To my notices/questions on Haskell-98, such as

>> FiniteMap implemented via balanced binary trees.
>> Such a useful thing - has not it to move to the Standard library? ...


Simon Peyton-Jones <[EMAIL PROTECTED]>  writes

> Perhaps, but I'm not willing to add anything to the library at this
> stage.  Two months ago maybe. Now, no.

But let it be for Haskell-2  - ?


>>           span, foldl, minimum, sort, and so on.
>> These definitions, they, probably, only specify what map has to be 
>> computed. And the implementation is free to substitute any correct 
>> definition. Right?

> Yes, and the introduction to Appendix A tries to say this:
> [..]
> Should I clarify this more?  How?
> [..]
> .. Same for foldl, and your other examples.
> [..]
> Since this point clearly did not come over to you, can you suggest
> how I might state it more clearly?


No, no, you are right. 
I am sorry for reading only .txt file that sketches the differences, and 
recalling the impressions of Haskell-1.4. I have a spot of difficulty in
printing/reading large non-ASCII texts.
Also, from the history, i got the impression that it is safer to hide
the standard  sort, foldl, minimum, maximum, sum ...,  
and now expected the implementors to confirm: "do not hide them now, 
rely on implementation".  Of course, one could as well look into the 
sorce ...


>> Some operations with association lists.
>> In Haskell-1.4 libraries i could not find a good replacement for 
>> the following operations:

> Same as FiniteMap.  Reasonable suggestions, but too late.

Let them be for Haskell-2.


------------------
Sergey Mechveliani
[EMAIL PROTECTED]






Reply via email to