In relation to the recent discussion about efficiencies
        vs. reduction count, timing, etc.
        Module Orthogonals has been completely revised,
        enhanced and generalized:
        (http://www.numeric-quest.com/haskell/Orthogonals.html)

        Excerpt:
        -------
        Arrays - with their indexing schemes and destructive
        updating are basic tools of imperative programming. But pure
        functional languages, which prohibit variable reassignments,
        cannot compete with imperative languages by using the same tools
        and following similar resoning and patterns. 
                With this in mind, we are attempting to demonstrate
        here that the standard indexing traps can be successfully avoided.
        This module implements afresh several standard problems from
        linear algebra. Standard Haskell lists are employed instead of
        arrays and not a single algorithm ever uses indexes for lookups or
        updates. 
                We do not claim high efficiency of these algorithms;
        consider them exploratory. However, we do claim that the clarity
        of these algorithms is significantly better than of those
        functionally similar algorithms, which employ indexing schemes. 

        Jan



Reply via email to