{- I am trying to understand constructor classes and their relationship with ordinary type classes. I wrote the code below to help me understand the distinction. The code is only for explanatory purposes, set operations use a tuple syntax. I use the naming convention of 'typeVar', 'typeCons', and 'dataCons' for type variables, type constructors, and data constructors respectively. Question1: Is my naming convention correct? I am particularly concerned about SetClass3 where the super class seems to use a type constructor but the subclass seems to use the same term as a data constructor. Q2: In this case constructor classes and type classes seem to provide similar functionality. In general what situation are each best suited? -} import Data.List data SetType typeVar = SetType [typeVar] deriving Show class SetClass1 typeCons where member1 :: Eq typeVar => (typeVar, typeCons typeVar) -> Bool intersect1 :: (Eq typeVar,Show typeVar) => (typeCons typeVar, typeCons typeVar) -> typeCons typeVar class SetClass2 dataCons typeVar where member2 :: (typeVar, (dataCons typeVar)) -> Bool intersect2 :: (dataCons typeVar, dataCons typeVar) -> dataCons typeVar class SetClass1 typeCons => SetClass3 typeCons dataVariable where union3 :: (Eq dataVariable,Show dataVariable) => (typeCons dataVariable, typeCons dataVariable) -> typeCons dataVariable instance SetClass1 SetType where member1 (x ,(SetType y)) = elem x y intersect1 ((SetType x),(SetType y)) = SetType (intersect x y) instance SetClass2 SetType Int where member2 (x ,SetType y) = elem x y intersect2 (SetType x,SetType y) = SetType (intersect x y) instance SetClass3 SetType Int where union3 (SetType x,SetType y) = SetType (union x y) test1a = member1 (1, (SetType [1,2])) test1b = intersect1 ((SetType [1,3,4]),(SetType [1,2])) test2a = member2 (1, (SetType [1::Int])) test2b = intersect2 ((SetType [1::Int]), (SetType [(1::Int)])) test3a = union3 ((SetType [1::Int,2::Int,3::Int]),(SetType [4::Int,5::Int])) This email originated from DIT. If you received this email in error, please delete it from your system. Please note that if you are not the named addressee, disclosing, copying, distributing or taking any action based on the contents of this email or attachments is prohibited. www.dit.ie Is ó ITBÁC a tháinig an ríomhphost seo. Má fuair tú an ríomhphost seo trí earráid, scrios de do chóras é le do thoil. Tabhair ar aird, mura tú an seolaí ainmnithe, go bhfuil dianchosc ar aon nochtadh, aon chóipeáil, aon dáileadh nó ar aon ghníomh a dhéanfar bunaithe ar an ábhar atá sa ríomhphost nó sna hiatáin seo. www.dit.ie Tá ITBÁC ag aistriú go Gráinseach Ghormáin – DIT is on the move to Grangegorman _______________________________________________ Beginners mailing list [hidden email] http://mail.haskell.org/cgi-bin/mailman/listinfo/beginners |
My intended use of the term 'subclass' is as the Haskell report: "class Num of numeric types is a subclass of Eq, since all numbers may be compared for equality" Of course what I have written may not be consistent with the report. Regards, Pat On 31 October 2016 at 09:26, Imants Cekusins <[hidden email]> wrote:
This email originated from DIT. If you received this email in error, please delete it from your system. Please note that if you are not the named addressee, disclosing, copying, distributing or taking any action based on the contents of this email or attachments is prohibited. www.dit.ie Is ó ITBÁC a tháinig an ríomhphost seo. Má fuair tú an ríomhphost seo trí earráid, scrios de do chóras é le do thoil. Tabhair ar aird, mura tú an seolaí ainmnithe, go bhfuil dianchosc ar aon nochtadh, aon chóipeáil, aon dáileadh nó ar aon ghníomh a dhéanfar bunaithe ar an ábhar atá sa ríomhphost nó sna hiatáin seo. www.dit.ie Tá ITBÁC ag aistriú go Gráinseach Ghormáin – DIT is on the move to Grangegorman _______________________________________________ Beginners mailing list [hidden email] http://mail.haskell.org/cgi-bin/mailman/listinfo/beginners |
In reply to this post by Patrick Browne
Hi Patrick,
I am not sure I understand what you are trying to do, but here are some thoughts. - your use of the names `typeVar` and `typeCons` make total sense to me. But your name `dataCons` in the definition of SetClass2 does not. You are only talking about types here. All you are doing is talking about *type* constructors. You are not constructing a specific value of that type. Maybe one way to see the difference between the type constructors and the data constructors is to change your first line to: data SetType typeVar = SetTypeDataConstructor [typeVar] deriving Show - I also don't understand why you named your type variable `dataVariable` instead of sticking to `typeVar` in SetClass3. It is a *type* variable in the type signature of the functions in the type class. - I don't understand what you mean by a "constructor class". My simple understanding of what a type constructors and data constructors are is here https://wiki.haskell.org/Constructor#Type_constructor. - I don't see the terms super class and sub class ever used in haskell land. Although I do understand they make sense, I think they are really OO terms. I've tried to map OO concepts into haskell and there are many different ways to do it. The best description of all those ways that I found was in Oleg Kiselyov's work https://arxiv.org/pdf/cs/0509027.pdf - I think parametric polymorphism as used in haskell forces us to compare types in terms of equality constraints (this type is "the same" as that), whereas subtype polymorphism as used in OO languages forces us to think in terms of inequalities (this type is "more specific" than that). I think those force fundamentally different ways to think about and model your problems. I would encourage you to use equality as much as possible, but if you want to model subtyping in haskell, I think Oleg's work above has a few approaches that I think are better than type classes. In particular, in OO there was always an ambiguity between modelling subtyping with inheritance or with a "has a" relationship when the subclass explicitly contains an object of the superclass. You can use the latter idea in haskell, Oleg's work has a good way to do it. Hope this helps. Cheers, Dimitri > From: PATRICK BROWNE <[hidden email]> > To: The Haskell-Beginners Mailing List - Discussion of primarily > beginner-level topics related to Haskell <[hidden email]> > Subject: [Haskell-beginners] Constructor classes and type classes. > Message-ID: > <CAGFLrKfQVaQQJybMTn-pJksV1VKUXrmBks14Jkm==Zqt=[hidden email]> > Content-Type: text/plain; charset="utf-8" > > {- > I am trying to understand constructor classes and their relationship with > ordinary type classes. > I wrote the code below to help me understand the distinction. The code is > only for explanatory purposes, set operations use a tuple syntax. > > I use the naming convention of 'typeVar', 'typeCons', and 'dataCons' for > type variables, type constructors, and data constructors respectively. > Question1: > Is my naming convention correct? > I am particularly concerned about SetClass3 where the super class seems to > use a type constructor but the subclass seems to use the same term as a > data constructor. > > Q2: > In this case constructor classes and type classes seem to provide similar > functionality. > In general what situation are each best suited? > -} > > import Data.List > data SetType typeVar = SetType [typeVar] deriving Show > > > class SetClass1 typeCons where > member1 :: Eq typeVar => (typeVar, typeCons typeVar) -> Bool > intersect1 :: (Eq typeVar,Show typeVar) => (typeCons typeVar, typeCons > typeVar) -> typeCons typeVar > > > class SetClass2 dataCons typeVar where > member2 :: (typeVar, (dataCons typeVar)) -> Bool > intersect2 :: (dataCons typeVar, dataCons typeVar) -> dataCons typeVar > > > class SetClass1 typeCons => SetClass3 typeCons dataVariable where > union3 :: (Eq dataVariable,Show dataVariable) => (typeCons dataVariable, > typeCons dataVariable) -> typeCons dataVariable > > instance SetClass1 SetType where > member1 (x ,(SetType y)) = elem x y > intersect1 ((SetType x),(SetType y)) = SetType (intersect x y) > > instance SetClass2 SetType Int where > member2 (x ,SetType y) = elem x y > intersect2 (SetType x,SetType y) = SetType (intersect x y) > > instance SetClass3 SetType Int where > union3 (SetType x,SetType y) = SetType (union x y) > > > test1a = member1 (1, (SetType [1,2])) > test1b = intersect1 ((SetType [1,3,4]),(SetType [1,2])) > test2a = member2 (1, (SetType [1::Int])) > test2b = intersect2 ((SetType [1::Int]), (SetType [(1::Int)])) > test3a = union3 ((SetType [1::Int,2::Int,3::Int]),(SetType [4::Int,5::Int])) > -- 2E45 D376 A744 C671 5100 A261 210B 8461 0FB0 CA1F -- 2E45 D376 A744 C671 5100 A261 210B 8461 0FB0 CA1F _______________________________________________ Beginners mailing list [hidden email] http://mail.haskell.org/cgi-bin/mailman/listinfo/beginners |
Dimitri, Thanks for taking the time and effort to respond. Your comments have clarified some of my confusion. My motivation is that I am trying to implement the code from a paper with the following data and classes: {-# LANGUAGE MultiParamTypeClasses #-} data TimeFacts t = Bef (t, t, TimeFacts t) | New class SetClass t where class SetClass t => Time l t where class Time l t => TimeE l t where class Time l t => TimeTO l t where However I cannot get the instances to compile: instance SetClass (TimeFacts t) where -- OK instance (SetClass t) => Time TimeFacts t where -- Expecting one more argument to 'TimeFacts' instance TimeE TimeFacts t where -- Expecting one more argument to 'TimeFacts' instance TimeTO TimeFacts t where -- Expecting one more argument to 'TimeFacts' Note, I am only using the headings for the classes and instances, which should be OK. Any pointers to a solution would be much appreciated. Regards, Pat This email originated from DIT. If you received this email in error, please delete it from your system. Please note that if you are not the named addressee, disclosing, copying, distributing or taking any action based on the contents of this email or attachments is prohibited. www.dit.ie Is ó ITBÁC a tháinig an ríomhphost seo. Má fuair tú an ríomhphost seo trí earráid, scrios de do chóras é le do thoil. Tabhair ar aird, mura tú an seolaí ainmnithe, go bhfuil dianchosc ar aon nochtadh, aon chóipeáil, aon dáileadh nó ar aon ghníomh a dhéanfar bunaithe ar an ábhar atá sa ríomhphost nó sna hiatáin seo. www.dit.ie Tá ITBÁC ag aistriú go Gráinseach Ghormáin – DIT is on the move to Grangegorman _______________________________________________ Beginners mailing list [hidden email] http://mail.haskell.org/cgi-bin/mailman/listinfo/beginners |
this builds. not sure if the instances are as intended {-# LANGUAGE MultiParamTypeClasses, FlexibleInstances #-} module SetClass where data TimeFacts t = Bef (t, t, TimeFacts t) | New class SetClass t where class SetClass t => Time l t where class Time l t => TimeE l t where class Time l t => TimeTO l t where instance SetClass (TimeFacts t) where -- OK instance (SetClass t) => Time (TimeFacts t) t where instance (SetClass t) => TimeE (TimeFacts t) t where instance (SetClass t) => TimeTO (TimeFacts t) t where _______________________________________________ Beginners mailing list [hidden email] http://mail.haskell.org/cgi-bin/mailman/listinfo/beginners |
In reply to this post by Patrick Browne
Imants Thanks for the sanity check on the instances.I am concerned about the use of the function immsucc' from the same paper, which is defined in instances the Time class (but omitted from the class). The function immsucc' is first used without arguments, which to me implies that it 'might' a function passed to the higher-order closure function. It is then used as just an ordinary function with arguments. I have found a version of closure based on Craft of Function Programming. which does not require a function as an argument. So basically my question is: does the usage of immsucc' make sense in the code below?, do the types match up? Regards, Pat data TimeFacts t = Bef (t, t, TimeFacts t) | New class SetClass t => Time l t where new :: (t, t) -> l t bef :: (t, t, l t) -> l t succ , pred :: (t, l t) -> Set t immsucc , immpred :: (t, l t) -> Set t before, after :: (t, t, l t) -> Bool immedbefore, immedafter :: (t, t, l t) -> Bool immedafter (a, b, t) = immedbefore (b, a, t) after (a, b, t) = before (b, a, t) before (a, b, t) = memberInSet (b, succ (a, t)) succ (a, tl) = closure ([a], immsucc' , tl) pred (a, tl) = closure ([a], immpred' , tl) immsucc (a, tl) = immsucc' ([a], [], tl) immpred (a, tl) = immpred' ([a], [], tl) instance (SetClass t) => Time TimeFacts t where immsucc’ (a, s, New ) = s immsucc’ (a, s, Bef (e, f, tt)) = if memberInSet (e,a) then immsucc’ (a, insertInSet (f,s), tt) else immsucc’ (a, s,tt) On 3 November 2016 at 18:05, Imants Cekusins <[hidden email]> wrote:
This email originated from DIT. If you received this email in error, please delete it from your system. Please note that if you are not the named addressee, disclosing, copying, distributing or taking any action based on the contents of this email or attachments is prohibited. www.dit.ie Is ó ITBÁC a tháinig an ríomhphost seo. Má fuair tú an ríomhphost seo trí earráid, scrios de do chóras é le do thoil. Tabhair ar aird, mura tú an seolaí ainmnithe, go bhfuil dianchosc ar aon nochtadh, aon chóipeáil, aon dáileadh nó ar aon ghníomh a dhéanfar bunaithe ar an ábhar atá sa ríomhphost nó sna hiatáin seo. www.dit.ie Tá ITBÁC ag aistriú go Gráinseach Ghormáin – DIT is on the move to Grangegorman _______________________________________________ Beginners mailing list [hidden email] http://mail.haskell.org/cgi-bin/mailman/listinfo/beginners |
Free forum by Nabble | Edit this page |