Webalizer GroupAgent config – round 2

A helpful reader pointer out that Opera has borked it’s user agent string, which unfortunately means that webalizer can no longer distinguish between Opera 9 & Opera 10. It is a bit of a shame, but not something I can do anything about.

I also took out the “Version/4″ detection for Safari, because it is a little stinky – any browser could put “Version/” in their UA string (which happens to be exactly what Opera has done). So here is my updated webalizer GroupAgent config, for use with version 2.01 of webalizer (which is the Debian stable version):

GroupAgent      Firefox/3               Firefox 3
HideAgent       Firefox/3
GroupAgent      Firefox/2               Firefox 2
HideAgent       Firefox/2
GroupAgent      Firefox/1               Firefox 1
HideAgent       Firefox/1
GroupAgent      Firefox                 Firefox
HideAgent       Firefox
GroupAgent      Chrome/4                Chrome 4
HideAgent       Chrome/4
GroupAgent      Chrome/3                Chrome 3
HideAgent       Chrome/3
GroupAgent      Chrome/2                Chrome 2
HideAgent       Chrome/2
GroupAgent      Chrome/1                Chrome 1
HideAgent       Chrome/1
GroupAgent      Chrome/0                Chrome 0
HideAgent       Chrome/0
GroupAgent      Chrome                  Chrome
HideAgent       Chrome
GroupAgent      iPhone                  iPhone
HideAgent       iPhone
GroupAgent      Safari                  Safari
HideAgent       Safari
# Instead of using "Opera/10", Opera stuck with "Opera/9.80" and added "Version/10" to their UA
# string - this means we cannot distinguish between Opera 9 and 10 using this version of webalizer
GroupAgent      Opera/8                 Opera 8
HideAgent       Opera/8
GroupAgent      Opera/7                 Opera 7
HideAgent       Opera/7
GroupAgent      Opera/6                 Opera 6
HideAgent       Opera/6
GroupAgent      Opera/5                 Opera 5
HideAgent       Opera/5
GroupAgent      Opera                   Opera
HideAgent       Opera
# TODO: When Debian finally upgrade webalizer, use "MSIE x"
GroupAgent      8.0                     Internet Explorer 8
HideAgent       8.0
GroupAgent      7.0                     Internet Explorer 7
HideAgent       7.0
GroupAgent      6.0                     Internet Explorer 6
HideAgent       6.0
GroupAgent      MSIE                    Internet Explorer
HideAgent       MSIE

Webalizer GroupAgent config

I’ve wanted Webalizer to group the UA strings for the major players in the browser market for a while now, and finally sat down this evening to poke around with things.

I don’t believe in “grouping” everything, because it just means I end up with a massive “grouping” list to maintain, and I just know I won’t be able to do that. So I’m just concentrating on the browsers who hold most market share for the browser demographic my client’s site’s are attracting. Besides, browsers with a small market share (I’m taking like less than a couple of percent here) aren’t really worth grouping.

As a developer, I’m interested in what versions of browsers visitors are using, but not that interested…and not so interested that I want to confuse the clients who are going to view the stats for their site. So I’m grouping by major version number, making sure to define a “catchall” style group for version numbers above/below any of the groups I have defined. This means that any new browsers versions released before I get a chance to update the list are still grouped.

The only problem I ran into was trying to define a GroupAgent string for IE. I wanted to match MSIE 8, MSIE 7, MSIE 6, etc. but unfortunately space characters delimit the GroupAgent string. Which is a bit of a pain. I noticed that the sample config file on the webalizer site used quotes to surround the string and encompass the space, but that didn’t work for me. Perhaps because Debian stable is still using version 2.01.10 from 2002. Gotta love Debian. I’ll revisit if they ever update the package.

Anyway, that means that I’m using just version numbers to group IE. Which isn’t great as we could potentially get some false positives…but hopefully not many as we’re matching all the other big players beforehand.

GroupAgent	Firefox/3		Firefox 3
HideAgent	Firefox/3
GroupAgent	Firefox/2		Firefox 2
HideAgent	Firefox/2
GroupAgent	Firefox/1		Firefox 1
HideAgent	Firefox/1
GroupAgent	Firefox			Firefox
HideAgent	Firefox
GroupAgent	Chrome/4		Chrome 4
HideAgent	Chrome/4
GroupAgent	Chrome/3		Chrome 3
HideAgent	Chrome/3
GroupAgent	Chrome/2		Chrome 2
HideAgent	Chrome/2
GroupAgent	Chrome/1		Chrome 1
HideAgent	Chrome/1
GroupAgent	Chrome/0		Chrome 0
HideAgent	Chrome/0
GroupAgent	Chrome			Chrome
HideAgent	Chrome
GroupAgent	iPhone			iPhone
HideAgent	iPhone
GroupAgent	Version/4		Safari 4
HideAgent	Version/4
GroupAgent	Version/3		Safari 3
HideAgent	Version/3
GroupAgent	Safari			Safari
HideAgent	Safari
GroupAgent	Opera/10		Opera 10
HideAgent	Opera/10
GroupAgent	Opera/9			Opera 9
HideAgent	Opera/9
GroupAgent	Opera/8			Opera 8
HideAgent	Opera/8
GroupAgent	Opera/7			Opera 7
HideAgent	Opera/7
GroupAgent	Opera/6			Opera 6
HideAgent	Opera/6
GroupAgent	Opera/5			Opera 5
HideAgent	Opera/5
GroupAgent	Opera			Opera
HideAgent	Opera
# TODO: When Debian finally upgrade webalizer, use "MSIE x"
GroupAgent	8.0			Internet Explorer 8
HideAgent	8.0
GroupAgent	7.0			Internet Explorer 7
HideAgent	7.0
GroupAgent	6.0			Internet Explorer 6
HideAgent	6.0
GroupAgent	MSIE			Internet Explorer
HideAgent	MSIE