(Message perl/tiefile:352)
/home/mjd/bin/mailpager 352
Return-Path: mjd@plover.com
Delivery-Date: Fri Apr 04 15:45:54 2003
Return-Path: <mjd@plover.com>
Delivered-To: mjd-filter-deliver2@plover.com
Received: (qmail 24716 invoked by uid 119); 4 Apr 2003 15:45:54 -0000
Delivered-To: mjd-filter@plover.com
Received: (qmail 24690 invoked by uid 119); 4 Apr 2003 15:45:50 -0000
Delivered-To: mjd@plover.com
Received: (qmail 24679 invoked by uid 119); 4 Apr 2003 15:45:50 -0000
Message-ID: <20030404154550.24678.qmail@plover.com>
To: bug-Tie-File@rt.cpan.org
cc: mjd@plover.com
Subject: Re: [cpan #2276] POD typo in Tie-File-0.93 
In-reply-to: Your message of "Mon, 24 Mar 2003 09:48:49 EST."
             <rt-2276-6150.1.59229200746395@cpan.org> 
Date: Fri, 04 Apr 2003 10:45:50 -0500
From: Mark Jason Dominus <mjd@plover.com>


> [mjd@plover.com - Sun Mar 23 18:50:36 2003]:
> > I guess it depends on whether you think the first line of the file is
> > line 0 or line 1.
> 
> Yes, of course. And to me, in plain English (or in my case Swedish...),
> the first line of something is line # 1. Maybe it's just me... ;-)

You will note that the documentation does not say "display 42nd line
of the file".  It says "Display line 42 of the file".  This was not an
accident.  The 42nd line of the file is line 41, just as the first
line of the file is line 0 and the third line of the file is line 2.

Of course, one could argue on philological grounds that "third" must
necessarily mean "item #3", regardless of where the numbering starts.
But this is undesirable.  It would introduce a discontinuity between
the cardinal and the ordinal numerals in cases where items are
numbered starting with 0, as is common in computer applications,
because the ordinal numerals 'first' and 'second' are *not* derived
from the cardinal numbers 'one' and 'two'.  The first ('principal')
item would be #0, the second ('following') item would be #1, and then
the third ('number three') item would be #3.  Then you would have no
way to refer to item #2.  If for some reason the iterms were numbered
starting with -57, the problem would be even bigger since then there
would be no word for items -55 through 2.

In any case, it doesn't seem clear that making the change you
suggested would lead to a net decrease of confusion in people reading
the manual, so I think I probably won't do it.

Best regards,
-D.

