Re: Non-declared Variables

Quinn Tyler Jackson <qtj-query@shaw.ca> writes:


>> I don't think that it is a good idea when language definition allows
>> you to use variables without declaration.
>
> I think languages should forbid the declaration of variables, for the
> following reasons:
> [...good reasons...]


But there's one thing where you need to declare the variables, it's
when you have lexical scope.


{
      x=0;
      {
            x=1;
            print x;
      }
      print x;
}


What does this print? Is there one or two variables? What about: {{
x=1; } print x;} and other such cases? Of course you can have only
dynamic variables and no lexical scope.


But I think the conclusion reached 20 years ago was that having only
dynamic scope without lexical scope was bad, and even that it'd be
better to have only lexical scoping...


That said, in lisp for example, we don't really declare the variables,
we directly bind them:


(let ((x 0))
      (let ((x 1))
            (print x))
      (print x))


and we get printed 0 and 1.






[Additionnal declarations can be inserted, but are optional.


  (let ((x 0))
        (declare (type integer x)
                          (special x))
        (let ((x 1.0))
              (declare (type float x))
              (print x))
        (print x))
]




--
__Pascal Bourguignon__ http://www.informatimago.com/


"Klingon function calls do not have "parameters" -- they have
"arguments" and they ALWAYS WIN THEM."

Return to the comp.compilers page.
Search the comp.compilers archives again.