Re: [Caml-list] Custom let bindings

[Jonathan Protzenko <jonathan.protzenko@gmail.com>]

Hi folks,

I've put a raw, not pretty dump of the discussion at 
<http://pauillac.inria.fr/~protzenk/let-bang.html>. We didn't reach any 
consensus there, and we've covered the subject extensively. I'm just 
publishing it if anyone's interested.

Please refrain from adding any extra messages to this thread, unless 
you have arguments that make one proposal *clearly, objectively better 
than the others*. There's been enough back-and-forth already, so any 
arguments such as "/me likes" or "just pick any one of them" will be 
happily dismissed.

Thanks,

jonathan

On Sun 22 Jan 2012 11:23:45 AM CET, Xavier Leroy wrote:
> On 01/20/2012 01:39 PM, Mehdi Dogguy wrote:
>
>> I noticed that Alain Frisch tried to add custom let bindings (see r11894
>> and r11906) but it was reverted later on (see r11960) because no
>> consensus was reached (among OCaml Core team, I guess).
>
> Yes.  We Caml developers spent a lot of time arguing about the merits and
> demerits of various syntaxes for monadic lets and other forms of
> generalized let bindings, discussing usability, generality, etc.
> There were some unexpected surprises, e.g. Lwt's concurrency monad
> needing a special monadic "let ... and ... in ..."  binding, which
> makes sense for this particular monad but not for others.
>
> In the end, we decided that none of the proposals is something we can
> commit on and put (forever) in the core language.  This is one of the
> cases that is currently best handled by Camlp4 syntax extensions,
> either specific to a monad (Lwt) or more general (pa_monad).
>
> Regards,
>
> - Xavier Leroy
>