Newbie Question: Are arguments in a function definition bound when function is called?



Robert Dodier wrote:
> On 8/26/09, Stavros Macrakis <macrakis at alum.mit.edu> wrote:
>
>   
>> Assignment (q: ...) and binding set the value slot; function
>> definition (q(x) := ...) sets the function slot. When evaluating the
>> function call q(...), Maxima first checks the function slot, and only
>> if there is no function value does it check the value slot.
>>     
>
> For the record, I know we've discussed it before, it seems like
> the existing function evaluation scheme is more complex than
> it needs to be. As I'm sure you know, it would be simpler to
> always look at the value slot or never. Today I'm inclined to
> say "always" but it seems to change depending on the weather.
>
> It occurs to me (and I would expect that it's been discussed
> in the Lisp-1 vs Lisp-2 literature) that we could have multiple slots
> but allow at most one of them to hold something.
>
> FWIW
>
> Robert Dodier
> _______________________________________________
> Maxima mailing list
> Maxima at math.utexas.edu
> http://www.math.utexas.edu/mailman/listinfo/maxima
>   
There are many items associated with symbols. The symbol cos has 
properties that pertain
to simplification, derivatives, integrals, perhaps inequalities, 
evaluation numerically,
etc.
If someone uses cos  as a local variable or a bound argument, do you 
want all of this
to disappear?

It is not that it would make a programming language unusable. After all, 
the Scheme
dialect of lisp does that.  But requiring

g(f,x):=    f(x)

to be written as g(f,x):= apply(f,[x])
   or perhaps (f)(x)

is pretty easy, and is apparent by looking at the code.

The other direction is pretty hard, and would, I think, require 
rethinking scope
(implementing lexical scope). Some programs would run just the same.  
Some would be very
subtly different. Some would be substantially broken, and some programs 
like subst
and ev would become (I fear) terribly confusing unless the users were 
first educated
about functions, functional arguments, closures, and scope.

Just off the top of my head though..
RJF