Does this typeclass have a name? - Haskell-Cafe - Haskell.org

newer
Re: [Haskell-cafe] "add ifM and...

Does this typeclass have a name?

older
Re: [Haskell-cafe] conduit zips

Bardur Arantsson

21 Apr 2014 21 Apr '14

5:09 a.m.

Hi all, I was wondering if anyone out there knows if this type class has a name? class Foo a e where foo :: e -> a -> a bar :: a (I also have a fundep a -> e, but that's not essential.) Essentially the usage is: You have a sequence of "events" e_i and a starting value "bar" and can use the "foo" function to apply all events to that starting value foo e_n $ foo e_{n-1} $ ... $ foo e_0 $ bar and thus get the final value of the a induced by the events e_i. (I've included the "bar" starting value in the type class, but I suppose it could be supplied by a Default instance.) Regards,

Reply

Sign in to reply online Use email software

Show replies by date

Ivan Lazar Miljenovic

21 Apr 21 Apr

7:04 a.m.

On 21 April 2014 15:09, Bardur Arantsson wrote:

Hi all,

I was wondering if anyone out there knows if this type class has a name?

class Foo a e where foo :: e -> a -> a bar :: a

(I also have a fundep a -> e, but that's not essential.)

Essentially the usage is:

You have a sequence of "events" e_i and a starting value "bar" and can use the "foo" function to apply all events to that starting value

foo e_n $ foo e_{n-1} $ ... $ foo e_0 $ bar

= foldr foo bar [e_n, e_{n-1} .. e_0] ?

and thus get the final value of the a induced by the events e_i.

(I've included the "bar" starting value in the type class, but I suppose it could be supplied by a Default instance.)

Regards,

_______________________________________________ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe

-- Ivan Lazar Miljenovic Ivan.Miljenovic@gmail.com http://IvanMiljenovic.wordpress.com

Reply

Sign in to reply online Use email software

Bardur Arantsson

8:08 a.m.

On 2014-04-21 09:04, Ivan Lazar Miljenovic wrote:

On 21 April 2014 15:09, Bardur Arantsson wrote:

...
Hi all,

I was wondering if anyone out there knows if this type class has a name?

class Foo a e where foo :: e -> a -> a bar :: a

(I also have a fundep a -> e, but that's not essential.)

Essentially the usage is:

You have a sequence of "events" e_i and a starting value "bar" and can use the "foo" function to apply all events to that starting value

foo e_n $ foo e_{n-1} $ ... $ foo e_0 $ bar

= foldr foo bar [e_n, e_{n-1} .. e_0] ?

Yes, indeed, but I want the type class to constrain things since there are other interacting constraints on the types "a" and "e" (plus there's the fundep that I mentioned). The "single instance" part of type classes is also very desirable. Regards,

Reply

Sign in to reply online Use email software

Ivan Lazar Miljenovic

9:25 a.m.

On 21 April 2014 18:08, Bardur Arantsson wrote:

On 2014-04-21 09:04, Ivan Lazar Miljenovic wrote:

...
On 21 April 2014 15:09, Bardur Arantsson wrote:

...
Hi all,

I was wondering if anyone out there knows if this type class has a name?

class Foo a e where foo :: e -> a -> a bar :: a

(I also have a fundep a -> e, but that's not essential.)

Essentially the usage is:

You have a sequence of "events" e_i and a starting value "bar" and can use the "foo" function to apply all events to that starting value

foo e_n $ foo e_{n-1} $ ... $ foo e_0 $ bar

= foldr foo bar [e_n, e_{n-1} .. e_0] ?

Yes, indeed, but I want the type class to constrain things since there are other interacting constraints on the types "a" and "e" (plus there's the fundep that I mentioned). The "single instance" part of type classes is also very desirable.

Typeclasses only really make sense if you're defining polymorphic functions that deal with a variety of these kinds of types. If that's the case, define your own typeclass (as it doesn't appear generic enough to be found in existing libraries) and define the instances you need. Otherwise, maybe just define an explicit dictionary record with values for every possible type pairing (and use an associated type instance if you want to explicitly state that there's a relation between a and e)? -- Ivan Lazar Miljenovic Ivan.Miljenovic@gmail.com http://IvanMiljenovic.wordpress.com

Reply

Sign in to reply online Use email software

Bardur Arantsson

9:28 a.m.

On 2014-04-21 11:25, Ivan Lazar Miljenovic wrote:

On 21 April 2014 18:08, Bardur Arantsson wrote:

...
On 2014-04-21 09:04, Ivan Lazar Miljenovic wrote:

...
On 21 April 2014 15:09, Bardur Arantsson wrote:

...
Hi all,

I was wondering if anyone out there knows if this type class has a name?

class Foo a e where foo :: e -> a -> a bar :: a

(I also have a fundep a -> e, but that's not essential.)

Essentially the usage is:

You have a sequence of "events" e_i and a starting value "bar" and can use the "foo" function to apply all events to that starting value

foo e_n $ foo e_{n-1} $ ... $ foo e_0 $ bar

= foldr foo bar [e_n, e_{n-1} .. e_0] ?

Yes, indeed, but I want the type class to constrain things since there are other interacting constraints on the types "a" and "e" (plus there's the fundep that I mentioned). The "single instance" part of type classes is also very desirable.

Typeclasses only really make sense if you're defining polymorphic functions that deal with a variety of these kinds of types.

If that's the case, define your own typeclass (as it doesn't appear generic enough to be found in existing libraries) and define the instances you need.

Of course, I'm already doing exactly that -- I really was just wondering if there might be a name for the type class :). Regards,

Reply

Sign in to reply online Use email software

Tom Ellis

8 a.m.

On Mon, Apr 21, 2014 at 07:09:13AM +0200, Bardur Arantsson wrote:

class Foo a e where foo :: e -> a -> a bar :: a [...] You have a sequence of "events" e_i and a starting value "bar" and can use the "foo" function to apply all events to that starting value

Do you *really* want a typeclass for this, rather than a value? Tom

Reply

Sign in to reply online Use email software

Tobias Brandt

9:46 a.m.

Your class looks like Unfoldable from Data.Collectionshttp://hackage.haskell.org/package/collections-api-1.0.0.0/docs/Data-Collect.... foo is insert and bar is empty. The other three methods can be defined in terms of insert and empty. On 21 April 2014 07:09, Bardur Arantsson wrote:

Hi all,

I was wondering if anyone out there knows if this type class has a name?

class Foo a e where foo :: e -> a -> a bar :: a

(I also have a fundep a -> e, but that's not essential.)

Essentially the usage is:

You have a sequence of "events" e_i and a starting value "bar" and can use the "foo" function to apply all events to that starting value

foo e_n $ foo e_{n-1} $ ... $ foo e_0 $ bar

and thus get the final value of the a induced by the events e_i.

(I've included the "bar" starting value in the type class, but I suppose it could be supplied by a Default instance.)

Regards,

_______________________________________________ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe

Reply

Sign in to reply online Use email software

Andras Slemmer

1:46 p.m.

I think what you're looking for are monoid actions: http://hackage.haskell.org/package/monoid-extras-0.2.2.0/docs/Data-Monoid-Ac... And the laws the typeclass should satisfy: foo bar = id foo f . foo g = foo (f `mappend` g) The identity law is optional, in whihc case your 'e' would be a semigroup without a default 'bar' On 21 April 2014 11:46, Tobias Brandt wrote:

Your class looks like Unfoldable from Data.Collectionshttp://hackage.haskell.org/package/collections-api-1.0.0.0/docs/Data-Collect.... foo is insert and bar is empty. The other three methods can be defined in terms of insert and empty.

On 21 April 2014 07:09, Bardur Arantsson wrote:

...
Hi all,

I was wondering if anyone out there knows if this type class has a name?

class Foo a e where foo :: e -> a -> a bar :: a

(I also have a fundep a -> e, but that's not essential.)

Essentially the usage is:

You have a sequence of "events" e_i and a starting value "bar" and can use the "foo" function to apply all events to that starting value

foo e_n $ foo e_{n-1} $ ... $ foo e_0 $ bar

and thus get the final value of the a induced by the events e_i.

(I've included the "bar" starting value in the type class, but I suppose it could be supplied by a Default instance.)

Regards,

_______________________________________________ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe

_______________________________________________ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe

Reply

Sign in to reply online Use email software

Andras Slemmer

1:49 p.m.

Oh, and bar = mempty On 21 April 2014 15:46, Andras Slemmer <0slemi0@gmail.com> wrote:

I think what you're looking for are monoid actions: http://hackage.haskell.org/package/monoid-extras-0.2.2.0/docs/Data-Monoid-Ac... And the laws the typeclass should satisfy: foo bar = id foo f . foo g = foo (f `mappend` g)

The identity law is optional, in whihc case your 'e' would be a semigroup without a default 'bar'

On 21 April 2014 11:46, Tobias Brandt wrote:

...
Your class looks like Unfoldable from Data.Collectionshttp://hackage.haskell.org/package/collections-api-1.0.0.0/docs/Data-Collect.... foo is insert and bar is empty. The other three methods can be defined in terms of insert and empty.

On 21 April 2014 07:09, Bardur Arantsson wrote:

...
Hi all,

I was wondering if anyone out there knows if this type class has a name?

class Foo a e where foo :: e -> a -> a bar :: a

(I also have a fundep a -> e, but that's not essential.)

Essentially the usage is:

You have a sequence of "events" e_i and a starting value "bar" and can use the "foo" function to apply all events to that starting value

foo e_n $ foo e_{n-1} $ ... $ foo e_0 $ bar

and thus get the final value of the a induced by the events e_i.

(I've included the "bar" starting value in the type class, but I suppose it could be supplied by a Default instance.)

Regards,

_______________________________________________ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe

_______________________________________________ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe

Reply

Sign in to reply online Use email software

Bardur Arantsson

6:23 p.m.

On 2014-04-21 15:46, Andras Slemmer wrote:

I think what you're looking for are monoid actions: http://hackage.haskell.org/package/monoid-extras-0.2.2.0/docs/Data-Monoid-Ac... And the laws the typeclass should satisfy: foo bar = id foo f . foo g = foo (f `mappend` g)

The identity law is optional, in whihc case your 'e' would be a semigroup without a default 'bar'

Ah, yes, thanks. That looks about right, but unfortunately I have even less structure in that I don't have a monoid -- I had previously looked at group actions, but those have various extra laws. Maybe I should just call it "Action", but avoid the monoid bits. Thanks all,

Reply

Sign in to reply online Use email software

Sjoerd Visscher

10:09 p.m.

Just for completeness another option: http://hackage.haskell.org/package/reducers-3.10.2/docs/Data-Semigroup-Reduc... But this probably has the same problem, as it requires a Semigroup. Sjoerd On 21 Apr 2014, at 20:23, Bardur Arantsson wrote:

On 2014-04-21 15:46, Andras Slemmer wrote:

...
I think what you're looking for are monoid actions: http://hackage.haskell.org/package/monoid-extras-0.2.2.0/docs/Data-Monoid-Ac... And the laws the typeclass should satisfy: foo bar = id foo f . foo g = foo (f `mappend` g)

The identity law is optional, in whihc case your 'e' would be a semigroup without a default 'bar'

Ah, yes, thanks. That looks about right, but unfortunately I have even less structure in that I don't have a monoid -- I had previously looked at group actions, but those have various extra laws.

Maybe I should just call it "Action", but avoid the monoid bits.

Thanks all,

_______________________________________________ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe

Reply

Sign in to reply online Use email software

4095

Age (days ago)

4095

Last active (days ago)

Download

10 comments

6 participants

tags

participants (6)

Andras Slemmer
Bardur Arantsson
Ivan Lazar Miljenovic
Sjoerd Visscher
Tobias Brandt
Tom Ellis