[Haskell-cafe] GHC Extension Proposal: ArgumentBlock

Sven Panne svenpanne at gmail.com
Mon Sep 7 07:53:44 UTC 2015

2015-09-07 7:39 GMT+02:00 Bardur Arantsson <spam at scientician.net>:

> (Please don't top-post... and please don't quote all of the previous
> message unless you really need to  It breaks the flow of conversation.)

That's the bane of almost all new mail clients, having the totally wrong
defaults, promoting threads of quadratic size, all in Yoda-style... :-P

> Haskell keeps evolving at a brisk pace, see the depenently typed stuff
> that's coming up, etc.

Exactly, and let's not forget all the library-related changes (AMP,
FTP/BBP, MonadFail, SemiGroup/Monoid, etc.): As much as the concrete
syntax, the standard libraries constitute the basis of a common language.
Given the already high velocity of changes in the libraries, let's no make
things more complicated by introducing tons of ad hoc changes for no real
value. Don't get me wrong, the library changes are really valuable and
urgently needed, but they already put some non-trivial burden onto
maintainers. Nevertheless, I'd like to see more cleanup/fixes in that area
(e.g. an overhaul of Num and friends), but this would be item 0 in
https://wiki.haskell.org/Wadler's_Law... ;-)

> I question the value of constantly changing the *syntax*, which is
> mostly a triviality. The other syntactic extensions that we brought up
> saved a lot more than one character, btw!

IMHO it's not so much about saving characters, but more about generality
and regularity: One can see e.g. LambdaCase as unification of, well, lambda
and case, making them effectively just syntactic sugar. So by this
reasoning, LambdaCase is a "good" extension. DoAndIfThenElse has a
different quality: It just removes a very common pitfall, so personally I
consider it "less good" than LambdaCase, but it still has some value. The
proposed extension neither generalizes things nor does it remove a common
pitfall => -1.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.haskell.org/pipermail/haskell-cafe/attachments/20150907/5eeec41a/attachment.html>

More information about the Haskell-Cafe mailing list