and therefore hard to justify
any substantial changes. That was true in 2010, and it is, if anything,
more true now.

> Therefore, please let us defer further discussion and
> ultimate decision on MRP to the resurrected
> HaskellPrime committee, which is where it properly
> belongs. Otherwise, the Haskell community itself might
> be one of the things that MRP breaks.

I hope not. Haskell has gained an awful lot from the researchers, teachers,
and developers who create it and use it. I hope we can work out an
appropriate balance of inclusion, caution, and speed.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.haskell.org/pipermail/haskell-cafe/attachments/20151006/c9d09a69/attachment.html>

More information about the Haskell-Cafe mailing list