Web Directions

Browsers and Operating systems

• What oper­at­ing sys­tems do they use in their day to day work?
• What browsers do they use?
• What browsers do they test with?
• Mobile Browsers

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.

We asked about the tech­nol­ogy use of the respon­dents them­selves as we felt it might give an insight into trends such as the adop­tion of par­tic­u­lar oper­at­ing sys­tems, par­tic­u­lar browsers, for exam­ple Google’s recently released Chrome, over time, and so on. It’s widely believed (though far from eas­ily proven) that early adopters like those who responded to this sur­vey give at least some indi­ca­tion of the future behav­ior of more main­stream users. Certainly, the now wide­spread adop­tion of Firefox was pre­fig­ured by a sig­nif­i­cant uptake by tech savvy peo­ple, par­tic­u­larly web design­ers and devel­op­ers, and sim­i­larly, the rise in the use of the Mac OS by main­stream com­puter users fol­lows a large increase in the use of that oper­at­ing sys­tem among web design­ers and devel­op­ers going back some years now.

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 around 90% accord­ing to Hitslink in November 2008), the fact that the sin­gle largest oper­at­ing sys­tem respon­dents reported using was Mac OS X 10.5 (at 41.5%), with 10.4 of the same OS reported as the OS of choice by a fur­ther 5.3% of respon­dents, mak­ing for a total of nearly 46% of all respon­dents should come as a sur­prise, despite the anec­do­tally high use of the Mac OS by web pro­fes­sion­als. Overall, Windows still leads as the sin­gle most widely used OS, with a com­bined XP and Vista per­cent­age of 46.8, less than a per­cent higher.

Among win­dows users, XP still dom­i­nates Vista by a fac­tor of 4 to 1, reflect­ing the gen­eral slow adop­tion of Microsoft’s new oper­at­ing sys­tem two years after its release. In fact, respon­dents who use Windows are less likely to use Vista than Windows users on the whole accord­ing to the Hitlink sur­vey quoted above, where the dif­fer­ence was about 3 XP users for every 1 Vista user.

To com­pare adop­tion rates, Mac OS X 10.5 was released around a year later than Vista, and has an adop­tion rate among respon­dents who use the Mac OS of around 87% (as opposed to Vista’s 20%).

The other sig­nif­i­cantly used OS is Linux, which around 4.5% of respon­dents use as their pri­mary OS. This is sig­nif­i­cantly higher than var­i­ous sources report for gen­eral Linux use.

OS by orga­ni­za­tion type

If we break down oper­at­ing sys­tem use by orga­ni­za­tion size, we find that as the size of the orga­ni­za­tion dimin­ishes, the per­cent­age of users using non-​​windows OSs increases markedly. While the use of Linux stays rel­a­tively sta­ble, as the size of the orga­ni­za­tion dimin­ishes, the per­cent­age of Windows Vista users actu­ally increases from 3% in the largest orga­ni­za­tions, to 15% in orga­ni­za­tions of 2 – 10 employ­ees. The pro­por­tion of Vista to XP users grows even more dra­mat­i­cally, from less than 1 in 10 in the largest orga­ni­za­tions, to more than 1 in 3 in small organizations.

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. The first two relate to the browsers respon­dents use them­selves, and we then asked which browsers respon­dents tested their sites in.

Among our respon­dents, all ver­sions of Internet Explorer com­bined were used sub­stan­tially less as a pri­mary browser than Firefox or Safari alone. Internet Explorer 7, the most widely used ver­sion of IE at just 3.2% was still used less than Google’s recently released Chrome (4.2%), and only mar­gin­ally more than Opera (3.1%).

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 70%, Firefox 21%. Safari 7.3%, and no other browser above 1% (includ­ing Chrome).

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. A ques­tion about devel­op­ment method­ol­ogy later in the sur­vey might also give some clue. When respon­dents were asked about their “approach to devel­op­ing for mul­ti­ple browsers”, 85% responded either that they

develop to W3C stan­dards, and then work around IE

or

develop to W3C stan­dards and expect browsers to sup­port these

Given the wide­spread belief that stan­dards sup­port in browsers other than Internet Explorer is bet­ter than in that browser, this fac­tor in itself might account for why both for prac­ti­cal and philo­soph­i­cal rea­sons, web pro­fes­sion­als pre­fer a browser other than Internet Explorer.

It must be kept in mind that there is no ver­sion of Internet Explorer for Mac OS X or for Linux, whereas there are ver­sions of Firefox for Windows, Mac OS X and Linux. A related point of pass­ing inter­est is that Safari for Windows is used by less than 1% of Windows using respon­dents as their pri­mary browser.

It’s not sur­pris­ing too that 75% of respon­dents replied that they used other browsers in addi­tion to their pri­mary browsers. It’s not clear from the responses (and the ques­tion could have been bet­ter worded to ascer­tain more clearly) whether these sec­ondary browsers are used for brows­ing gen­er­ally, or test­ing an devel­op­ment specifically.

Mobile Browsing

In future sur­veys, it would be worth ask­ing respon­dents about their use of the web with a mobile device. But for now, we can observe the fol­low­ing about mobile browser use.

• 9 respon­dents men­tioned using the Blackberry browser
• just under 200, or more than 15%, men­tioned using Mobile Safari
• 20 men­tioned using a Nokia browser
• 42 men­tioned Opera Mini
• 20 men­tioned Opera Mobile

It’s clear that Mobile Safari (the browser in the iPhone and iPod Touch) is cur­rently the dom­i­nant browser for web design­ers and devel­op­ers in this sur­vey. But Opera is sur­pris­ingly well rep­re­sented. It’s inter­est­ing to note too that 75% of respon­dents don’t use the mobile web, or at least not suf­fi­ciently fre­quently for it to occur to them to respond.

Despite the hype sur­round­ing the iPhone and its web brows­ing capa­bil­i­ties, if early adopters like our respon­dents are any indi­ca­tion, wide­spread mobile web use is still some way off.

Other plat­forms

In addi­tion to PC and mobile browsers, 28 respon­dents men­tioned test­ing print ver­sions of their sites, 5 with the Wii, and 4 referred to test­ing on “TVs”. It’s clear, as we’ll see from later sur­vey answers, that the focus of most devel­op­ers con­tin­ues to be pre­dom­i­nantly on tra­di­tional laptop/​PC based brows­ing envi­ron­ments, with some focus on mobile plat­forms, and lit­tle out­side that.

Other thoughts on addi­tional browsers

While Internet Explorer has a sur­pris­ingly small per­cent­age of pri­mary users, a fur­ther 240 or so respon­dents men­tioned it as an addi­tional browser they use. Given that around 90% of respon­dents (as we’ll see in a moment) say they test their sites in this browser, I’d sug­gest that this fig­ure of 240 (or about 20% of respon­dents) rep­re­sents some kind of use other than sim­ply test­ing. It may for exam­ple be for vis­it­ing a site, or appli­ca­tion which only works in Internet Explorer (a still too fre­quent occur­rence), for intranet or inter­nal appli­ca­tions which require Internet Explorer, or so on. In future sur­veys it may be inter­est­ing to try and get some fur­ther under­stand­ing about why respon­dents use the browsers they do, and what they use them for.

Google Chrome, recently released to con­sid­er­able fan­fare, rates less than might be expected among this audi­ence as a pri­mary browser. It fares bet­ter in regards to alter­na­tive browsers, with about 10% of users report­ing it as an addi­tional browser. Given 40% of respon­dents report test­ing their sites in Chrome, we might once again con­clude that this 10% fig­ure rep­re­sents uses other than sim­ply test­ing. Keep in mind too that Chrome is avail­able at present only for Windows, and so, par­tic­u­larly given the high num­ber of non Windows using respon­dents, one would expect an increase in its share once it is avail­able on other platforms.

We’ve already seen the dom­i­nance of Firefox among respon­dents as their pri­mary browser. A fur­ther 300 or so respon­dents, or about 25% also list it as an addi­tional browser. Couple this with the per­cent­ages who test with one or more ver­sions of Firefox, which we’ll take a look at next, and among the respon­dents of this sur­vey, Firefox is still very much the dom­i­nant browser.

What browsers do they test with?

Perhaps the sin­gle biggest chal­lenge for web design­ers and devel­op­ers going back to the mid­dle 1990’s has been the fact that pages appear dif­fer­ently in dif­fer­ent browsers. Often this is by design — CSS is designed to give pref­er­ence to the user’s options for things like font size over the the site’s own style sheets. In other instances, the dis­crep­an­cies are to do with plat­form dif­fer­ences (for exam­ple the dif­fer­ence in default DPI of Windows and the Mac OS makes for appar­ently dif­fer­ent font sizes on these dif­fer­ent oper­at­ing sys­tems). But many of these dis­crep­an­cies occur because of bugs in the way the CSS stan­dards are imple­mented, or dif­fer­ences in imple­men­ta­tion of these stan­dards due to ambiguities.

So, devel­op­ers and design­ers have found it nec­es­sary to test across a range of browsers (includ­ing dif­fer­ent ver­sions of the same browser) in order to ensure that pages work accept­ably across widely used browsers.

A decade ago, devel­op­ers and design­ers might test in two ver­sions of Internet Explorer and Netscape Navigator (the cur­rent, and pre­vi­ous ver­sions). Five or six years ago, before the mete­oric rise of Firefox, and the resur­gence of the Mac plat­form, they might even test only in the cur­rent and pre­vi­ous widely used ver­sions of Internet Explorer on Windows. Few would test in any kind of device other than a PC or laptop.

But as of the end of 2008, where are respon­dents test­ing their sites?

We first asked about the com­monly used browsers that respon­dents tested their sites in.

Glancing at the results, what’s prob­a­bly most sur­pris­ing is how few peo­ple test in Internet Explorer 8 (29.5%, com­pared with 40% for Chrome, 50% for Opera, and around 20% for both the long out­dated Safari 2, and mobile Safari). While IE8 is still in beta, there had been ver­sions released for nearly 10 months at the time of the sur­vey, and with sig­nif­i­cantly improved stan­dards sup­port, a new stan­dards mode, and related issues around tar­get­ing dif­fer­ent ren­der­ing modes for devel­op­ers to con­tend with, it’s sur­pris­ing more devel­op­ers haven’t already started test­ing their sites in this ver­sion of IE.

Some rea­sons which might account for this low uptake include

• the dif­fi­culty of run­ning mul­ti­ple ver­sions of IE on the same system
• a sense that sites which work fine in IE7 will also work fine in IE8
• since the cur­rent ver­sion of IE8 is not final, the sense that test­ing and address­ing issues before a final can­di­date release might be a waste of effort
• per­haps a sense with Microsoft’s efforts to pro­vide bet­ter sup­port for stan­dards based ren­der­ing, IE8 will ren­der pages a lot more like Firefox, Opera and Safari
• A still tiny (less than 1%) mar­ket share in late 2008 for the browser

This low test rate nonethe­less took us by surprise.

Other results of inter­est include the fact that less than 80% of respon­dents now test their sites with IE6, despite its reported mar­ket share of around the same as Firefox, which has a test rate of over 90%, and nearly 3 times Safari’s mar­ket share, which has a sim­i­lar test rate to IE6.

This would sug­gest that respon­dents favor the browsers they use over more pop­u­lar browsers when it comes to test­ing. A later ques­tion we asked, which we’ve already referred to, regard­ing devel­op­ment method­ol­ogy, might throw some light on this as well. As we’ve seen, when respon­dents were asked about their “approach to devel­op­ing for mul­ti­ple browsers”, 85% responded either that they

develop to W3C stan­dards, and then work around IE

or

develop to W3C stan­dards and expect browsers to sup­port these

The trend it would seem, is for devel­op­ers to focus on cur­rently ship­ping browsers, and less so on legacy browsers. This will be an inter­est­ing trend to fol­low in future surveys.

Mobile Browsers

With the mobile web an increas­ing real­ity, it’s inter­est­ing to see that only around 20% of respon­dents tested their sites in Mobile Safari, and less than 5% in Opera mobile. Looking into what “other” browsers respon­dents were asked they tested in, we find only 2 ref­er­ences to Nokia browsers, one to OpenWave mobile browsers, 6 to Blackberry, 4 to Opera Mini, and only 2 for Windows Mobile IE.

So, it’s clear that for all the talk of the mobile web finally arriv­ing, even the sig­nif­i­cant major­ity of early adopter devel­op­ers don’t test in even the most com­monly used mobile browsers. Coupling this result with a later ques­tion “Do you opti­mize your sites for devices other than laptops/​PCs?”, where around 25% of respon­dents answered yes, with some kind of mobile browser being the over­whelm­ing major­ity of browsers respon­dents men­tioned opti­miz­ing for, this sug­gests there’s still a way to go before it’s com­mon that devel­op­ers focus on even test­ing for com­mon mobile browsers, let alone specif­i­cally develop for them.

This might be in part due to the per­cep­tion that since the over­whelm­ing major­ity of mobile web users use the iPhone or iPod Touch, with a ver­sion of the Safari browser, (and other mobile plat­forms like Nokia’s S60, and Android based phones use a browser based on the same under­ly­ing Webkit tech­nol­ogy), that if it’s fine in Safari, it will also be in these related browsers as well. This might also account for the very high (80% +) per­cent­age of respon­dents who test in Safari 3.

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.

Next

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