From: Chad Perrin <perrin@...>
Date: 2007-07-20T00:17:47+09:00
Subject: Re: The Pebble in the Ruby Shoe

On Thu, Jul 19, 2007 at 06:13:56PM +0900, Adrian Howard wrote:
> (taken off-list since wandering way OT :-)

Understandable.


> 
> On 17 Jul 2007, at 20:03, Chad Perrin wrote:
> 
> >On Tue, Jul 17, 2007 at 10:43:35PM +0900, Adrian Howard wrote:
> >>
> >>Perl's misfeature is pretty different - it's a global rather than a
> >>per-array thing. Being able to say "this particular array has indices
> >>-10 -> 10 is considerably more useful than saying "every array starts
> >>at -10" :-)
> >
> >News to me.  Would you point me at more information on the topic,  
> >please?
> 
> In the perl world we have the $[ variable (see perldoc perlvar for  
> gory details) that resets the index of the first element. As the  
> documentation says "Its use is highly discouraged.". I'd imagine that  
> it's primary motivation for inclusion was to make compatability with  
> awk easier in early perl's (which also has 1 indexing).

I didn't know about the likely awk connection in the existence of $[, and
didn't know early Perl used 1-indexing.  I guess I learn something new
every day.

When I said something was news to me, however, I meant specifically your
statement that the "fix" was global rather than per-array.  I somehow
read that as "it applies globally, and you can't make it apply per
array", rather than how I now realize you meant it.  When you first
mentioned that, I thought you were saying there was something like
perhaps a pragma that changed the way Perl handled arrays, or something
like that.

. . . but yeah, I see now you were just referring to the fact that $[ is
a global variable.  Of course, there's always the local hack:

  {
    local $[ = 1;
    # do stuff
  }

> 
> For example:
> 
>     my @array = ( 1 .. 10 );
> 
>     $[ = 0;
>     print @array[5], "\n";
> 
>     $[ = 1;
>     print @array[5], "\n";
> 
>     $[ = 2;
>     print @array[5], "\n";
> 
> would produce
> 
>     6
>     5
>     4
> 
> Spooky action at a distance. Scary.

A little weird, to be sure -- but awfully flexible if you need it, I
suppose.  Probably the most disturbing thing about it for me is the fact
that, for each of the changes to $[ in your code snippet, that changes
the behavior of *all* arrays.  Downright terrifying.


> 
> Fortran (or those versions that support it anyway) allow you to  
> specify the indices of a particular array - so for example you could  
> have
> 
>       real temperature(-100:100)
> 
> to make an array of real values with indices from -100 to 100. So you  
> can make the indices of your array match the indices of your domain -  
> if that would make your life easier :-)

I don't know any Fortran, so the talk of Fortran and arbitrary array
ranges is entirely new to me -- and fascinating.  The various differences
of feature sets between languages really teaches some interesting
principles, I find.


> 
> See <http://www.stanford.edu/class/me200c/tutorial_77/10_arrays.html>  
> for some examples.

Thanks.

-- 
CCD CopyWrite Chad Perrin [ http://ccd.apotheon.org ]
Dr. Ron Paul: "Liberty has meaning only if we still believe in it when
terrible things happen and a false government security blanket beckons."