Re: [Axiom-developer] "has" and "with" (was curious algebra failure)

[Ralf Hemmecke]

---BEGIN aaa.as
#include "aldor"
#include "aldorio"

main(): () == {
        import from Integer;
        T1 == List Integer;
        T2 == List Integer;
        t1: T1 := [1];
        t2: T2 := t1;
        stdout << "t1 = " << t1 << newline;
        stdout << "t2 = " << t2 << newline;

#if DoesntWork
        N1: ListType Integer == List Integer;
        N2: ListType Integer == List Integer;
        n1: N1 := [1];
        n2: N2 := n1;
        stdout << "n1 = " << n1 << newline;
        stdout << "n2 = " << n2 << newline;
#endif
}
main();
---END aaa.as

woodpecker:~/scratch>aldor -laldor -DDoesntWork -fx aaa.as
"aaa.as", line 17:         n2: N2 := n1;
                   ..................^
[L17 C19] #1 (Error) There is no suitable interpretation for the 
expression n1
  The context requires an expression of type N2.
     The possible types of the right hand side (`n1') are:
          -- N1

If you remove "-DDoesntWork" the code compiles and runs fine.
In Aldor T1 and T2 are identical. Constant assignment does not define a 
new type. That is different if you add the type (as is done for N1 and 
N2). The compiler introduces two new domains N1 and N2 which are like 
but different from List Integer.

Maybe that is confusing, but that's the way it is. (And it is sometimes 
useful.)

Ralf

On 08/13/2007 12:35 AM, Weiss, Juergen wrote:
> I found the old document about type equivalence in Scratchpad. It's:
> A New Algebra System, May, 29 th 1984, James H Davenport. I have only
> found a paper version. Maybe someone has an online version.
>
> It states that:
> 1. Two named types are only equivalent if the names are the same.
> 2. Anonymous types are equivalent when stucturally equivalent
> 3. An anonymous type is never equivalent to a named type.
>
> So following 1.
>
> t1 == List Term
> t2 == List Term
> x : t1
> y : t2
> y := x      
>
> is not supposed to work,
>
> but following 2.
>
> x : List Term
> y : List Term
> y := x    
>
> is ok,
>
> and following 3.
>
> t == List Term
> x : List Term
> y : t
> y := x   
>
> is not supposed to work as well
>
> All examples are taken from the paper.
>
> I am not sure how much of this design is preserved in the current
> system. But without having had an intense look at the examples,
> I got the impression, that they follow the rules above.
>
> Regards
>
> Juergen Weiss
>
> Juergen Weiss     | Universitaet Mainz, Zentrum fuer Datenverarbeitung,
> address@hidden| 55099 Mainz, Tel: +49(6131)39-26361, FAX:
> +49(6131)39-26407
>  
>
> > -----Original Message-----
> > From: address@hidden 
> > [mailto:address@hidden
> >  On Behalf Of Gabriel Dos Reis
> > Sent: Monday, August 13, 2007 12:16 AM
> > To: Bill Page
> > Cc: address@hidden
> > Subject: Re: [Axiom-developer] "has" and "with" (was curious 
> > algebra failure)
> >
> > On Sun, 12 Aug 2007, Bill Page wrote:
> >
> > | On 8/12/07, Gabriel Dos Reis wrote:
> > | > On Sun, 12 Aug 2007, Bill Page wrote:
> > | >
> > | > | If you prefer, now that we have this page nearly 
> > finished we could
> > | > | rename it to remove the SandBox prefix, then all 
> > subscribers would be
> > | > | notified of any further changes.
> > | >
> > | > That would be great!
> > | >
> > | 
> > | Ok, now please refer to the page:
> > | 
> > | http://wiki.axiom-developer.org/AnonymousCategories
> >
> > Many thanks.
> >
> > -- Gaby
> >
> >
> >
> > _______________________________________________
> > Axiom-developer mailing list
> > address@hidden
> > http://lists.nongnu.org/mailman/listinfo/axiom-developer
>
>
> _______________________________________________
> Axiom-developer mailing list
> address@hidden
> http://lists.nongnu.org/mailman/listinfo/axiom-developer