suspicious Accuracy
|
| Author | Content |
|---|---|
| Abe Aug 12, 2005 2:48 PM EDT |
That means FireFox lost 0.64%. At the same time IE gained 0.64%. Woo, stats with this accuracy must be suspect. Made up that is. |
| TxtEdMacs Aug 12, 2005 4:25 PM EDT |
Odd, I just happened to look at a sites statistics today where I would expect very few developers would visit other than those having strong political interests. Firefox 39% Mozilla 2% Safari 8% with IE (several versions) total in the high 40's. What was even more surprising, was the OS distribution: Linux/Unix 2% Mac 15% XP 61% 2000 15% (?) 98 less, but significant and other version of Windows in the 3 - 1 % range. Hence, it is not our guys that are flooding some user type sites with FF. Good point about the exact matching of loss/win percentages. The measurement accuracy is not believable. Moreover the studies I have viewed FF and IE status are not well correlated. |
| hkwint Aug 13, 2005 4:07 AM EDT |
Another interesting party researching use of Firefox is Xitimonitor. It's interesting this (Xitimonitor) site mentioned 12,78% Firefox usage in (whole) North America on a sunday in July ('snapshot'). Maybe because this site is French? It's still interesting anyway, and it's a pity this article doesn't mention the 'total number of hits researched' as well as in July as in August. Maybe many people are on holiday? On which day and which time did they measure? (Xitimonitor showed in the weekend Firefox market share is about 2% higher than on weekdays). Without these knowledge, this statistics are a bit void in my opinion. The Firefox-counter is at 79Million by the way (that's good, now waiting for 100Million). |
| Abe Aug 13, 2005 6:14 AM EDT |
Like I always said, all these measurements by various monitoring companies are for the birds. The only way we will be able to get a pretty good idea about browser percentages is by having many of the Internet sites publish their stats. We don't need every site, we just need google, yahoo, for search engines sites, PCMag, Computerworld, ZDnet for the lousy commercial magazines, Slashdot and Lxer, LinuxToday and such for Linux sites. The larger the sample the more accurate the results will be. All they have to do is publish the raw data and we will take care of the rest. Methods and algorithms should be openly published by the analysts. Many sites have polls for various subjects and they publish their data. We keep hearing that IE is dominant. With all the sites that are pro-MS, I wonder why they don't publish their brwoser-ID data?!!! Could it be that they don't want to show the trend for some reason? Could it be they want money for it? Well MS is the only one who would pay, and it must not be of their interest to see the trend. Therefore, we will have to continue reading this kind of cra**y articles. |
| TxtEdMacs Aug 13, 2005 6:59 AM EDT |
Abe
Quoting: "... [just] publish the raw data and we will take care of the rest ..." It's not quite that simple. When you really look at the raw data you will find too that calculated one way, e.g. by unique viewers differs markedly from those employing page views. Then there is the issue of bots, spiders, and other weird digital beasts. Finally there are those that leave a seemingly senseless trail. I have made inquires at various sites where I have seen usage data. One that made absolutely no sense claimed most of the Windows OS usage was attributable to what seemed to be "Longhorn", perhaps its needless to say that I received no response. Most others also failed to respond when I asked how they calculated their numbers and what techniques they employed to separate valid user data from those with significant ambiguities. However, I did get one response that just re-enforced my view that the accurate measurement of browser and OS by any site is very problematic. I have code that I wrote that attempts to compile usage statistics on the basis of unique users by browser/OS combinations. Moreover, I am trying to learn how I can publish/deposit the source where is would be freely available under a GPL license. |
| MESMERIC Aug 14, 2005 10:26 PM EDT |
I look after a website
and despite me often advertising on strongly Linux-oriented sites.
and my own hits (sometimes I forget to hide the UserAgent) I get something like IE 88% Firefox 6.5% Safari 1% something like that. Windows takes 93% but what is curious is the split: XP 65% NT 1.5% Me 1.5% 2000 21% !! 98 5% !! A lot of companies are still using Windows 2000, and a lot of homes still use Windows 98. Mac takes about 2% Linux takes about 1.5% I fear that if it wasn't my advertising on linux channels. the percentage drop would decrease considerably. What Linux needs is strong media advertising. Firefox had that. |
Posting in this forum is limited to members of the group: [ForumMods, SITEADMINS, MEMBERS.]
Becoming a member of LXer is easy and free. Join Us!