I've basically got a 'web bug' on one of my sites to get better activity stats (the log reports I get with my hosting aren't great), but it does slow the site very slightly, sets a cookie and gives data away so I'm considering ditching it.

One of my employers' major clients seems to have been sold some kind of analysis service which uses JavaScript and an embedded Java applet from the provider's server, I spotted it as I wondered why the site set my PC's hard disk going. It causes an irritating pause on low spec PCs and seemingly crashes a few systems. Nice.

Posted by: Matt Round on December 15, 2002 05:26 PM

---

 

I'm not sure exactly what is meant by using referrer logs to map traffic. If it's what I think, the method has a weakness, because users' hosts can sometimes use dynamic IP addresses within
session, and the IP address changes.

Here's what I do every day on my Samuel Johnson site: everyday, I save the server file as a text file using Netscape (oddly, the other browsers won't produce a text file I can parse in excel) and then sort by IP address and time. Frequently I see strings of similar-but-different IP addresses with nearby times which, when -resorted- by access time become a coherent session.

So I'm not sure how this kind of traffic monitoring can be done -easily- without using something like cookies or web bugs.

Posted by: Frank on December 15, 2002 09:05 PM

---

 

I think the reason you don't hear about it is because everyone who understands the issue knows how to work around it. You can set Mozilla to
only load images from the same server as the page and there are other workarounds.

Posted by: Dave on December 16, 2002 08:55 AM

---