>then I call >im = new image(1,1); >im.src = "http://domainname.com/script"+arguments ; > >so that the web server logs the whole line, which gives more infos than >those sent along with a GET http request. > >the logs files of the web server will then be processed by batch programs, > >which will separate the relevant bits of information for doing >statistics. Good idea! The only thing I advice you is to put the data in a separate file. Log files can get tremendously huge and if you have to sort through all thousands of lines to find the few hundred that are generated by the script the program can get quite slow. What I do is calling a Perl script in the same way as you do. It sanitizes my data (browser goes into browser file, referrer goes into referrer file etc) and then sends a transparent image back to the browser. Anyway, take care that your real logfiles don't get too huge. If you wish I can send you a Perl browser detect module I wrote. The coding is not too good, but it's up to date (Omniweb is the last browser added). ppk _________________________________________________________________________ Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com.