Jakob Voss writes:
 > So 1-based should not be used because of additional non-standard
 > fancy stuff that nobody will use, but because of it's standard in
 > modern query languages.

Strong argument.

So how about this:

	substring := <start1> [ ':' <length> ]

Where <start1> is a one-based character-count (0 is illegal) and the
:<length> is optional.  If <length> is negative, it counts backwards
from the end of the string.  So:

	1:10 -- First ten characters
	0:10 -- *error*
	1 -- Whole string (from first to end)
	2 -- All of string except the first character
	1:-4 -- All of string except last four characters
	2:-1 -- All of string except first and last characters

 _/|_	 ___________________________________________________________________
/o ) \/  Mike Taylor  <[log in to unmask]>  http://www.miketaylor.org.uk
)_v__/\  "If it were necessary to tolerate in other people everything
	 that one permits in oneself, life would be unbearable" --
	 Georges Courteline.