It would output "\`\n" if it acted upon the ordinal value.

But since ++ doesn't act on ordinal values, why should --? Furthermore, what should:

    my $a = "a";
    $a --;
    print $a;
[download]

print on an EBCDIC platform?

Abigail

It depends upon which EBCDIC variant the platform was running but IIRC with Western European platforms it would be a speech mark. But I take your point.

But since the sequence is a b c d e f g h i etc. why not apply magic to the -- operation and use the numbers as Base 26.

So a decrement would of ab would become:

'ab' = (1*(25*1))+(2*(25*0)) 

'ab'-- = (1*(25*1))+((2*(25*0))-(1*(25*0))) 
       = (1*(25*1))+((1*(25*0))
       = 'aa'
[download]

Then 'a'-- would always = -1 and decrements become OK with a predictable behaviour.

Sorry, totally wrong, it should be:

'ab' = (1*(26**1)) + (2*(26**0))

'ab' -- = (1*(26**1)) + ((2*(26**0)) - (1*(26*0)))

        = (1*(26**1)) + (1*(26**0))

        = 'aa'
[download]

But ideally a string containing only alphanumerics with an alphabet in upper and lower case could be handled as a Base 62 number which would also allow decrement, because the question with the above is how does it handle numerics and upper case. Base 62 would solve that.