Slashdot Mirror
Mozilla Gains on Internet Explorer
Alizarin Erythrosin writes "PCWorld is reporting that Internet Explorer's share of the browser market has dropped 1%, the 'first noticeable decline since WebSideStory began tracking the browser market in late 1999.' With all the exploits and security holes in IE recently, it's no wonder! Google News has related stories, including many on the recently disclosed (and patched!) bug in Mozilla on Win2k/XP machines (documented on Slashdot on Thursday)"
"but one percentage point is simply not statistically significant."
It's extrememly significant. When is the last time a dominant MS end-user product *lost* marketshare? Coming at a time when the "Life Around MS Campus Is Going To Get Tough" memo is released, I think it shows that MS is for once (and largely thanks to open source) finding itself with an actual fight on its hands.
Go penguins! And little BSD daemons. And that... Mozilla lizard thing. :-)
"A loss of 1 percent of the market may not mean much to Microsoft, but it translates into a large growth, proportionately, in the number of users running Mozilla and Netscape-based browsers. Mozilla and Netscape's combined market share has increased by 26 percent, rising from 3.21 percent of the market in June to 4.05 percent in July, Johnston said." A 26% increase for Mozilla and Netscape in less than one month is significant to them. I'll bet MS considers it significant also.
If you read the friendly article you see that there was a 26% share increase for mozilla (from 3.21 to 4.05).
Posted anon cause whores quoting the article getting insightful annoy me.
Different sites attract different audiences. w3schools.com has a much higher percentage climb for Mozilla based browsers and Opera which is good because it means web development types are starting to prefer (and test their sites in) alternative browsers.
Google is also showing a slow and steady climb of Mozilla based browsers. It seems that the only people who are moving to IE6 are ones upgrading from 5.x.
It doesn't look that much when you see how much IE is used but I'd say we're getting to the stage where there's enough users of alternative browsers that any company would be crazy to ignore them.
Once Firefox 1.0 is released I'd expect to see even more people using it.
Check out the W3Schools browser statistics. Mozilla based browsers have grown from just a few percents to over 12% in a little over year with steady increases every month. Now W3Schools is hardly your average Joe's website, but they are pretty representative of average Joe web developer I'd say.
I'm also seeing Firefox evanglism and enthusiasm in new places all the time, on gamer boards (Shacknews) and other unlikely places. It's because Firefox is the new cool thing, something regular Mozilla never achieved. The Firefox branding effort has paid of big time. Having the best browser doesn't bring in the users, having the best browser with a slick look does.
It's like deja vu all over again.
The stats I have seen show IE having at best 90% but more like 80% of the browser market. Some even show lower than that. Of course, it very much depends on the web sites demographic. The Google zeitgeist is pretty good and it certainly shows a significant uptick in Mozilla usage of late. I would trust Google more than any other site as it is a site that anyone on any platform will find useful so should be more reliable.
"I have the attention span of a strobe lit goldfish, please get to the point quickly!"
In all seriousness I don't understand why Mozilla hasn't taken over the browser market already.
I am absolutely in love with Firefox and I use it almost exclusively, but I'm sitting here looking at slashdot and parts of the stories overlap the menus on the left. Not so in IE. If you want to see a *really* egregious example of this, go to liquidations.com.
Now I don't know a thing about why this is (if there is some adjustment that I can make to fix it, I'd love to know), but if that happens with slashdot--which is ALL ABOUT open source--what do we expect?
It's easier to wear the spandex than to do the crunches. --David Lee Roth
If you had actually taken a GOOD look at the graphs on this site you would have noticed that the browser graph has the following caption:
Web Browsers Used to Access Google
March 2001 - May 2004
As the article is talking about the last month (june) you couldn't really expect it to show up. However you are right that 1 percent will not be very obvious on the graph.
It would not be unreasonable to suggest that those sites that would install such software tend to be more computer oriented and thus visited by more tech savvy visitors
Yes, it would be unreasonable. Websitestory is one of those pagecounter services (add an image at the bottom of your page kind of deal). If anything, it would not be unreasonable to suggest that the sites that use it are NOT computer oriented (since a site like slashdot knows how to implement their own counters).
So it only measures visitors to sites that have specifically installed the software. It would not be unreasonable to suggest that those sites that would install such software tend to be more computer oriented and thus visited by more tech savvy visitors, people pre-disposed to have an alternative browser. That may not be the case of course, but in no way is this a true random sample of websurfers.
as someone who works for another major web analytics provider i can tell you that they would in fact have a wide variety of clients, so their numbers would not be slanted toward tech sites.
IMHO their stats would actually be one of the best indicators of a market trend it would be possible to get.
When your sample size is large enough, your error margin gets vanishingly small. They can look at the logs of important web sites and see what browsers are hitting them; that way they can "survey" a million users, which makes the sampling error .1%. And they number is probably more like 10 million.
That assumes, of course, that their methodology for picking users is correct. If last month they chose MSN.com, and this month they swapped it out for slashdot, that would skew their results far more than the sampling error would. But methodological errors are hard to put error bars on.
The most severe bug however is the Software Update feature. I installed 0.9.1 last week and almost immediately I saw an article on Slashdot about a Firefox security hole and fix. I didn't immediately attempt to install the fix. So a few days later I went to mozilla.org and saw that 0.9.2 was the latest version. Help->About shows I'm still at 0.9.1.
The software update feature was first introduced in 0.9 and therefore probably has bugs of its own, the good thing is that all the problems with the automatic update are being found quickly and will be fixed for 1.0
You don't need 0.9.2 - it's just 0.9.1 with the security patch applied. If you've applied the patch you don't need to upgrade as there's no other differences
Where I work (Porsche, Germany. Ok, not this large a corporation, but still larger than a mom-and-pop company), using IE is deemed a major security risk and thus forbidden.
We are all using Netscape 7 as our main browser.
So yes, sensible corporations have been listening.
Heep an eye on the Mozilla ActiveX project. On there is a number of things:
- An activex plugin that lets you run activex controls in Mozilla (be very careful with this, read the docs - you can lock it down to host just the controls you need)
- A way of embedding Mozilla into other browsers using activex
- A means of making IE support Netscape type plugins (which it used to at one time)
- An activex plugin for legacy browsers like Netscape 4
Unfortunately they're having problems getting this to work in Firefox 0.9 but keep an eye on that page for what you want.
In statistics what matters is choosing a random sample---an SRS (simple random sample). Sites such as windowsupdate and kernel.org don't matter. To choose a statistically-valid SRS, start by defining the population: all web sites on the Internet. Then choose a sample size n by random from these. Then determine the browser usage from the SRS. BTW, as the sample size n increases toward the size N of the population, the precision increases also.
Personally, I wouldn't consider using anything other than Opera. The 7 series is near perfect.
That said, Opera identifies itself as Internet Explorer 6 by default. As much as I would like to see these surveys reflecting an increase in its market share (as is surely happening, just inconspicuously), the developer's paranoia about the recent MSN style-sheet incident isn't helping in this respect.
New Opera users who are unaware of this setting can change it by pressing F12 and selecting 'Identify as Opera'.
1% is not really any news. Seriously, it is pathetic that /. is jumping up and down, all giddy, for one percentage point.
If you like rejoicing over a diminishing marketshare for Microsoft, then you should go here.
IIS had its record market share some time around april 2002, and has since lost about 14%, mostly to Apache.
IIS has 35% and went down to 21%, Apache had 56% and went up to 67%
The Internet is full. Go Away!!!
And so at last the beast fell and the unbelievers rejoiced. But all was not lost, for from the ash rose a great bird. The bird gazed down upon the unbelievers and cast fire and thunder upon them. For the beast had been reborn with its strength renewed, and the followers of Mammon cowered in horror.
from The Book of Mozilla, 7:15
"The number of Unix installations has grown to ten, with more expected." (Unix Programmer's Manual, 2nd ed.; june 1972)
Uh, not really. Taco has always said that IE figures heavily into Slashdot's stats. Everyone at work, after all.
When I log into citibankcards.com (using mozilla of course) there is a message in bright red that comes up warning users that they should not use IE. It seems to come up no matter what browser I use.
This is probably because Citibank was specifically targeted by that password-sniffing exploit of IE that came up recently. The exploit installed something via IE that send passwords directly via HTTP, which would bypass firewalls entirely. The security problems in IE have finally become dangerous to their users--this is beyond simple spyware or adware, this is real no-holds-barred computer crime.
-- "Makes Little Debbie look like a pile of puke!" - Moe Szyslak
If you believe that Netscape was slower than IE (each browser on its own), grap a 16-bit copy of IE 4 (the last one I know that had a 16-bit version) and a 16-bit copy of Netscape Navigator 4, and a copy of MicroSoft Windows 3.x.
Netscape was noticeably faster (although horribly slow on a 386 with 16MB of RAM). While Navigator was usable once loaded, IE was another story entirely. Once you began using MicroSoft Windows 4.x, parts of IE were loaded whether you wanted them to be or not which created the appearance that it was less resource intensive (which worked, since that's all people cared about, and it was too difficult to prevent it from doing so). Mozilla addressed this on MicroSoft platforms by doing something similar, by having it launch at start up and stay in the tray (although this can be enabled/disabled easily).
Most likely, she doesn't like the Firefox skin... But show her how awesome tabbed browsing is and she might start to change her tune.
The text overlap on /. is known about:
/. referrers though, so you'll have to go to the URL yourself.
bugzilla
They don't like
Yeah, but everyone does that more than they do other things and should be using free software for it. Microsoft should not be used for anything that touches the web and should be run virtualized of firewalled heavily if not blinded to your network.
Dual booting solves the game problem and more often than not, the games are moving to Linux anyway. Dual booting is a pain and best done with bios or swapable hard drives unless Winblows requires a rebuild. At the same time, Windows gaming, with all of it's Direct X dll hell has always been a pain under winblows, so gamers should not have a big problem with your proposition.
Until it is possible to run practically any Windows software under Linux with no problems, the most you are going out of the majority of home users is a dual-boot, if that. Certainly not complete swap-outs.
Bah. I've been swapped out for years now. There's enough "good enough" free software for everything. Let me tell you, it all runs much more practically and easier than Windoze junk ever did.
A nice little program called Bosch can solve the rest of the world's windoze problems. Check out these screenshots for yourself:
Official shots
XP running in Bosch under Linux
It may be slow and hurt, but it's way easier to do that than it is to keep a real windoze machine going with email, web browsing and all of those other things Windoze is not good at.
Friends don't help friends install M$ junk.
I can't find any information on how they've collected these statistics, but this change could just be a change in the number of people that are spoofing the user-agent string. It could be some email circulating telling people to visit WebSideStory or one of the sites they track.
Without some measure of the accuracy of that 1% figure, it's hard to judge how significant this report is.
Just for the sake of karma whoring, here are the instructions for patching Mozilla and Firefox.
Remember that this is only for those folks running Windows 2000 or Windows XP.
Yes, it's me again.
Want small businesses to move to Mozilla? It can happen, but the lack of a fully-featured calendar sticks them to Outlook.
Take a look at some of my prior comments. This is a huge issue. I do consulting on the side and you don't know how many times I hear:
"Can the Calendar in Mozilla act like Outlook?"
"Can it import Outlook meeting requests?"
"Can I sync it with my PDA?"
"Can I email requests to other users so they can just double-click it and add it to their Mozilla calendar?"
The answers to the above are pretty much NO.
Yet, there's time to bicker about the default theme for Firefox.
Screw themes. Let's work on functional features.
Small businesses just won't give a shit about Mozilla unless basic calendaring features like the above find their way into the software. Instead, they'll do what they've been doing - Windows Server 2003 Small Business Edition, and Outlook. Just to send calendar invites back & forth.
Sad, but true.
Netscape Calendar used to work well. What the hell happened?
Let s = sample standard deviation.
Let w = +/- width of the confidence interval.
Let n = sample size needed.
Let k = multiplier for the confidence level... Use 2 for 95% confidence, 3 for 99% confidence.
k(s/(n^0.5)) = w (for 95% C.I.)
Solving for n:
n = 4(s^2)/(w^2) (95% C.I.)
n = 9(s^2)/(w^2) (99% C.I.)
This result is also intuitively satisfying: You need a larger sample if
- You want a higher confidence level
- The sample standard deviation is larger
- You want a smaller confidence interval
Source: Statistics, 4th Ed., McClave & Dietrich, Dellen Publishing Co., © 1988. Page 327, "Determining the Sample Size Necessary for Making Inferences About a Population Mean".