Latest Stats as at 2/9/2008. These figures cover the 16-day period 16/8/2008 to 31/8/2008. Vista Market Share ================== For period 16th to 31st August 2008, figures based on visits:- XP 71.6 (was 71.8) Vista 15.9 (was 15.8) Other Windows 4.0 (was 4.5, 5.6) Mac 6.3 (was 5.8) Linux 1.4 (same) Others 0.7 (was 0.6) The decline of older Windows versions continues. Growth in Vista is essentially at the expense of earlier Windows versions, but it's all $$$ to M$. In the "others" category, the big three are Nokia, DoCoMo and iPhone. I am watching these three carefully. Browser Breakup =============== For period 16-31 August: IE 70.5 (was 71.4 and 72.0) FF 21.7 (was 21.3) Safari 4.8 (same) Opera 1.0 (was 1.1) All Others 1.9 (was 1.7) These figures are based on IP addresses. IE is loosing market share to Firefox gradually. Firefox 3 Takeup ================ These percentage figures are based on visits, and only from an OS where the breakup is clear. For example FF3 is not available for Windows 98, so Win98 is excluded. FF1 3.7 (was 3.8) FF2 55.9 (was 61.2, 65.1, 73.1) FF3 40.4 (was 35.1, 30.5) The big takeup since early June continues. I'm not 100% sure about these latest figures as the FF3 result on XP seem a bit high to me, but if correct they are a huge rebuff to M$. The market loves Firefox but does not like Vista in the same way. Marketers probably cannot see that one is a browser and the other is an operating system. The latest Ubuntu DVD distros have come with a Beta of FF3 not FF2. They took a punt on FF3 which has proven correct, and users can painlessly upgrade to later versions. Search Engines Share ==================== August 16-31. Google 87.2 (was 84.4, 85.7, 86.6, 86.2) Yahoo 7.3 (was 7.3, 8.0, 7.5, 7.8) Microsoft 2.0 (was 3.3, 2.6, 2.1, 2.3) All others 3.5 (was 5.0, 5.0, 3.7, 3.7) This is the lowest figure ever recorded for M$, expect lots of bad press on this, as it translates into less advertising revenue. Already there have been articles saying the proposed purchase of Yahoo was foolish and timewasting. Neither M$ nor Yahoo seem to be able to figure out what Google is doing right. Still, nothing compares with sensis - if you search for "Bondi Beach Motel" one of the results on the first page is "Things to Do in Bourke". Actually there are no motels in Bondi and nothing to do in Bourke. No sign of Telstra's masterpiece "sensis", but a couple of searches from WebWombat (it's dreadful) and one from Anzwers. Linux Distro Analysis ===================== I am still experimenting with Linux distro analysis, but the results are far from satisfactory, especially since so many Linux operating systems are identified as just i686 or 64-bit, without the distro name. The Linux version of the Asus EEE-PC from Xandros is not visible. Hacker Analysis =============== Every day it goes on and on. Robots and bogus humans with weird agent strings looking for weaknesses in PHP pages, looking for various blogs and bulletin board and looking for CMS software, and especially looking for PHPbb, to see if they can break in. Many are using the lib-www-perl library, and for some unknown reason, many of them are using Opera as their browser. They keep visiting even after they are denied access and are given 403 errors instead. Most 404 errors are caused by these people, these errors far out number your own mistakes on the website, or ordinary people misstyping things. The sooner I automate the entire banning process the better. The proposal is for the database to write a new .htaccess file which would then be loaded onto the server say once a week. All offenders would be banned for a month from the date of their last visit - some would thus be permanently excluded. They are well under 1% of the total hits. ================================================================