Thanks. I had assumed that in

data A = ...
data B = ...
data AorB = A | B

the A | B in the definition of AorB referred to the types A and B. But now I gather that they are completely unrelated -- except for the fact that they are spelled the same. 

Thanks.

-- Russ



On Tue, Dec 14, 2010 at 1:29 PM, Hector Guilarte <hectorg87@gmail.com> wrote:
But did you already understand what's happening?

> data AorB = A | B

is pretty much like an enumeration the same as

> data DayOfTheWeek = Monday | Tuesday | Wednesday | ... | Sunday

it stores no values of any type, you could even have DayOfTheWeek as
a constructor in the same declaration of the data DayOfTheWeek, because they are
in different scopes:

> data DayOfTheWeek = Monday | Tuesday | Wednesday | ... | Sunday | DayOfTheWeek

and by no means, that A or B on the right hand side of your data AorB are referencing the
data A nor the data B...

In the end, you could have:

data A = Aconstructor Int
data B = Bconstructor Int
data AorB = A A | B B

f :: Int -> AorB
f x 
  | even x     = A (Aconstructor x)
  | otherwise = B (Bconstructor x)

I just added a Value to each constructor of the data AorB, which happens to be
named the same as the value they store...

What I'm about to do is something I haven't tried, but I don't see why it shouldn't
compile:

> data Try = Int Int

in data Try you have a constructor named Int, and it has a value of type Int,

I'm trying to explain it the best I can, but I don't know if I managed to do it
clearly, please let me know if there's something where I wasn't clear enough.

Hector

 
On Tue, Dec 14, 2010 at 4:40 PM, Russ Abbott <russ.abbott@gmail.com> wrote:
My typo started this most recent confusion. When I wrote

data AorB = A | B

compiles with error.
That raises the question of what it really means!

I meant to say  

data AorB = A | B

compiles without error.
That raises the question of what it really means!

-- Russ 


On Tue, Dec 14, 2010 at 1:04 PM, Tobias Brandt <tob.brandt@googlemail.com> wrote:
On 14 December 2010 22:02, Hector Guilarte <hectorg87@gmail.com> wrote:
> Tobias you replied at the same time I was answering, you did explained what
> is happening,
> however, you said something which isn't right. As I mentioned before, "Type
> Constructors" and
> "Value Constructors" are in different scopes, and so, their names can be
> used again...
> if you don't believe me, give it a try with this example.

I believe you. As I said, I was talking crap :-)