Robert Sanderson writes:
 > >Referring back to the original email:
 > > > substring="-2:0"  -- the last two characters  // not possible currently
 > > > substring="0:0"   -- the last character
 > >
 > >Shouldn't substring="-2:0" be the last *three* characters, as
 > >substring="-1:0" would be the last two?
 > 
 > Yes.
 > 
 > The fault is mine -- I'm a python programmer at heart, and the python
 > syntax is very similar:
 > 
 > string[start:end]
 > 0 based, where negatives count backwards, end is *not* inclusive and
 > start or end omitted means start or end of the string.
 > 
 > So in python,  "string"[-2:] == "ng"
 > 
 > The other issue with my poor examples was that I was also considering
 > not having 0 at all, and the last character of the string would then be
 > -1, as the first character of the string would be 1, but then thought
 > that changing as little of the current proposal as needed would be
 > better.

Urrgh.  Yikes.

 > >The proposed substring function is really many functions in
 > >disguise. I reckon it is at least five distinct functions:
 > 
 > At least.  But we don't really want 5 or more relation modifiers that
 > each cover a small part of the functionality of substrings, do we?

I sort of agree with Ashley that this is overloaded; but I also agree
with Rob that breaking it into five is not the solution.  I think that
going to a 1-based index is a mistake, and that the consequent benefit
of getting to use 0 to mean "the end" is too small to be worth the
cost.  I think the proposal as it stands is just about right.

 _/|_	 ___________________________________________________________________
/o ) \/  Mike Taylor  <[log in to unmask]>  http://www.miketaylor.org.uk
)_v__/\  "If all our misfortunes were laid in one common heap, whence
	 everyone must take an equal portion, most people would be content
	 to take their own and depart" -- Socrates.