From: Arnaud Spiwack <arnaud.spiwack@gmail.com>
To: caml-list@yquem.inria.fr
Subject: Re: [Caml-list] Re: Void type?
Date: Sun, 29 Jul 2007 14:58:23 +0200 [thread overview]
Message-ID: <46AC8EEF.1040102@gmail.com> (raw)
In-Reply-To: <20070729124340.GA18564@furbychan.cocan.org>
Here is what you can do with void1 and not with void2 :
type void1 = { v: 'a. 'a };;
# let void1_elim x = x.v;;
val void1_elim : void1 -> 'a = <fun>
Of course there are various ways to get the same for void2 :
# let void2_elim1 (x:void2) = assert false;;
val void2_elim1 : void2 -> 'a = <fun>
# let rec void2_elim2 (x:void2) = void2_elim2 x;;
val void2_elim2 : void2 -> 'a = <fun>
# let void2_elim3 (x:void2) = Obj.magic x;;
val void2_elim3 : void2 -> 'a = <fun>
But none of these are really functional/well typed except maybe
void2_elim2 but it looks like cheating. Of course these are all safe
because void has no closed instance. But I'd consider void1 to be
morally much more satisfying. It really contains the essence of what a
void type should be.
Arnaud Spiwack
Richard Jones a écrit :
> On Sun, Jul 29, 2007 at 01:36:24PM +0200, Arnaud Spiwack wrote:
>
>> Jon Harrop a écrit :
>>
>>> On Sunday 29 July 2007 12:05:47 Arnaud Spiwack wrote:
>>>
>>>
>>>> It is the good solution if you work with the original syntax (and it's
>>>> absolutely equivalent to the dual definition in term of empty variant
>>>> which you can write in the revised syntax).
>>>>
>>>>
>>> I don't quite understand this "empty variant from the revised syntax
>>> thing". How is:
>>>
>>> type void
>>>
>>> not an empty variant?
>>>
>>>
>>>
>> Well, not technically I believe. It's a type with no definition. I
>> wouldn't be adamant about that but I reckon it's not considered as a sum
>> type by OCaml type system.
>> Plus you cannot write the empty matching :
>> match x with []
>> in the original syntax, preventing you from writing a function of type
>> void -> 'a without using exceptions or Obj.magic or an obviously
>> looping function or such.
>>
>> Thus it does not really have the logical behavior of an empty variant.
>>
>
> Can you explain what you mean a bit more?
>
> type void1 = { v: 'a. 'a };;
> let f (_ : void1) = 1;;
> --> val f : void1 -> int = <fun>
> let f () : void1 = failwith "error";;
> --> val f : unit -> void1 = <fun>
>
> type void2;;
> let f (_ : void2) = 1;;
> --> val f : void2 -> int = <fun>
> let f () : void2 = failwith "error";;
> --> val f : unit -> void2 = <fun>
>
> They seem to be fairly similar to me.
>
> Rich.
>
>
next prev parent reply other threads:[~2007-07-29 12:58 UTC|newest]
Thread overview: 38+ messages / expand[flat|nested] mbox.gz Atom feed top
2007-07-28 4:14 Stefan Monnier
2007-07-28 4:33 ` [Caml-list] " Erik de Castro Lopo
2007-07-28 4:51 ` Chris King
2007-07-28 18:49 ` Stefan Monnier
2007-07-28 18:53 ` Basile STARYNKEVITCH
2007-07-29 0:48 ` Stefan Monnier
2007-07-28 18:57 ` Arnaud Spiwack
2007-07-28 6:12 ` Daniel de Rauglaudre
2007-07-28 6:15 ` Chung-chieh Shan
2007-07-28 8:22 ` [Caml-list] " rossberg
2007-07-29 6:31 ` Chung-chieh Shan
2007-07-29 11:05 ` [Caml-list] " Arnaud Spiwack
2007-07-29 11:16 ` Jon Harrop
2007-07-29 11:36 ` Arnaud Spiwack
2007-07-29 12:43 ` Richard Jones
2007-07-29 12:58 ` Arnaud Spiwack [this message]
2007-07-29 17:02 ` Richard Jones
2007-07-29 20:06 ` Arnaud Spiwack
2007-07-29 22:55 ` Brian Hurt
2007-07-30 4:40 ` skaller
2007-07-30 23:13 ` Brian Hurt
2007-07-31 8:52 ` Richard Jones
2007-07-31 13:08 ` Chris King
2007-07-31 15:27 ` Markus Mottl
2007-08-01 11:37 ` Tom
2007-08-01 16:23 ` Markus Mottl
2007-07-30 4:44 ` Geoffrey Alan Washburn
2007-07-30 13:11 ` [Caml-list] " Brian Hurt
2007-07-30 13:32 ` Christopher L Conway
2007-07-30 13:35 ` Geoffrey Alan Washburn
2007-07-30 13:41 ` [Caml-list] " Chris King
2007-07-30 17:43 ` Christophe Raffalli
2007-07-30 17:58 ` Markus Mottl
2007-07-30 14:27 ` Jeff Polakow
2007-07-28 7:58 ` Sébastien Hinderer
2007-07-28 8:13 ` [Caml-list] " Basile STARYNKEVITCH
2007-07-28 12:29 ` Christophe TROESTLER
2007-07-28 13:36 ` Brian Hurt
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=46AC8EEF.1040102@gmail.com \
--to=arnaud.spiwack@gmail.com \
--cc=caml-list@yquem.inria.fr \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox