On Tue, Dec 2, 2008 at 4:40 AM, Mukul Gandhi <> wrote:
> Hi Mike,
> Thanks for your thoughts.
>
> I agree that functional languages are quite useful. But I think the
> problem is, many of the underlying hardware architectures are von
> Neumann based. von Neumann machines work best with imperatives
> programs (languages).
This is broad based conclusion ... I agree that at a lower level
things very much tend to be imperative (machine code) but the picture
is very unclear in higher order abstraction.
I offer perl as a classic example of ambiguity in that one program in
an imperative style as you could a declarative or OO style ... and its
been said that perl shares several of the attributes of lisp once you
start digging around making it a 'sleeper' functional language.
I guess the point in categorization is the ability to bring a set of
recognized techniques to bear... for example I think that imperative
programs excels wherever one has a problem that is easily
characterized by a state machine.
This however does not guarantee that a developer will then create a
standard state machine in favorite language .. we all tend to reinvent
yet another slightly different version of a state machine instead of
reuse.
> I think building machine architectures which can support functional
> paradigm is challenging ...
I might be wrong, but once again is not machine code imperative and on
top of that there can exist several functional programming languages ?
IMO the most compelling consideration in choosing a programming
approach revolves around the requirements of the data structure
needed.
If XML is being used, then using programming languages that support
XML or things like XSLT/XQuery.
You may find that this is not the right choice for manipulating images
(and blobs in RDBMS are just as bad) or for doing low level binary
operations on a hard drive.
cheers, Jim Fuller
