If your page actually _is_ a script, it runs upon opening, i.e. CGI or mod_perl, depending on your Apache setup. You can set it up to run a background job which does whatever, including reading from the Apache environment variables the information you want from the user's system. Once it's launched the background job, you have it print the next HTML page, which has a meta-refresh embedded in it. In the mean time, your background job completes creating its page, and the refresh sends your user to it. That's the long way, to be used only if you have lots of stuff to do. In reality, Perl can do the simple stuff you ask of it so that the dynamic stuff happens between the click and the new page appearing. It's only if you need the pages to stay visible long enough for human eyeballs to read and minds to process that you need to resort to refreshes et al.

Short way:

• User clicks on link http://www.mydomain.org/myperl.cgi
• myperl.cgi reads apache ENV variables,
  generates new page,
  and displays it
• done

or long way:

• User clicks on link myperl.cgi
• myperl.cgi reads Apache ENV variables
• launches background job with system('foo.pl', 'USERIP', '&')
  and displays intermediate page with meta-refresh tag to URL of dynamically generated page
• bg job does whatever else it needs to do
  and generates new page as a file on filesystem
• refresh sends user from intermediate page to new page
• done

A caveat is that the referrer and user variables that Apache maintains are often only those of proxies, so you will often get the ip's of AOL proxy machines as opposed to the end-user ip.

Some links:

http://httpd.apache.org/docs/env.html
http://www.apachefreaks.com/apache2/env.html
http://cgi.resourceindex.com/Documentation/Environment_Variables/