I'd support the map=fmap thing, On Fri, Jun 21, 2013 at 3:00 PM, Daniel Díaz Casanueva < dhelta.diaz@gmail.com> wrote:

Hello!

To be honest, I always write "fmap" instead of "map" in all my code. Sometimes I change the container I am using, and if I write "map", then I would have to change that "map" to "Something.map" or just "fmap". I know for sure that fmap is going to work in any Functor instance, so I don't have to deal with these things. In some situations (using functions from other classes as well), I only change the type signature of my functions, and the program keeps working as expected, which is awesome.

Therefore, I support the proposal.

By the way, how strong is the reason "it would be harder to learn"? Is it really that hard to use type classes? I don't think you need any deep math to understand how to work with type classes in Haskell. Even if they are called "Functor" or "Monad", you don't need to know about categories to use them. In fact, any Haskell student faces type classes in some point.

In the other hand, I remember that at the beginning somebody showed me that "+" sums integers. And eventually, it worked with floats too. He called this "overloading" and, presented in that way, type classes appeared like something natural in my mind. Can't the same thing been done in this case?

Best regards, Daniel Díaz.

On Fri, Jun 21, 2013 at 11:59 AM, Andrew Butterfield < Andrew.Butterfield@scss.tcd.ie> wrote:

...

On 21 Jun 2013, at 10:51, Nikita Volkov wrote:

...

Since there's been so many suggestions concerning radically changing the Prelude recently, I'd like to start a poll on a subject.

I know that "map" is limited to lists for the beginners, but then we're already making a lot of unbeginnerish changes to Prelude. Also it's not a monad transformer, but just a functor - how can a person learn Haskell without understanding what a Functor is?

!!!! No one needs to know what a functor is in order to learn Haskell. No one needs to know any category theory ideas to learn Haskell.

Haskell is a very good vehicle to introduce such category-theoretic concepts, and as someone learns Haskell, learning about Haskell classes(*) such as Functor, etc, is very important and useful.

My main reason for this sudden and abrupt response is that I would be worried if any Prelude re-design was predicated on the idea that Haskell beginners need to understand Functors - they don't, but hopefully they will learn...

(*) as distinct from learning about the category theoretic notion of functor.

...

_______________________________________________ Libraries mailing list Libraries@haskell.org http://www.haskell.org/mailman/listinfo/libraries

-------------------------------------------------------------------- Andrew Butterfield Tel: +353-1-896-2517 Fax: +353-1-677-2204 Lero@TCD, Head of Foundations & Methods Research Group Director of Teaching and Learning - Undergraduate, School of Computer Science and Statistics, Room G.39, O'Reilly Institute, Trinity College, University of Dublin http://www.scss.tcd.ie/Andrew.Butterfield/ --------------------------------------------------------------------

_______________________________________________ Libraries mailing list Libraries@haskell.org http://www.haskell.org/mailman/listinfo/libraries

_______________________________________________ Libraries mailing list Libraries@haskell.org http://www.haskell.org/mailman/listinfo/libraries

On Fri, 21 Jun 2013, Matthias Kilian wrote:

On Fri, Jun 21, 2013 at 04:53:49PM -0400, Carter Schonwald wrote:

...

I'd support the map=fmap thing,

Does it cause any consequences for existing libraries and programs? Does it break any existing Haskell98 (or Haskell2010) code?

If not: fine. But if yes: please don't do it.

My main concern is that we reduce type-safety. If you have a lot of nested Functors, and you apply a nested fmap like in (fmap (fmap (fmap f))) with a polymorphic f, then after adding or removing a functor level the code can probably still be compiled, but is wrong. It is also more difficult to read, since in order to understand (fmap (fmap (fmap f))) you have to infer types mentally, in order to find out that, say, the second fmap is a List.map. These are my general concern with always using the most general function available. Thus a -1 for this proposal, from me.

On Sun, 23 Jun 2013, Nikita Volkov wrote:

Henning, don't forget that noone suggests to remove Data.List.map, which you will still be perfectly able to use qualified when you feel the need to.

If in future Prelude moves more and more to one most general function that does everything given the right type, then I am really concerning to use NoImplicitPrelude in all my modules and use a custom Prelude or none at all.

To me having the most general functions imported unqualified from Prelude

... are you sure, "fmap" is the most general function?

and specific ones qualified seems most natural.

On Sun, Jun 23, 2013 at 2:20 PM, Nikita Volkov wrote:

...

If in future Prelude moves more and more to one most general function

On Jun 23, 2013, at 10:11 PM, Henning Thielemann < lemming@henning-thielemann.de> wrote: that does everything given the right type, then I am really concerning to use NoImplicitPrelude in all my modules and use a custom Prelude or none at all.

That's exactly what I am doing now and the very existence of a dozen of "prelude" projects on hackage suggests that I'm hardly alone in that.

Except that most of them are making things *more* general, whereas Henning believes polymorphism is evil and confusing. Is there anything else we need to remove from Haskell to make it "appropriate"? Within the past 24 hours I have heard impassioned arguments for jettisoning polymorphism, monads, and arrows. Maybe strong typing should be removed next? Currying? -- brandon s allbery kf8nh sine nomine associates allbery.b@gmail.com ballbery@sinenomine.net unix, openafs, kerberos, infrastructure, xmonad http://sinenomine.net

Frankly, the proposal text does not make clear what exactly you want to do. I would guess the name 'fmap' was chosen for 'Functor' because 'map' was already taken for lists. So it would make sense to 1. remove Prelude.map 2. make map an alias for fmap Then, in 50 years from now we can deprecate fmap and remove it in 100 years from now. ;-) Cheers, Andreas On 21.06.2013 21:00, Daniel Díaz Casanueva wrote:

Hello!

To be honest, I always write "fmap" instead of "map" in all my code. Sometimes I change the container I am using, and if I write "map", then I would have to change that "map" to "Something.map" or just "fmap". I know for sure that fmap is going to work in any Functor instance, so I don't have to deal with these things. In some situations (using functions from other classes as well), I only change the type signature of my functions, and the program keeps working as expected, which is awesome.

Therefore, I support the proposal.

By the way, how strong is the reason "it would be harder to learn"? Is it really that hard to use type classes? I don't think you need any deep math to understand how to work with type classes in Haskell. Even if they are called "Functor" or "Monad", you don't need to know about categories to use them. In fact, any Haskell student faces type classes in some point.

In the other hand, I remember that at the beginning somebody showed me that "+" sums integers. And eventually, it worked with floats too. He called this "overloading" and, presented in that way, type classes appeared like something natural in my mind. Can't the same thing been done in this case?

Best regards, Daniel Díaz.

On Fri, Jun 21, 2013 at 11:59 AM, Andrew Butterfield mailto:Andrew.Butterfield@scss.tcd.ie> wrote:

On 21 Jun 2013, at 10:51, Nikita Volkov wrote:

> Since there's been so many suggestions concerning radically changing the Prelude recently, I'd like to start a poll on a subject. > > I know that "map" is limited to lists for the beginners, but then we're already making a lot of unbeginnerish changes to Prelude. Also it's not a monad transformer, but just a functor - how can a person learn Haskell without understanding what a Functor is?

!!!! No one needs to know what a functor is in order to learn Haskell. No one needs to know any category theory ideas to learn Haskell.

Haskell is a very good vehicle to introduce such category-theoretic concepts, and as someone learns Haskell, learning about Haskell classes(*) such as Functor, etc, is very important and useful.

My main reason for this sudden and abrupt response is that I would be worried if any Prelude re-design was predicated on the idea that Haskell beginners need to understand Functors - they don't, but hopefully they will learn...

(*) as distinct from learning about the category theoretic notion of functor.

> > > > _______________________________________________ > Libraries mailing list > Libraries@haskell.org mailto:Libraries@haskell.org > http://www.haskell.org/mailman/listinfo/libraries

-------------------------------------------------------------------- Andrew Butterfield Tel: +353-1-896-2517 Fax: +353-1-677-2204 Lero@TCD, Head of Foundations & Methods Research Group Director of Teaching and Learning - Undergraduate, School of Computer Science and Statistics, Room G.39, O'Reilly Institute, Trinity College, University of Dublin http://www.scss.tcd.ie/Andrew.Butterfield/ --------------------------------------------------------------------

_______________________________________________ Libraries mailing list Libraries@haskell.org mailto:Libraries@haskell.org http://www.haskell.org/mailman/listinfo/libraries

_______________________________________________ Libraries mailing list Libraries@haskell.org http://www.haskell.org/mailman/listinfo/libraries

-- Andreas Abel <>< Du bist der geliebte Mensch. Theoretical Computer Science, University of Munich Oettingenstr. 67, D-80538 Munich, GERMANY andreas.abel@ifi.lmu.de http://www2.tcs.ifi.lmu.de/~abel/

On Fri, Jun 21, 2013 at 4:57 PM, Andreas Abel wrote:

Frankly, the proposal text does not make clear what exactly you want to do.

I would guess the name 'fmap' was chosen for 'Functor' because 'map' was already taken for lists.

About 15 years ago (in Haskell 1.4), 'map' was the name of the method of the Functor class. Then it was changed to 'fmap' and the monomorphic 'map' was added. (Hint: go read the Haskell 1.4 report if you want to see a version of Haskell that is better than the one we have today in a dozen or so ways. :))

So it would make sense to

1. remove Prelude.map 2. make map an alias for fmap

Then, in 50 years from now we can deprecate fmap and remove it in 100 years from now. ;-)

I suspect that is the proposal, although it probably entails getting rid of Data.List.map (and re-exporting the Functor version) as well to not be a significant pain. It is possible to break some things by doing this, but the corner cases involved ought to be pretty rare. -- Dan