Do webbrowser stats lie?

Why webbrowser statistics lie is an article by Björn Jacke which gives some reasons why web-server logs are unreliable, this isn’t news to many of us Jeffrey Golberg had a good rant on the subject 8 years ago.

Unfortunately the new article contains some misleading information, and what could’ve been a useful article is made to look like an anti-IE rant, rather than a use the information wisely rant. For example, it says that you can’t disable images in IE, not only can you in all versions (even pocketIE) there was even the Toggle Images power toy, to make it a simple one click operation long before mozilla was on the scene.

It also has an intriguing little fact that IE splits requests for documents into multiple byte range requests to speed loading, I can’t see any other indication of this, none of my logs (on servers which support range getting) show the behaviour, and it surprises me since all MS installations I’ve seen have been strict on the HTTP/1.1 two connections limit (unless you hack the registry)

The favicon part is also misleading, as IE only requests the icon if the site is boomarked, (or a link stored elsewhere in the filesystem), so the actual number of extra IE connections would not be as high as suggested.

There’s also a strange attack on the users of IE as a browser (suggesting it’s used by the thick, and naive internet user) and that these people spend all their time on web-discussion boards, which would skew the statistics. It ignores the fact it would only skew the statistics of those web-discussion board sites, not any others.

Then there’s a proxy argument, that the more advanced user is more likely to come through a proxy, I find this surprising if the thick users use IE, since the most aggressive users of caches are the big ISPs targetting the naive user. Also, I’d’ve thought the more experienced user is likely to have a better connection and therefore be less likely to be using a proxy.

that is not a joke, this can be proven from httpd log entries

Is a quote from the document, the think that can be proven is that IE users don’t know what they’re looking for and spend a lot more time wandering around a site aimlessly. I can’t actually see how this could be proven. (a person with a proxy cache in use for example would never be seen despite going over the same page 10 times in a row looking for something…) Also if it is the case, it shows that us web-developers have a lot bigger problem than accessible authoring, if we can’t even get our users to the content.

It’s a shame, the conclusions of the article are noble, and completely correct, but the anti-IE rant helps no-one, and devalues the message.