Browsers and Operating Systems

The next set of questions we asked focussed on the respondents day to day technology use — the browsers and operating systems they use themselves, and then asked which browsers they test the web sites and applications they build with.

What operating systems do they use in their day to day work?

Given the still very high reported market share of the Windows operating system (as high as 92% according to Hitslink in November 2009), the fact that the single largest operating system respondents reported using was Mac OS X 10.6 (at 37%), with Mac OS X 10.5 reported as the OS of choice by a further 13% of respondents, and 10.4 still at 1.3% making gave a total of over 50% of all respondents using Mac OS X. This is up from around 46% market share for Mac OS X among respondents last year. In the 13 months between surveys, Mac OS X 10.6 was released, and yet this is now the OS of choice for 70% of Mac OS users — showing how rapidly Mac users tend to upgrade.

Last year, Windows through all versions slightly lead Mac OS X in market share terms, but Mac OS X now leads Windows 50.7% to 41%.

Last year we observed that Windows XP had a fourfold lead over the much newer Windows Vista, and while XP is still the number 1 version of Windows used (24.5% of all users), Windows 7 has shown much higher rates of adoption at 10.3% than Vista (down to 6.2% from 9.2% last survey).

The other significantly used OS is Linux, which around 4.2% of respondents use as their primary OS. This is down slightly (4.5%) from 2008, but significantly higher than various sources report for general Linux use.

Which OS do you primarily use for day to day work?

2010
OS Count %
Windows XP 343 24.47%
Windows Vista 87 6.21%
Windows 7 145 10.34%
Mac OS X 10.4 20 1.43%
Mac OS X 10.5 184 13.12%
Mac OS X 10.6 507 36.16%
Linux 59 4.21%
Unix 0 0%
Other 4 0.29%
2008
OS Count %
Windows XP 465 37.68%
Windows Vista 113 9.16%
Mac OS X 10.4 65 5.27%
Mac OS X 10.5 512 41.49%
Linux 56 4.54%
Unix 1 0.08%
Other 7 0.57%

Overall OS share

2010
OS Count %
Windows 575 41%
Mac OS X 711 50.71%
Linux 59 4.21%
Unix 0 0%
Other 4 0.29%
2008
OS Count %
Windows 578 46.84%
Mac OS X 577 46.76%
Linux 56 4.54%
Unix 1 0.08%
Other 7 0.57%

What browsers do they use?

Of more practical interest particularly to web professionals, is the question of which browsers respondents use. Here we asked several related questions. We asked respondents which browsers they use themselves, and in addition for the first time, what factors are important in that decision. We then asked which browsers respondents tested their sites in which should reflect the sense of which browsers remain relevant in the mind of web professionals

As with last year, among our respondents, all versions of Internet Explorer combined (including IE8, released since the last survey) were used substantially less as a primary browser than Firefox or Safari alone. IE 6, 7 and 8 combined had a share of under 2% (last year IE6 and 7 combined were around 3.5%) Internet Explorer 8, the most widely used version of IE at under 1% (last year IE7 was used by 3.2%) was used less than any other browser we asked about, other than Safari 2. But there were other big movements as well.

Last survey, Firefox 3 was the 800lb Gorilla, with 60.5% market share. Now, Firefox 3.5 (16.3%) and 3.6 (37.8%) combined have a share somewhat lower, at 54%. In the interim Safari 4 was released and now has a nearly identical share at 20.26% that Safari 3 had last survey (20.91%). Evidence for how rapidly Safari users upgrade is to be found in the fact that only 2 respondents our of over 1400 now user Safari 3. Opera has seen a drop off from around 3% to under 2%, but the big mover is Google Chrome, up from 4.2% to 17.2%.

As Chrome and Safari share a very similar underlying core, Webkit, among our respondents, Webkit with 37% has gained dramatically on Gecko, Firefox’s engine in the last 13 months.

These are of course profoundly different results from general browser market share. For example as of late 2008, Hitslink was reporting IE market share of 61.5% (down from around 70% in late 2008), Firefox 24.2% (up from 21%). Safari 4.5% (down from 7.3%), and Chrome at 5.6% (up from less than 1%). In general browser share Opera is also up considerably to 2.4%, from under 1%.

Which browser do you primarily use for day to day browsing?

2010
Browser Count %
IE6 0 0%
IE7 11 0.78%
IE8 13 0.93%
Firefox 3.5 229 16.33%
Firefox 3.6 530 37.8%
Safari 3 2 0.14%
Safari 4 284 20.26%
Google Chrome 241 17.19%
Opera 24 1.71%
Other 10 0.71%
2008
Browser Count %
IE6 15 1.22%
IE7 40 3.24%
IE8 3 0.24%
Firefox 2 44 3.57%
Firefox 3 741 60.05%
Safari 2 1 0.08%
Safari 3 258 20.91%
Google Chrome 52 4.21%
Opera 38 3.08%
Other 27 2.19%

Rendering Engine market share

2010
Engine Count %
Trident 24 1.7%
Gecko 759 54.1%
Webkit 527 46.6%
Opera 24 1.71%
2008
Browser Count %
Trident 58 4.7%
Gecko 785 63.6%
Webkit 311 25.2%
Opera 38 3.08%

Summarizing last years results, we observed

Exactly why such a high percentage of respondents opt for a browser other than the default platform browser (predominantly Firefox, though Opera on Windows is nearly as widely used as Internet Explorer) on both the Mac and Windows is hard to say, and would be worth investigating in follow up surveys. Reasons might include the widespread use of developer tools like Firebug for Firefox, and DragonFly for Opera.

So, this year, we went ahead and asked. We identified several factors that might be relevant to a respondent’s choice, and then gave them the option of saying how important (on a scale of 1 to 5) this factor is. We then grouped these into three bands — not important (1 or 2 out of 5), important (3 out of 5) and very important (4 or 5 out of 5).

We were somewhat surprised by the results. Based on some ad hoc surveying via Twitter, we thought that developer tools (like Firebug), would have played perhaps the most significant role. In fact, built-​in and plug-​in developer tools received the the lowest percentage of “very important” (and combined “important” and “very important” responses) (though between 40 and 50% of respondents still reported that these tools were very important to their decision). At the top of the list was “standards support”, with 81.5% of respondents saying this was a very important factor in their decision of primary browser. Stability and performance also proved popular, while CSS3 support (with 62% rating it very important and a further 19.6% rating it important in their decision) comfortably outpointed HTML5 support, with 55.8% and 23.3% respectively.

What features are important in your decision?

2010
Feature not important important very important
Built in Developer Tools 30.16% 23.62% 42.73%
Plug in Developer Tools 30.4% 14.3% 51.84%
Performance 10.06% 12.08% 74.41%
stability 9.98% 11.71% 74.84%
Standards Support 9.01% 6.1% 81.41%
CSS3 Support 15.4% 19.61% 61.51%
HTML5 Support 17.45% 23.24% 55.83%
User Experience Design 14.59% 20.61% 61.32%
Familiarity 18.57% 20.3% 57.67%

Mobile Browsing

While we didn’t ask explicitly which mobile browsers respondents use, we did ask developers what other browsers they use in addition to their primary browser. From the results of this question (which required users to enter the names of the browsers, and not merely select from a list, and so if anything under-​represents the nominated browsers), Mobile Safari was mentioned by 29% of respondents (up from 16% the previous survey). Android was mentioned by 3.8% of respondents, Opera Mini 2.6%, Opera Mobile 1.5% (both more or less holding steady), while Blackberry increased from .75% to around 1.1%. A Nokia browser (which specific one it is often not possible to determine) was mentioned by .8%, down from 1.6% last year. Palm’s Pre, while being in many ways the most web developer friendly, was mentioned by only 3 of the respondents to the survey.

Mobile Browser Summary

2010
Browser Count %
Mobile Safari 413 29%
Android 52 3.8%
Pre 3 .2%
Blackberry 15 1.1%
Opera mini 36 2.6%
Opera mobile 20 1.43%
Nokia 11 .8%
2008
Browser Count %
Mobile Safari 200 16%
Blackberry 9 .75%
Opera mini 42 3.4%
Opera mobile 20 1.6%
Nokia 20 1.6%

What browsers do they test with?

While even in the last two or three years browsers have made big strides in more consistently supporting underlying web standards, meaning fewer cross browser headaches for developers, respondents still found it important to test their sites in multiple browsers. Which browsers they choose reflects the collective judgement of the industry as to the relevance of particular browsers.

In the 13 months between our two surveys, there’s been considerable change in the browsers respondents test in. In the first survey, the single most widely tested in browser was Firefox 3, at 92%. IE7 was tested in by 89.5%, with IE6 still occupying the attention of 78.5% of respondents. The then in late beta IE8 only had a surprisingly low 29.5% of respondents concerned enough to test with it. Meanwhile Safari 3 then had over 80% of respondents testing their sites in it.

In this most recent survey, IE8, and IE7 top the list, with 83.5% and 82.5% of respondents testing in them respectively. Safari 4 still has 80% of respondents testing with it, but Firefox 3.6, in comparison with Firefox 3’s 92% has 69% of respondents testing in it, along with 61% testing in Firefox 3.5. This in part reflects the fragmentation we saw among Firefox users — split between versions 3.5 and 3.6, rather than all using the same version as in the previous year. IE6 is following IE5.5 off the radar as a browser which respondents feel needs to be tested in, dropping from 78.5% to 59.7%.

Reflecting its increase in use by respondents (and not least its arrival on the Mac platform since the last survey) Chrome increased from a healthy 39.5% to now 76.9%.

Opera, for its relatively low market share and use as a primary browser among respondents, managed to more or less hold steady among respondents who test their sites with it at just under 50%.

On the mobile front, Mobile Safari jumped from 21% to 38%, while Opera Mobile edged up from 4.9% to 5.7%. Android recorded a surprisingly low .8%, given its much higher mention by respondents as a browser they actually use! Blackberry is mentioned even less, with the Pre receiving only a single mention. In the case of the Pre, and perhaps Android, being webkit based, perhaps developers consider that a site which works on the iPhone will also work fine on these other webkit based browsers. There’s also the challenge that only actually physically using a device allows for adequate testing.

Which browsers do you test your web sites with?

2010
Browser Count %
IE5 18 1.28%
IE6 837 59.7%
IE7 1158 82.6%
IE8 1171 83.52%
Firefox 2 157 11.2%
Firefox 3 522 37.23%
Firefox 3.5 859 61.27%
Firefox 3.6 973 69.4%
Safari 2 36 2.57%
Safari 3 358 25.53%
Safari 4 1124 80.17%
Mobile Safari 533 38.02%
Google Chrome 1078 76.89%
Opera 660 47.08%
Opera Mobile 80 5.71%
2008
Browser Count %
IE5 53 4.29%
IE6 968 78.44%
IE7 1104 89.47%
IE8 364 29.50%
Firefox 2 604 48.95%
Firefox 3 1135 91.98%
Opera 616 49.92%
Opera Mobile 61 4.94%
Safari 2 237 19.21%
Safari 3 992 80.39%
Mobile Safari 259 20.99%
Google Chrome 488 39.55%

In the last survey wrap up, we observed

It will be interesting to see how the number of people testing in mobile browsers changes in the coming 12 months, and it would be expected that a considerably higher percentage of developers will be testing in some kind of mobile browser 12 months from now.

While testing on the iPhone has increased dramatically, testing on other devices, particularly Android, has not seen anything like this increase. As we’ll see later in the survey, the number of respondents optimizing their sites for mobile is still surprisingly low. As that percentage increases, we’d expect to see a related increase in the number of respondents focussing their testing specifically for mobile devices.

Next

Next we’ll look at the markup techniques and philosophies that respondents use to develop for the web.