I used LiveHTTPHeaders on my Firefox and: yes, in the end the
javascript is request a 1 pixel image.
The request I recorded on the site http://travelshop.autostadt.de is:
GET
/__utm.gif?utmwv=1&utmn=1142882144&utmsr=1920x1200&utmsc=24-bit&utmul=en-us &utmje=1&utmfl=7.0%20r25&utmdt=Autostadt%20%3E%20TravelShop&utmhn=travelsho p.autostadt.de&utmp=/Startseite&utmac=UA-xxxxx-x&utmcc=__utma%3D5829923.131 9782024.1132062099.1132223841.1132230211.8%3B+__utmb%3D5829923%3B+__utmc%3D 5829923%3B+__utmz%3D5829923.1132062099.1.1.utmccn%3D%28direct%29%7Cutmcsr%3 D%28direct%29%7Cutmcmd%3D%28none%29%3B
HTTP/1.1
(Hope most of the numbers are just cookie information and don't reveal
too much about myself :-) )
The response is a gif containing 35 bytes.
So I would assume you can also add a noscript tag to the javascript
code. The problem is, that it just counts the hits on the site. Since
the code has no access to the cookie, it can not track visitors, nor
any browser related stuff (including resolution, etc.).
But maybe if you record some path like /noscript you at least get an
overview over spiders, etc. However, your visitors statistic may crap
up.
The code should read something like:
(your tracking JS code)</script>
<noscript>
<img src="http://www.google-analytics.com/__utm.gif?utmdt=[Title of
page]&utmhn=[your domain name]&utmp=[path to
record]&utmac=UA-xxxxx-x"/>
</noscript>
I have not tested this (since I don't want to spoil up my statistics).
But maybe this is an idea of also tracking RSS feeds?
-t
