CGI applications often need to generate output depending on input/parameters. One way of achieving this is to use a dispatch table. The following highly contrived example demonstrates how a dispatch table could be implemented (not using CGI or modules to keep it simple).

#!/usr/bin/perl

use warnings;
use strict;
use Data::Dumper;

my $params = {
  param1 => 1,
  param2 => 2,
};

my ($case, $args) = process_params($params);

my %dispatch_table = (
  case1   => \&case1,
  case2   => \&case2,
  default => \&default,
);

my $page_header = page_header();
my $body        = $dispatch_table{$case}->($args);
my $page_footer = page_footer();

print $page_header, $body, $page_footer;

sub process_params{
  
  my $params = shift;
  my $case = q{default};
  my $args = {};
  
  if ($params->{q{param1}} eq 1){
    $case = q{case1};
  }
  elsif ($params->(q{param2}) eq 2){
    $case = q{case2};
  }
  else{
    $case = q{default};
  }
  
  return $case, $args;
  
}

sub case1 {
  my $args = shift;
  my $output = qq{case one\n};
  
  return $output;
}

sub case2 {
  my $args = shift;
  my $output = qq{case two\n};
  
  return $output;
}

sub default {
  my $args = shift;
  my $output = qq{default\n};
  
  return $output;
}

sub page_header {
  my $ph = qq{page header\n};
  return $ph;
}

sub page_footer {
  my $pf = q{page footer};
  return $pf;
}
[download]

The "case" subs could call subs/methods in different (application) modules.

If this approach is suitable then something like CGI::Application would be worth looking at. It can help look after and simplify all this and much more. Combined with, say, HTML::Template, CGI applications, imo, become a whole lot easier and reusable. A little extra effort up front but well worth it.