Re: [Axiom-developer] Question concerning types...

[Ralf Hemmecke]

Ah... I forgot to say... if ever indefinite integers come to life then a 
syntax like

i: Integer

would be quite acceptable/preferrable.

Maybe I am wrong, but is there use for indefinite integers in library 
code or is that rather a use case for an interactive session?

If it's only for an interactive session, then I believe the interpreter 
should be made so smart to deal with "programming variables" (like the i 
above) that have no values as indefinite objects.

Note that even if SingleInteger just has 32 bits to store a value, the 
interpreter adds some knowledge anyway (name of the variable that holds 
the SingleInteger, the type of the value, etc.). So it would be simple 
for the interpreter to add another bit saying that this particular 
variable does not have a value.

If I look at

(1) -> n: Integer

Type: Void
(2) -> sin(2*%Pi*n)

   n is declared as being in Integer but has not been given a value.

then I believe the interpreter already does that. It just is not smart 
enough to apply the information it has.

Ralf