Thanks for your suggestions.

I switched to file globbing. It seems simpler and is built into perls greater than 5.6.0

I eliminated one of the map lines as the glob now returns the full path. I think this sped things up a bit

I switched to mod ing by three as you suggested. % is probably 100 times more expensive than addition and eliminated but 1/100000 of a second isn't much more than a 1/1000 and the modulo is easier to read.

I left the error checking in. --A few times it was handy to check to see that stat actually succeeded

I left the sort in, I do want the most recent images first. I also suspect that the $a $b variables supplied by sort are quicker than $_[0] and $_1

I left the printed comments and the print "content-type: text/html\n\n"; in as the former help me figure out problems when looking at HTML source and the later <large>:-></large> is needed for the HTML to display.

#!/usr/local/bin/perl 

require 5.6.0;
use warnings;
use strict;
use File::Spec::Functions;
use File::Basename;



my $gHTMLPath='/go/pics/thumbs/';

my $gthumb_dir=catdir('..','go','pics','thumbs');    


##################################
#                                #
# no user servicable parts below # 
#                                #
##################################
if ($gthumb_dir=~m|/|){
    $gthumb_dir.='/';
}elsif ($gthumb_dir=~m|\\|){
    $gthumb_dir.='\\';
}elsif ($gthumb_dir=~m|:|){
    $gthumb_dir.=':'; # we're on a Mac ??? not tested

}else{ 
    die "on unknown OS or couldn't get path at start \n"
}




# FILE: listthumbs.pl  REVISION DATE: 09-13-2001
#
# CGI to list display jpeg, gif & png files in a directory

#####################
#  Declarations     #
#####################

sub get_files($);    # get a list of jpeg files;

sub size_cmp();        # used in sort function compares two values    


sub main();        # 

######################
# Run our program    #
######################

main();

######################
#    Definitions     #
######################

sub size_cmp(){
    my @a_stat=stat($a) or die "couldn't stat\n $a \n$!\n";
    my @b_stat=stat($b) or die "couldn't stat \n $b\n$!\n";

    #7 is size  so
    # sorting by size puts the duplicates next to each other

    #9 is modification time
    # sort reverse order
    return $b_stat[9] <=> $a_stat[9];
}


sub get_files($) {
    my @result=glob("$_[0]*.{jpg,jpeg,png}");

    if (0==@result) {
        die "couldn't glob in get_Files()";
    }
    return @result;
    
}


sub main(){
    my @pics=get_files($gthumb_dir);
    @pics=sort size_cmp @pics;

    for (my $i=0;$i<@pics;$i++){
        $pics[$i]=basename($pics[$i]);
    }

    print "content-type: text/html\n\n";
    print "<!-- start code by $0 -->\n" ;

    my $count=0;

    print '<TABLE>';
    print '<TR>';
    foreach (@pics){
        $count++;
    
        print "<TD width=200 align=left>\n";
        print '<IMG SRC='. $gHTMLPath . $_ .'><BR>'."\n";
        print ;
        print "</TD>\n";
        if (0==$count% 3){
            print "</TR>\n";
            print '<TR>';

        }

    } # end loop
    print "\n</tr>" if 0 != (3 % $count);
    print '</TABLE>';
    print "there are $count thumbs<BR>\n";
    print "<!-- end code by $0 -->\n" ;
[download]

--mandog