Re: Data.FiniteMap proposed addition, bug fix

10 Nov 2004

      At 23:35 09/11/04 +0100, Jean-Philippe Bernardy wrote:
...
On Tue, 09 Nov 2004 15:22:53 +0000, Graham Klyne  wrote:
...
At 13:06 09/11/04 +0000, Simon Marlow wrote:
...
On 09 November 2004 12:45, Graham Klyne wrote:
...
I'd like to propose an addition to the FiniteMap module in the form
of a monadic version of plusFM_C.  The proposed implementation is
pretty much a copy of the existing implementation within a do block.
Data.FiniteMap will shortly be deprecated in favour of DData.Map (which
will be renamed to Data.Map when it is imported).  Perhaps you'd like to
reformulate the proposal using Data.Map instead?
Proposed addition:
[***]
Ok, but in order to have a consistent interface, we'll need monadic
versions of a lot of functions: there are Set, IntMap, IntSet types,
and intersection, difference, etc. Might be a lot of work. A solution
may be to make standard versions a special case of the monadic ones,
still I wonder about performance issues.
I hadn't thought about that specifically, but I did think about other 
similar functions on Map values, and it raises questions of both utility 
and consistency.

My suggestion comes from a specific and real requirement in some software 
I'm writing.  Because the actual type is abstract (hidden), I can't do it 
externally to the module, other than by using several function calls to 
extract, merge, delete and re-insert elements, which doesn't seem desirable.

In this case, I decided to push for the particular function that I knew I 
wanted, and to not expend energy on other functions whose actual utility 
was uncertain.

FWIW, other candidate functions for this treatment that I see would be:
             , insertWith, insertWithKey, insertLookupWithKey
             , differenceWith
             , differenceWithKey
             , intersectionWith
             , intersectionWithKey
             , fromListWith
             , fromListWithKey
             , fromAscListWith
             , fromAscListWithKey

Hmmm, I also note:
             , mapAccum
             , mapAccumWithKey
and wonder why these aren't done with a monadic function?   It looks a bit 
like a state monad operation to me.

#g

------------
Graham Klyne
For email:
http://www.ninebynine.org/#Contact