Last modified: 2005-07-23 02:40:29 UTC
As per http://www.w3schools.com/browsers/browsers_stats.asp , wikipedia should display usage statistics for browser, OS, screen resolution, and other useful pieces of research data. Wikipedia is not simply useful to a small audience such as Linux users, Web developers, or gamers. It houses information useful to a diverse audience, including students, hobbiests in several fields, academics, and a range o fexperts looking to share their knowledge. Wikipedia also gets a large number of visits per day, so it supplies a useful and diverse sample to gather information from.
browser statistics are already available at http://www2.knams.wikimedia.org/stats/ screen resolution is a pain to collect and i doubt it's worth the effort...
Those stats are rather useless as a research tool. What is the market growth of the Linux and OSX operating systems, how many people still have Windows 98, what browsers are most common, which is most used on OSX? TBH I'd rather somewhere useful like Google supply this info; but I know a lot of people who use Wikipedia with various browsers and OSes, so it looks like a good information source.
if you'd like to provide some more constructive input than "useless", perhaps we could look at implementing the statistics you want. (describing what exactly you'd like to see would be a good start).
Wikipedia is not your market research tool, and insulting the project isn't likely to make people want to do free work for you.
re Comment 3: The stats there pretty much show which Wikipedia pages are most accessed, how much traffic Wiki gets hourly and daily, who refers to Wikipedia, and what the user agent is. Although the first three are useful in the engineering of Wikipedia's servers and are perhaps interesting to some people, they're on the whole disinteresting applied to any other task or Web site. Wikipedia gets a lot of hits on Harry Potter, perhaps Harry Potter should be given attention and cleaned up as best possible; that doesn't mean some random Web site should design based around Harry Potter, though ;) The user agent stats are pretty bland. Are we mostly inable to tell that Internet Explorer comes from Windows 95 versus Windows XP? Or that Firefox (the second user agent) comes from Windows, Linux, or OSX? Based on that chart, I sure can't tell. Some useful stats would be measurements of which browser and which OS are in use; screen resolution is pretty much accepted as being 800x600 or higher, although having numbers would be nice, but you said it's a pain to collect. Re Comment 4: I don't see why you're being hostile. Look at all the information you can gather from Wikipedia about operating systems, Web browsers, computers, stem cell research, food, sex, and TV shows. There are articles that talk about Linux, Windows, Firefox, Internet Explorer, Web browsers in general; there are comparisons of features in some articles too. For example: - The 'Linux' article: "Its market share for desktops is rapidly growing. According to market research company IDC, 25% of servers and 2.8% of desktop computers were already running Linux in 2002. However, argued advantages ... have spurred a growing number of high-profile cases of mass adoption of Linux by corporations and governments. The Linux market is among the fastest growing and is projected to exceed $35.7 billion by 2008." - The 'Operating system advocacy' article claims "estimated" market shares, although the estimates aren't exactly backed well or updated continuously. - The 'Mozilla Firefox' article estimates the market share of Mozilla Firefox While it's nice that these articles talk about market share of OSes and Web browsers, it's similar to opening an old book you bought; the data was iffy when it was written, and it's not getting any more accurate as time passes. You have to wait for the next release to get updated stats. Even on Wikipedia, the "next release" of the article tends to not be the next day, at least not with pertainence to statistics. Being able to refer to something like 'Special:Usage statistics' would be nice. Looking at the stats link given in Comment 2, the data needed to derive this information from is not all there. Hence, for this purpose the data given is effectively 'useless', about as much as using a single equation to solve for two or more unknown values is 'useless'.
