I check on a lot fewer browser versions than I used to, now that I always validate my code. It's amazing what a difference that makes. Still, I do check out the first template pages pretty closely.
On Windows I check IE 5.5 and IE 6, Netscape 4.7 and Netscape 6.2, Opera 6 and Moz (well, until it got in a fight with Netscape 6 and blew up!). Also, for the template pages at least, I check AOL 6.0. AOL 7.0 is so well behaved I don't worry about it unless the site has very complex functionality.
On Mac I check Netscape 4, IE 5, Opera 6 (because that's all we have installed). But as I said, lately I've had fewer and fewer cross-browser snarls, so after the templates check out, I tend to lay back a lot more than I used to.
Latest browser statistics figs for July 2002 according to W3Schools web site...
Browser Version July 2002
Internet Explorer 6.x 39%
Internet Explorer 5.x 51%
Internet Explorer 4.x 2%
Netscape 4.x 3%
Other Netscape compatible 1%
Browsers that count for less than 0.5% are not listed.
| the last 14hrs here:|
|Page Views ||Percent ||Browser ||Chart || || |
|22390 ||39.82 % ||MSIE v6.x || || || |
|16994 ||30.22 % ||MSIE v5.x || || || |
|5202 ||9.25 % ||Opera v6 || || || |
|4572 ||8.13 % ||Netscape v5+.X (gecko) || || || |
|3721 ||6.62 % ||Unknown Browsers || || || |
|1552 ||2.76 % ||Opera v5 || || || |
|504 ||0.90 % ||Netscape v4.X || || || |
|394 ||0.70 % ||Konqueror v2.0 || || || |
|281 ||0.50 % ||Konqueror v3.0.0-10 || || || |
|184 ||0.33 % ||MSIE v4.x || || || |
|149 ||0.26 % ||Opera v4 || || || |
|115 ||0.20 % ||Konqueror v2.2.1 || || || |
|110 ||0.20 % ||Lynx || || || |
|22 ||0.04 % ||Netscape v3.X || || || |
|11 ||0.02 % ||Konqueror v3 || || || |
|10 ||0.02 % ||Konqueror v2.2-11 || || || |
|6 ||0.01 % ||Opera v7 || || || |
|3 ||0.01 % ||Konqueror v2.1.1 || || || |
|3 ||0.01 % ||MSIE v3.x || || || |
|2 ||0.00 % ||Opera v3 || || || |
|2 ||0.00 % ||WebTV || || || |
|1 ||0.00 % ||Konqueror v2.1 || || || |
|1 ||0.00 % ||Konqueror v2.2.2-2 || || || |
|1 ||0.00 % ||Konqueror v3.0 || || || |
|1 ||0.00 % ||MSIE v2.x || || || |
Interesting chart, Brett, but I don't think it is representative of general browser use on the web. There are too many geeks here, Opera for instance would never get that percentage on a 'normal' site.
As always, it depends heavily on who the target group of the site are. Techies vs. people who use IE because it is installed automatically for example.
Who's the poor soul on Internet Explorer 2.x!
Okay Brett, who's got the Opera v.7? ;)
I want it! lol!
On one ot the large commercial stat sites, the one that W3Schools gets their reports from, Opera is nearing the 1% mark ( doubled in the last two months) - this percentage is likely under reported due to Opera's default ID settings. A number of indicators point to Opera's accelerating market penetration.
And not just Opera, how many people out there in the real world are actually using Konqi to browse the web?! If you add up Brett's stat there will be around 1.5% ... hmm all these Linux/xBSD geeks :)
I notice the stats show page views. This is not always terribly reliable, because it's always possible that some of the low ratings are produced by people logging on with their choice of browser, seeing an unnavigable mess and moving on (unlikely on this particular site, I know, but still).
For example, www.noncompatible.com shows the following figures:
200 page views with Browser X;
150 page views with Browser Y.
That seems as if Browser X has a bigger market share than Browser Y -- but does it? If the site is unusable in Browser Y, then in fact the stats may be broken down like this:
20 visits with Browser X, each visitor viewing 10 pages;
150 visits with Browser Y, each visitor logging on and then logging straight back out again.
Now which browser has the largest market share?
Of course, that's deliberately exaggerated, but shows how "page view" stats can sometimes skew the data. Unfortunately, I don't know of a reliable method of counting the number of actual visitors as opposed to individual page views.
Page views must remain a rough guide only.
>who's got the Opera v.7
That's pretty good for Linux. I've not parsed Opera and netscape strings for Linux ids. Plus, most of the *nix browsers allow you to do your own agent string. I'm all but positive atleast 5 to 8% of those reporting as IE are actually something else.
>seeing an unnavigable mess and moving on
Exactly. Worth remembering.
"how many people out there in the real world are actually using Konqi to browse the web?!"
It comes bundled with Suse Lunix 8.0
I hope this question isn't too tangential to this thread - but we've been having a few discussions here regarding how many visually disabled people actually surf the web with 'talking browsers'.
What user agents would I be looking for in logs to quantify this traffic - I assume I'm looking for text browsers like lynx - any other suggestions??
MS nirator seams to do fairly well reading text from internet explorer. My guess is that sight impaired peopel will be using standard browsers with either software to read text or possibly browser plugins???
Chris_D: according to the Royal National Institute for the Blind, IE is the most popular browser among visually impaired users - just like "normal" people ;)
Something the w3c appears to be unaware of is that folks don't have aural browsers. They have screen readers that work at the OS level. If you only had an aural browser how would you navigate the OS to start it?
Download the trial version of Jaws (win32) to try it for yourself.