newbie type signature question
Please excuse my newbiness, but in this snippet: data (Monad m) => DataType m = DataType { f :: Char -> m () } test_function :: (Monad m) => DataType m -> m () ^^^^^^^^^^^^ test_function d = f d 'C' Why is "(Monad m) =>" required, when the definition of DataType already implies it? Is there an easier way to do this or will I have to have it in all signatures containing DataType? Thanks, Brock
Brock Peabody wrote:
Please excuse my newbiness, but in this snippet:
data (Monad m) => DataType m = DataType { f :: Char -> m () }
test_function :: (Monad m) => DataType m -> m () ^^^^^^^^^^^^ test_function d = f d 'C'
Why is "(Monad m) =>" required, when the definition of DataType already implies it? Is there an easier way to do this or will I have to have it in all signatures containing DataType?
This is exactly the same question I had when I started learning Haskell. The reason is that you are explicitly giving a type annotation for test_function, and whenever you give a type annotation what you see is what you get: the compiler doesn't add any extra things to it. If you had not given any type signature the inferred type would have been Monad m => DataType m -> m () as expected, whereas when you wrote test_function :: DataType m -> m () you are actively telling the compiler there is no context at all, but the compiler needs the context in this particular function in order to apply f. The context is not needed in all functions which use DataType - only in those functions which actually make use of m's Monad'ness, so you could write: test2 :: DataType m -> DataType m test2 x = x (See http://haskell.org/onlinereport/decls.html#user-defined-datatypes section 4.2.1 for more info) There was a post a while back (unfortunately I can't seem to locate it) where someone posted a link to some guidelines on haskell coding style where one guideline was never to use contexts in data declarations. Regards, Brian. -- Logic empowers us and Love gives us purpose. Yet still phantoms restless for eras long past, congealed in the present in unthought forms, strive mightily unseen to destroy us. http://www.metamilk.com
Brian Hulley wrote:
There was a post a while back (unfortunately I can't seem to locate it) where someone posted a link to some guidelines on haskell coding style where one guideline was never to use contexts in data declarations.
I would love to see that guideline. What is the correct way to express a constraint for a data declaration if this way is wrong? Thanks, Brock
Brock Peabody wrote:
Brian Hulley wrote:
There was a post a while back (unfortunately I can't seem to locate it) where someone posted a link to some guidelines on haskell coding style where one guideline was never to use contexts in data declarations.
I would love to see that guideline. What is the correct way to express a constraint for a data declaration if this way is wrong?
It is at http://www.informatik.uni-bremen.de/agbkb/forschung/formal_methods/CoFI/hets... in the section under "Types": "Don't put class constraints on a data type, constraints belong only to the functions that manipulate the data." So according to this guideline you're not supposed to think of associating contraints with data: constraints are only relevant for functions which manipulate the data, therefore (if you agree with this view) the very idea of associating constraints with data is wrong. It is possible that this feature was added to the language for the benefit of people who prefer not to use explicit type signatures but afaiu this goes against best practice where everything should always have an explicit signature to make code easy to understand and facilitate debugging of type errors. Of course this does not apply to the completely different use of constraints when defining existential types. Regards, Brian. -- Logic empowers us and Love gives us purpose. Yet still phantoms restless for eras long past, congealed in the present in unthought forms, strive mightily unseen to destroy us. http://www.metamilk.com
Hello Brian, Saturday, June 10, 2006, 3:05:25 AM, you wrote:
It is possible that this feature was added to the language for the benefit of people who prefer not to use explicit type signatures but afaiu this goes against best practice where everything should always have an explicit signature to make code easy to understand and facilitate debugging of type errors.
when you work with C++ or some other OOP language, you can define that some field in structure should some some specific interface and this allows to use functions of this interface on this field. i required the same feature in Haskell, for example: data UTF8Stream h = (ByteStream h) => UTF8Stream h instance TextStream (UTF8Stream h) ... addUTF8Encoding :: h -> (UTF8Stream h) and so on. currently i should add type constraint to each and every class and function i declared. i don't tried using existential types here, so i'm not sure that they will give the same high speed for inlined functions. moreover, they "lose" information about other type classes that 'h' supports, although functions from these classes may be required for application that use "UTF8Stream h" and even by other definitions in my lib: instance (ByteStringStream h) => ByteStringStream (UTF8Stream h) ... -- Best regards, Bulat mailto:Bulat.Ziganshin@gmail.com
From: Bulat Ziganshin [mailto:bulat.ziganshin@gmail.com]
when you work with C++ or some other OOP language, you can define that some field in structure should some some specific interface and this allows to use functions of this interface on this field. i required the same feature in Haskell, for example:
I come from a C++ background, and there is no formal way in the language to specify constraints for type (template) parameters; a major weakness for generic programming. Some interesting work is being done to add this ability though, ex: http://www.boost.org/libs/concept_check/concept_check.htm. Regards, Brock
Bulat Ziganshin wrote:
Hello Brian,
Saturday, June 10, 2006, 3:05:25 AM, you wrote:
It is possible that this feature was added to the language for the benefit of people who prefer not to use explicit type signatures but afaiu this goes against best practice where everything should always have an explicit signature to make code easy to understand and facilitate debugging of type errors.
when you work with C++ or some other OOP language, you can define that some field in structure should some some specific interface and this allows to use functions of this interface on this field. i required the same feature in Haskell, for example:
data UTF8Stream h = (ByteStream h) => UTF8Stream h
instance TextStream (UTF8Stream h) ...
addUTF8Encoding :: h -> (UTF8Stream h)
and so on. currently i should add type constraint to each and every class and function i declared.
Hello Bulat - Thanks for the example. Afaiu, if you don't write any type signature at all, type inference will infer the constraint, at least this is what I understand from http://haskell.org/onlinereport/decls.html section 4.2.1: data Eq a => Set a = NilSet | ConsSet a (Set a) Pattern matching against ConsSet also gives rise to an Eq a constraint. For example: f (ConsSet a s) = a the function f has inferred type Eq a => Set a -> a. At the moment when you have an explicit type signature you can see exactly what the function needs, so that if you later refactored the code you could see which functions actually used the Eq a constraint and which ones didn't. For example a function to find the size of a set probably wouldn't need Eq a. So if the compiler was allowed to add a context to an explicit type signature you would no longer be able to just look at the type signature for a function to find out what it was making use of - you'd need to also look at all the data declarations and gather all the constraints together, to arrive at the full type signature. Nevertheless I suppose it would be useful to have some kind of compiler flag to allow this - certainly given that contexts are allowed on data declarations it is kind of awkward having to "repeat" all the info on each function that uses that type, since at the moment the only way to avoid this repetition is to avoid giving functions that use that type any type signatures at all, which seems worse. Or perhaps there could be a special syntax to indicate a partial type signature, that the compiler would complete by adding the contexts eg: f ::: Set a -> a ^^^ ? 3 colons to indicate that the compiler will add the relevant context(s)
i don't tried using existential types here, so i'm not sure that they will give the same high speed for inlined functions. moreover, they "lose" information about other type classes that 'h' supports, although functions from these classes may be required for application that use "UTF8Stream h" and even by other definitions in my lib:
instance (ByteStringStream h) => ByteStringStream (UTF8Stream h) ...
Yes I also think they'd be slower and not recommended unless they're actually needed eg for getting a common interface to objects whose concrete types aren't known at compile time (being like virtual functions in C++). I just mentioned them as an example of something not relevant to the guideline. (Whoever wrote the guideline might have had more reasons but that's all I can think up so far.) Best regards, Brian. -- Logic empowers us and Love gives us purpose. Yet still phantoms restless for eras long past, congealed in the present in unthought forms, strive mightily unseen to destroy us. http://www.metamilk.com
Hello Brian, Tuesday, June 13, 2006, 1:11:37 AM, you wrote:
data UTF8Stream h = (ByteStream h) => UTF8Stream h
instance TextStream (UTF8Stream h) ...
addUTF8Encoding :: h -> (UTF8Stream h)
and so on. currently i should add type constraint to each and every class and function i declared.
Hello Bulat - Thanks for the example. Afaiu, if you don't write any type signature at all, type inference will infer the constraint,
of course, but 1) it is considered as bad programming style 2) haddock currently can't infer function types 3) it doesn't work for class headers 4) functions that create UTF8Stream don't use ByteStream functions
Or perhaps there could be a special syntax to indicate a partial type signature, that the compiler would complete by adding the contexts eg:
f ::: Set a -> a ^^^ ? 3 colons to indicate that the compiler will add the relevant context(s)
partial signatures is also among the proposals for future Haskell, and Oleg has technique to make partial signatures even in H98
i don't tried using existential types
Yes I also think they'd be slower
Simon Marlow once said it can be faster. i don't know details, though. something like the dictionaries for my types are constructed on each call (?), while dictionaries for existential type constructed only once, on the moment of encapsulation -- Best regards, Bulat mailto:Bulat.Ziganshin@gmail.com
Brian Hulley wrote:
"Don't put class constraints on a data type, constraints belong only to the functions that manipulate the data."
So according to this guideline you're not supposed to think of associating contraints with data: constraints are only relevant for functions which manipulate the data, therefore (if you agree with this view) the very idea of associating constraints with data is wrong.
I wonder what the reasoning for this guideline is. It seems that it would require one to repeat the constraint over and over. For the Prelude Data.Map type, for instance, wouldn't almost every function have to have the constraint that the key type is ordered? Also, I think it would push detection of the error further from the place where it could have been detected. Doesn't it make more sense to detect that there is an error when a user tries to instantiate a map with an unordered type rather than later when an attempt is made to use it? Regards, Brock
Hello Brian, Friday, June 9, 2006, 9:50:30 PM, you wrote:
data (Monad m) => DataType m = DataType { f :: Char -> m () }
test_function :: (Monad m) => DataType m -> m () ^^^^^^^^^^^^ There was a post a while back (unfortunately I can't seem to locate it) where someone posted a link to some guidelines on haskell coding style where one guideline was never to use contexts in data declarations.
afaiu, this is considered as drawback of H98, and may be even addressed some day. at least in h' mail-list it was discussed: adding the datatype context to the context of functions that use it. we can even ask GHC team to add this as one more glasgow extension -- Best regards, Bulat mailto:Bulat.Ziganshin@gmail.com
Sorry, I meant to send this to the whole list. Brock Peabody wrote:
Please excuse my newbiness, but in this snippet:
data (Monad m) => DataType m = DataType { f :: Char -> m () }
test_function :: (Monad m) => DataType m -> m () ^^^^^^^^^^^^ test_function d = f d 'C'
Why is "(Monad m) =>" required, when the definition of DataType already implies it? Is there an easier way to do this or will I have to have it in all signatures containing DataType?
Because class constraints on data types are a bit silly. It just restricts the types of the constructors so you can only apply them if m is a Monad. It doesn't actually package up the evidance that m is a monad inside the value to make this stuff work. Existential types do package up the class instance in the value, but they hide the type. data Showable = forall a . (Show a) => Showable a showShowable :: Showable -> String showShowable (Showable x) = show x Getting them both is tricky, but you can do it if you use a GADT to write a type that means "exists a such that a = m and a is a Monad": {-# OPTIONS -fglasgow-exts #-} data TyEq (a :: * -> *) (b :: * -> *) where Refl :: TyEq a a data DataType m = forall m' . (Monad m') => DataType (TyEq m m') (Char -> m' ()) buildDataType :: (Monad m) => (Char -> m ()) -> DataType m buildDataType = DataType Refl test_function :: DataType m -> m () test_function (DataType Refl f) = f 'C' -- try let x = buildDataType putChar -- :t x -- :t test_function -- test_function x Brandon
Thanks,
Brock
_______________________________________________ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
participants (5)
-
Brandon Moore -
Brian Hulley -
Brock Peabody -
Bulat Ziganshin -
J. Garrett Morris