Re: LSP

"Dmitry A. Kazakov" <mailbox@xxxxxxxxxxxxxxxxx> wrote:
On Fri, 22 Feb 2008 20:15:01 -0500, Daniel T. wrote:
S Perryman <q@xxxxx> wrote:

Completeness and decidability are the ultimate arbiters of
proof systems.

Given any two types, it is easy to discern wither one is
substitutable for the other and once a type is defined, the
properties that determine if it is substitutable for another type
are immutable, so I see no problem.

The problem is that your statement is wrong:

No, it's not.

class T
{
public :
int One (P& x) { return 1; }
};

class S
{
public :
int One (P& x) { return HALT (x); }
};

Is S substitutable for T?

First, you have to properly specify the types. All you have above is
some random code, not type specifications. Specify the types according
their Liskov contracts and it will be easy to tell.
.

Relevant Pages

  • Re: importNode and root node name
    ... When dealing with elements in a namespace you should instead use the ... method getElementsByTagNameNSand properly specify the namespace in ... You would select '/*' to get the root element as opposed to the root ... i'd have to specify the ns: ...
    (comp.lang.javascript)
  • RE: calculation accuracy
    ... also properly specify the literal as double precision, ... only 8 digits of it. ... Can we have a standardized way to specify pi? ...
    (comp.lang.fortran)