From: Michal 'hramrach' Suchanek <hramrach@...>
Date: 2004-10-02T05:23:12+09:00
Subject: Re: Range syntax theory

Interesting discussion

I will just add my little bit :)

On Fri, Oct 01, 2004 at 11:17:11AM +0900, Zach Dennis wrote:
> Gavin Sinclair wrote:
> 
> >
> > 
> >
> >
> >I'll try to come up with a good reasoning.  Pretend I'm designing a 
> >language.
> >
> >x..y is known to programmers of some other languages as representing a
> >"range"  (i.e. not necessarily what we'd call a Range, but something close
> >to the concept).  I decide that this concept will have its own
> >representation as a Range object.  So:
> >
> > 2..6        # -> Range object
> >
> >So far so good.  And familiar to some, but by no means all, programmers of
> >other languages.
It is also used that way in 'normal' or mathematical writing, mostly with
two but also with arbitrary number of dots.
> > 
> >
> >Now I decide, for the good reasons put forth by others, that end-exlusive
> >ranges are useful enough to be treated explicitly as well.  I go hunting
> >for a way to represent them, and settle on

Or we need another range different from the 'normal'. Using just one
more dot makes it easy to type, using an operator that has the third
character different would be easier to remember. Actually I have to try
the ranges in irb when I am not using them for a while to make sure I
get the right one.
As I am familiar with the <2,6>, <2,6), (or '[' instead of '<')
mathematical notation I would use something that resembles that, the
round braces are easy to remember in that they 'avoid' the extreme
value.  but 1[-]6 and 1[-)6 is quite hard to type and the braces would
be hard to parse as they are already used differently..

As the mathematical notation is used even in the last years of elementary
school around here I would consider it very widespread. But I have no
knowledge of the rest of the world.

My suggestion would be to use some cleverly English named functions for
clarity :)

Thanks

Michal Suchanek