Since I was a C programmer for so long before Perl, I occasionally mistype my code and put == when I ought to use eq. When it does happen, I get a message like Argument "" isn't numeric in ... because I use warnings;, and then it's easy enough to fix. That said, and compiler dependencies aside, there is a purely practical argument for having two equivalence operators: they do different things.
in reply to irritation makes perls
To understand that, consider the following toy program:
The string "0.0" is different from the string "0", but it's the same numerically. Yes, we could get the same thing with an overloaded equivalence operator and strong typing on all variables -- but to be honest, I find all that extra code to be a nuisance. Also, the dual-purpose nature of scalars allows us to return "0 but true" from a subroutine and let the caller use it as a boolean pass/fail or numeric count value (see Re: True or False? A Quick Reference Guide for details).
my $aval = "0"; # string number zero
my $bval = "0.0"; # also a string zero, spelled differently
for my $compare ( $aval, $bval )
print "Comparing '$compare' to 0...\n";
print "string match\n" if ( $compare eq 0 );
print "number match\n" if ( $compare == 0 );