Posted by J.O. Aho on 02/25/07 13:39

Red wrote:
> Im pretty interested in seeing some reliable data on browser market
> share. It has no effect on how I design my pages, but its useful stuff
> to know.
>
> I've found many sites that publish this, but the data varies widely. For
> example I've recetntly seen firefox having anywhere from between 13 and
> 39%. Im guessing that a lot of sites that publish these figures use
> their own server logs, so may already be biased to one type of browser
> (a web design orientated site may show higher firefox figures than say a
> site like zdnet.)

Usually they have a number of sites logs to check which browsers been used,
but then the result depends much on which sites they use, say a company that
has www.microsoft.com, support.microsoft.com as the main sites would get a
really high value on MSIE, while if another company had www.mozilla.org and
mozdev.org as the main sites would get a high result on gecko based browsers.

There aren't any good way to measure which browser is the most used or how
large share it has, logfiles has the disadvantage that those who often visits
the sites generates a lot of hits for their browser, while people who does
visit less those sites will generate a low amount of hits for their browser.
The same for cookies and as people are extremely careful with allowing 3rd
party cookies, those market shares based on such methods are quite lousy.

If you manage to design your site so it looks nice in MSIE a Gecko based
browser and a khtml based browser, then you will cover a good part of the most
common browser out there. There will always be some browsers that won't work well.


--

//Aho

[Back to original message]