Re: What's the best language to learn? [was Re: Should I learn C or Pascal?]

[rgilbert@unconfigured.xvnews.domain (Bob Gilbert)]

In article <01bb8a24$88d46fe0$87ee6fce@timpent.airshields.com>, "Tim Behrendsen" <tim@airshields.com> writes:
> 
> All of that is very interesting in a theoretical sense.  But
> if we're talking about educating a student, you have to start
> somewhere, and to just hit them full-bore with "abstraction
> of algorithm" theory, I think you're going to lose of lot of
> them, *and we are!*

Some people are probably more comfortable learning from the bottom
up and do better when they can understand all the little details
that are involved when implementing some algorithm.  However, many
others may find that the reverse is better for them, that they
like to see the big picture first and avoid getting bogged down in
all the little details which they can later become familiar.
I would guess that your preference is the bottom up approach,
and that is fine, but it doesn't work for everybody.

> It seems to me that that focusing on the more practical
> aspects of implementation gives them a solid foundation to
> build the rest of it on, including building up of the concepts
> of languages and why their valuable.

Certainly there needs to be focus on the "aspects of implementation",
but that is separate from learning the concepts of algorithms.
After all, much of the field of discrete and iterative mathematics 
was invented and developed long before there was any machine on
which to implement the algorithm(s).  If everybody was limited to
only thinking about that which could currently be implemented on
available machines, we might never develop better algorithms that
could possibly be implemented on machines that will come in the
future.  In fact, the design of the machines should be driven by
the algorithms we have developed, not the other way around.
Of course we have to be realistic, and many times we have to 
learn work with what we have, so we still need to be able to 
adapt the algorithm, or even develop algorithms, to the 
constraints of the available machine.

So if we always take a bottom up approach, we will always be 
slaved to the constraints of the available machines which will
serve as the procrustean beds in which our algorithms must fit,
and we won't advance.  If we always take the top down approach,
we may never design algorithms for which we can build a machine
to implement it, and nothing will get done.  I really think the
tunnel needs to be dug from both ends of the mountain so that 
there is a greater chance of meeting somewhere in the middle.

-Bob