Memory efficiency and compile-time I/Os

[Andreas C. Doering]

Hi,
I am currently playing around whereby the following questions arose:

I use currently Hugs (1.4.)
1) My application has a huge amount of memory (from parsing a 
   text representation of an PCB, i.e. CAD data)
   Since the data is used in a totally different sequence 
   than it is read, I guess lazy evaluation does not help too much, 
   and therefore all this data has to be present during computation.
   What is the internal overhead (not only in Hugs) for Integers, 
   e.g. stort in lists, arrays etc. 
   That is, if I notice that my process gets too large and starts to swap 
   what can I do to reduce the size of the physical representation
2) In order to experiment I would like to have my input data like a static 
   data object. I know that my file is not changed (or better I _very_ strongly
   hope so!). 
   If I use readFile I have to put everything "into the loop". 
   For the final program (propably translated with ghc before execution) 
   this is no problem, but for playing around I really would like to have 
   an option for getting rid of the action.
   I mean if could load the CAD data when the program is loaded this would not 
   contradict the "Functional Philosophy" because I could translate the data
   into Haskell syntax and load it along with the code, though I would not
   like this with a file 4 MByte size. 
   We do not put "Action"-restrictions on functions coming from the source
   file, do we?
   I am aware that a general solution would be hard for a compiler, but in a
   interpreter?

Andreas
---------------------------------------------------------------
                        Andreas Doering
                        Medizinische Universitaet zu Luebeck
                        Institut fuer Technische Informatik
                        Ratzeburger Allee 160
                 
                        D-23538 Luebeck
                        Germany

                        Tel.: +49 451 500-3741
                        Fax:  +49 451 500-3687
                        Email: [EMAIL PROTECTED]
----------------------------------------------------------------