From: Richard A. O'Keefe <ok_at_cs.otago.ac.nz>

Date: Fri, 17 Nov 2006 15:35:50 +1300 (NZDT)

Date: Fri, 17 Nov 2006 15:35:50 +1300 (NZDT)

Jan Wielemaker <wielemak_at_science.uva.nl> wrote:

I vote for log10.

Why? What advantage does log10(Y) have over log(10,Y)?

min/max are *arithmetic* functions.

They are arithmetic functions ONLY in the sense that that's where they

will be put in the Prolog standard. Considered mathematically,

max and min belong to lattice theory (partial orders, total orders,

semi-lattices, lattices) not to arithmetic (groups, rings, fields, &c).

Floating-point contagion is a property of arithmetic operations NOT by

natural law or necessity but ONLY if we SAY it is. We have already seen

floating-point contagion blindly applied in a context where it ought not

to have been (numeric comparison); we should be fixing that blunder, not

handing it new conquests.

I'd say the result of min(2, 2.0) is undefined (i.e. implementations

can return either 2 or 2.0).

That's not undefined. That's _incompletely_ specified, but not undefined.

That might be a necessary compromise.

I note that Ada 95 manages very nicely without floating-point contagion

in any form, and is the better for it. Prolog would have been better if

we had required explicit conversions.

_______________________________________________

prolog-standard mailing list

prolog-standard_at_neve.di.ubi.pt

http://neve.di.ubi.pt/mailman/listinfo/prolog-standard

Received on Mon Jul 14 2008 - 13:01:02 EST

*
This archive was generated by hypermail 2.2.0
: Wed Sep 08 2010 - 23:28:17 EST
*