more flexible partial application - Haskell-prime - Haskell.org

newer
an example of why users should be...

more flexible partial application

older
forwarded message on the...

Sebastian Sylvan

23 Jan 2006 23 Jan '06

12:38 p.m.

Are there any subtle reasons for why something like the following couldn't be allowed?

foo x y z w = ... bar x w = foo x _ _ w

I.e. a more flexible version of partial application. This would be translated too

bar x w = \y z -> foo x y z w

I.e a function which takes the "_" parameters in the same order they were encountered in the function application. Some other languages allow this, such as Nemerle. Quite handy. /S -- Sebastian Sylvan +46(0)736-818655 UIN: 44640862

Reply

Sign in to reply online Use email software

Show replies by date

lennart＠augustsson.net

23 Jan 23 Jan

12:46 p.m.

I think this is doable, but you have to specify carefully where those lambdas are inserted. Where are they inserted in your suggestion? It has to be a syntactic transformation, because using type information would make it too complex in my opinion. In my opinion, it's a useful feature, but I'm not convinced it's useful enough to add to Haskell. -- Lennart Quoting Sebastian Sylvan :

Are there any subtle reasons for why something like the following couldn't be allowed?

...
foo x y z w = ... bar x w = foo x _ _ w

I.e. a more flexible version of partial application. This would be translated too

...
bar x w = \y z -> foo x y z w

I.e a function which takes the "_" parameters in the same order they were encountered in the function application.

Some other languages allow this, such as Nemerle. Quite handy.

/S

-- Sebastian Sylvan +46(0)736-818655 UIN: 44640862 _______________________________________________ Haskell-prime mailing list Haskell-prime@haskell.org http://haskell.org/mailman/listinfo/haskell-prime

Reply

Sign in to reply online Use email software

Manuel M T Chakravarty

24 Jan 24 Jan

9:41 p.m.

lennart@augustsson.net wrote:

In my opinion, it's a useful feature, but I'm not convinced it's useful enough to add to Haskell.

I agree. BTW, IIRC the language Hope+ had a similar feature. I think it was question marks that identified left out parameters in Hope+. Manuel

Quoting Sebastian Sylvan :

...
Are there any subtle reasons for why something like the following couldn't be allowed?

...
foo x y z w = ... bar x w = foo x _ _ w

I.e. a more flexible version of partial application. This would be translated too

...
bar x w = \y z -> foo x y z w

I.e a function which takes the "_" parameters in the same order they were encountered in the function application.

Some other languages allow this, such as Nemerle. Quite handy.

/S

-- Sebastian Sylvan +46(0)736-818655 UIN: 44640862 _______________________________________________ Haskell-prime mailing list Haskell-prime@haskell.org http://haskell.org/mailman/listinfo/haskell-prime

_______________________________________________ Haskell-prime mailing list Haskell-prime@haskell.org http://haskell.org/mailman/listinfo/haskell-prime

Reply

Sign in to reply online Use email software

Andres Loeh

23 Jan 23 Jan

12:46 p.m.

Are there any subtle reasons for why something like the following couldn't be allowed?

...
foo x y z w = ... bar x w = foo x _ _ w

I.e. a more flexible version of partial application. This would be translated too

...
bar x w = \y z -> foo x y z w

I.e a function which takes the "_" parameters in the same order they were encountered in the function application.

I just want to make sure I understand this. What would

foo x y z = \w -> ... bar x w = foo x _ _ w

mean?

bar x w = \y z -> foo x y z w

or

bar x w = (\y z -> foo x y z) w

Cheers, Andres

Reply

Sign in to reply online Use email software

Sebastian Sylvan

1:49 p.m.

On 1/23/06, Andres Loeh wrote:

...
Are there any subtle reasons for why something like the following couldn't be allowed?

...
foo x y z w = ... bar x w = foo x _ _ w

I.e. a more flexible version of partial application. This would be translated too

...
bar x w = \y z -> foo x y z w

I.e a function which takes the "_" parameters in the same order they were encountered in the function application.

I just want to make sure I understand this.

What would

...
foo x y z = \w -> ... bar x w = foo x _ _ w

mean?

...
bar x w = \y z -> foo x y z w

or

...
bar x w = (\y z -> foo x y z) w

I'm not sure. :-) The first one seems to make more sense to me, but the second does make sense as well. So maybe the expansion should occur at the application site. I.e. walk along all the parameters to the function lifting each _ to an outer lambda expression (a la the first one). I'm not sure exactly the best way for this to behave. /S -- Sebastian Sylvan +46(0)736-818655 UIN: 44640862

Reply

Sign in to reply online Use email software

lennart＠augustsson.net

2:05 p.m.

I don't think the second one makes sense. You would have to know something about foo to expand it into the second. But you might not, e.g., bar f x w = f x _ _ w -- Lennart Quoting Sebastian Sylvan :

On 1/23/06, Andres Loeh wrote:

...
...
Are there any subtle reasons for why something like the following couldn't be allowed?

...
foo x y z w = ... bar x w = foo x _ _ w

I.e. a more flexible version of partial application. This would be translated too

...
bar x w = \y z -> foo x y z w

I.e a function which takes the "_" parameters in the same order they were encountered in the function application.

I just want to make sure I understand this.

What would

...
foo x y z = \w -> ... bar x w = foo x _ _ w

mean?

...
bar x w = \y z -> foo x y z w

or

...
bar x w = (\y z -> foo x y z) w

I'm not sure. :-) The first one seems to make more sense to me, but the second does make sense as well. So maybe the expansion should occur at the application site. I.e. walk along all the parameters to the function lifting each _ to an outer lambda expression (a la the first one).

I'm not sure exactly the best way for this to behave.

/S -- Sebastian Sylvan +46(0)736-818655 UIN: 44640862 _______________________________________________ Haskell-prime mailing list Haskell-prime@haskell.org http://haskell.org/mailman/listinfo/haskell-prime

Reply

Sign in to reply online Use email software

Dinko Tenev

24 Jan 24 Jan

3:53 a.m.

I created a feature request for this while GHC was still on SF, it has propagated to trac and can be seen here: http://hackage.haskell.org/trac/ghc/ticket/315 Cheers, Dinko On 1/23/06, Sebastian Sylvan wrote:

Are there any subtle reasons for why something like the following couldn't be allowed?

...
foo x y z w = ... bar x w = foo x _ _ w

I.e. a more flexible version of partial application. This would be translated too

...
bar x w = \y z -> foo x y z w

I.e a function which takes the "_" parameters in the same order they were encountered in the function application.

Some other languages allow this, such as Nemerle. Quite handy.

/S

-- Sebastian Sylvan +46(0)736-818655 UIN: 44640862

Reply

Sign in to reply online Use email software

Jeffrey Yasskin

10:51 p.m.

I think it's a neat feature, but: Using _ seems to conflict with a Jhc extension in which "Using underscore in an expression expands to bottom with an error message giving the current file and line number." http://repetae.net/john/computer/jhc/jhc.html Scheme has a similar feature called (cut) documented at http://srfi.schemers.org/srfi-26/srfi-26.html. I haven't read the SRFI, but it would make sense to incorporate their experience in any new feature for Haskell. Jeffrey Yasskin On 1/24/06, Dinko Tenev wrote:

I created a feature request for this while GHC was still on SF, it has propagated to trac and can be seen here:

http://hackage.haskell.org/trac/ghc/ticket/315

Cheers, Dinko

On 1/23/06, Sebastian Sylvan wrote:

...
Are there any subtle reasons for why something like the following couldn't be allowed?

...
foo x y z w = ... bar x w = foo x _ _ w

I.e. a more flexible version of partial application. This would be translated too

...
bar x w = \y z -> foo x y z w

I.e a function which takes the "_" parameters in the same order they were encountered in the function application.

Some other languages allow this, such as Nemerle. Quite handy.

/S

-- Sebastian Sylvan +46(0)736-818655 UIN: 44640862

_______________________________________________ Haskell-prime mailing list Haskell-prime@haskell.org http://haskell.org/mailman/listinfo/haskell-prime

Reply

Sign in to reply online Use email software

Dinko Tenev

25 Jan 25 Jan

7:40 a.m.

On 1/25/06, Jeffrey Yasskin wrote:

I think it's a neat feature, but:

Using _ seems to conflict with a Jhc extension in which "Using underscore in an expression expands to bottom with an error message giving the current file and line number." http://repetae.net/john/computer/jhc/jhc.html

For one thing, JHC clearly goes beyond Haskell '98 here, and for another, it clearly has no point using syntax where a reserved word would do just fine.

Scheme has a similar feature called (cut) documented at http://srfi.schemers.org/srfi-26/srfi-26.html. I haven't read the SRFI, but it would make sense to incorporate their experience in any new feature for Haskell.

For what I understand from the document, their experience suggests that it is a useful feature, if anything :)

Jeffrey Yasskin

Cheers, Dinko

Reply

Sign in to reply online Use email software

Aaron Denney

2:44 p.m.

New subject: OT: Re: more flexible partial application

On 2006-01-25, Dinko Tenev wrote:

a huge base64 blob.

Which, sadly, makes it harder for me to read your messages. I don't suppose there's anyway to get gmail to only label messages as UTF-8 only if they contain non-ascii? -- Aaron Denney -><-

Reply

Sign in to reply online Use email software

7056

Age (days ago)

7058

Last active (days ago)

Download

9 comments

7 participants

tags

participants (7)

Aaron Denney
Andres Loeh
Dinko Tenev
Jeffrey Yasskin
lennart＠augustsson.net
Manuel M T Chakravarty
Sebastian Sylvan