nanog mailing list archives

Quick fix


From: Bandy Rush <bandyrush () altavista com>
Date: 1 Mar 2000 20:11:04 -0800


I noticed that a number of individuals who post regularly to the NANOG mailing list have keyboards with malfunctioning 
shift keys.  Since working keyboards are so expensive, far beyond the financial means of most poorly-paid technical 
workers, I humbly offer the following Perl program, which you can use to filter your NANOG e-mail to correct 
capitalization errors (procmail should do the trick for that).  For example:

    i abhor cross-posting.  but sometimes ...

Gets converted to:

    I abhor cross-posting.  But sometimes ... 

It's not perfect, but it catches most occurrances.  If you have a slow computer, you probably want to disable the 
dictionary.  You can do this by setting @DICT_FILES=().  As is, it is configured to search for the word list in the 
default locations in Linux and Solaris.

And oh yes, it's a joke.  ;)

#!/usr/bin/perl -w  
 
# use good Perl form
use strict;
use English;
 
# location of plain-text word list
my @DICT_FILES = ("/usr/dict/words", "/usr/share/lib/dict/words");
 
# slurp in all input  
undef $INPUT_RECORD_SEPARATOR;
my $input = <>;
exit unless (defined($input));
 
# Fix end of paragraph
$input =~ s/([a-z])(\n\n.)/$1.$2/g;
 
# Fix start of sentence
$input =~ s/(\n\n+)([a-z])/$1\u$2/g;
$input =~ s/(\.\s+)([a-z])/$1\u$2/g;
 
# Fix self-referental problems
$input =~ s/(\b)i(\b)/$1I$2/g;
$input =~ s/(\b)randy(\b)/$1Randy$2/g;
$input =~ s/(\b)bush(\b)/$1Bush$2/g;
 
# Fix words that should be capitalized (based on the dictionary)
$INPUT_RECORD_SEPARATOR = "\n";
foreach my $dict_file (@DICT_FILES) {
    if (open(DICT, "< $dict_file")) {
        while (defined (my $uc_word=<DICT>)) {
            if ($uc_word =~ /[A-Z]/) {
                chomp($uc_word);
                my $lc_word = lc($uc_word);
                $input =~ s/(\b)$lc_word(\b)/$1$uc_word$2/g;
            }
        }
        last;
    }
}
 
# output our fixed text
print $input;
 
# The author releases this software into the public domain.  


--------------------------------------
AltaVista     Smart is Beautiful     http://www.altavista.com

Raging Bull?  Sleeping Bear?  Live stock quotes at AltaVista Live!
http://money.altavista.com

--------------------------------------



Current thread: