Description: Use Algorithm::Diff to show the differences as color-coded HTML between two text sequences.
use Algorithm::Diff qw(traverse_sequences);

my ($old, $new) = split /\n--\n/, join "", <DATA>; # get strings

$_ = [split /\s+/, $_] for ($old, $new); # turn them into arrayrefs

traverse_sequences($old, $new, {
  MATCH => sub { print "$old->[shift]\n"},
  DISCARD_A => sub { print "<font color=red>$old->[shift]</font>\n"},
  DISCARD_B => sub { print "<font color=green>$new->[shift,shift]</fon

We start. Let's begin. This is the old text. Then we end.
We start. This is the new text.  And here's some stuff I added. Then w
+e end.
which produces:
<font color=red>Let's</font>
<font color=red>begin.</font>
<font color=red>old</font>
<font color=green>new</font>
<font color=green>And</font>
<font color=green>here's</font>
<font color=green>some</font>
<font color=green>stuff</font>
<font color=green>I</font>
<font color=green>added.</font>
which results in:
We start. Let's begin. This is the old new text. And here's some stuff I added. Then we end.
Replies are listed 'Best First'.
Beware Web Colors!
by chromatic (Archbishop) on Mar 03, 2002 at 18:17 UTC
    Sorry to raise the dead here, but one thing bears repeating. Since a large percentage (25%) of men are red-green colorblind, it's better to pick two highly different colors. I have to read the original and the modified lines to figure out what's changed -- both "change" colors look nearly the same to me.

    Update: I confused the incidence statistic with the inheritance probability. Read a good description of genetics and recessive traits to find out what I was thinking.

      While I agree that colorblindness is important (particularly for you), you are overstating the incidence rates by a factor of 3.

      A good description of the design problem and solutions may be found here.