Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!sdd.hp.com!elroy.jpl.nasa.gov!decwrl!ucbvax!GARGOYLE.UCHICAGO.EDU!goer%sophist
From: goer%sophist@GARGOYLE.UCHICAGO.EDU (Richard Goerwitz)
Newsgroups: comp.lang.icon
Subject: RE:  string indexing peculiarity
Message-ID: <9102261550.AA09774@sophist>
Date: 26 Feb 91 15:50:05 GMT
Sender: daemon@ucbvax.BERKELEY.EDU
Distribution: inet
Organization: The Internet
Lines: 35


	Recently I found out the hard way about something I consider to be
	an anomaly in string indexing.  I consider it an anomaly  because
	it was, at least for me, totally unexpected.  To wit:

		s := "-D"
		x := (s[3:0] | "default")

	"Obviously" accessing the 3rd character in a 2-character string
	should fail and x would get the string "default".  It turns out,
	non-intuitively enough, that I get an empty string, exactly as
	if I had tried s[1:1].

Icon's string indexing methods are definitely non-intuitive for people
who acquire it as a second, third, etc. language (which means virtually
everyone).  I guess the idea that people need to catch on to is that
subscripts deal with string positions, and not the index of characters
within strings.  What throws people off is that s[3] would work as in
some other languages (producing the third character in s).  In reality,
though, this is shorthand in Icon for s[3+:1], i.e. s[3:4].

Of course, if we want to talk about unintuitive schemes, let's not for-
get languages where s[0] indexes the _first_ array element in s.

        s := "-D"
	s ? { x := move(3) | "default" }

        s := "-D"
        s ? { x := (tab(-3), tab(0)) | "default" }

        s := "-D"
        x := ("" ~== s[3:0]) | "default"


-Richard (goer@sophist.uchicago.edu)