Re: [colorforth] explicit vs. implicit
- Subject: Re: [colorforth] explicit vs. implicit
- From: "Samuel A. Falvo II" <kc5tja@xxxxxxxx>
- Date: Wed, 18 Aug 2004 03:27:12 -0700
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
On Wednesday 18 August 2004 02:57 am, stephen white wrote:
> On 18/08/2004, at 4:23 AM, Mark Slicker wrote:
> > + * dup +
> >
> > I get:
> >
> > ____+_____
> > / \
> > * *
> > / \ / \
> > x + x +
> > / \ / \
> > y z y z
>
> Aren't you only DUP'ing the result so far?
The nice thing about a functional programming language is that there is
absolutely *zero* distinction between 3+4, 7, and x+y, where x=2 and
y=5. This is mandated by the principle of referencial (referential?)
transparency.
Well, Forth isn't exactly a pure functional language. But even so,
DUPing the top of stack has the effect of computing the same function
with the same arguments twice -- in effect, replicating the entire
dataflow through the function tree as if it were a unique branch.
Therefore, by the principle of referential transparency, Mark's example
still holds.
> If you DUP then change the value of z before +'ing, you would get the
> same result as it's not really two trees being evaluated...
DUPing replicates the precise state of the dataflow -- changing z after
the DUP creates a new dataflow. Remember: changing z occurs after the
DUP has already been completed.
- --
Samuel A. Falvo II
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)
iD8DBQFBIy8GvwDm/l0jx/IRAnNrAKCFwAQSi5AB3VMw36bWKi1PTSm4YQCggcfA
h8+9vi3TkFHMEmnZKvNothE=
=zpio
-----END PGP SIGNATURE-----
---------------------------------------------------------------------
To unsubscribe, e-mail: colorforth-unsubscribe@xxxxxxxxxxxxxxxxxx
For additional commands, e-mail: colorforth-help@xxxxxxxxxxxxxxxxxx
Main web page - http://www.colorforth.com
