Friday, January 9th, 2009

State of the Web 2008

Category: Survey

John Allsopp has put together the data from his extensive State of the Web 2008 survey. In it you will find details on who was surveyed to set context, the browser / os usage claimed, and answers on the use of technology such as:

John and Scroll Magazine have a nice conclusion:

In this first of its kind study, over 1200 web professionals from across the globe, in a broad range of industry sectors and organization types, were surveyed on the browsers, operating systems, development technologies and practices they use when building web sites and applications.

Some surprising results from the survey include

  • Nearly half the respondents use Mac OS X Leopard, and over half use a non Windows Operating system. Windows XP still outweighs Windows Vista among these users by a factor of 4 to 1 as their operating system of choice.
  • Just a small majority, less than 5%, use any version of Internet Explorer as their primary browser, while Firefox dominates as the browser for choice, with over 60% market share. Safari 3 follows with 21%, and the much talked about Chrome on just 4%.
  • Only a tiny handful use Internet Explorer 8 beta as their browser of choice.
  • Despite the hype of the iPhone, less than 20% of respondents use the mobile web, and a similar number develop sites optimized for mobiles.

Web Development technologies

When it comes to web technology use, standards based and open source technologies dominate.

  • Only 3% of respondents say they never validate their sites while 70% say that they frequently or always do.
  • Only 10% of respondents say they use tables for layout, while well over 90% use CSS for styling their pages.
  • 35% of respondents say they use microformats in their markup.
  • 95% of respondents use JavaScript, and of these, almost all use libraries.
  • JQuery is the dominant library used by some way, with 60% of respondents saying they use it.

With plug-in technologies, Flash continues to dominate, with a market share of around 60%. Silverlight still has a lot of work to do to catch the long time industry leader, with a bare 2%, little more than the Real format. Apple’s Quicktime has a surprising 20% of the market.
Java applets have all but disappeared from the toolset of these early adopter developers.

On the back end, open source accounts for the majority of technologies used. Among server operating systems Linux at nearly 60% is used more than twice as often as Windows at 28%, with Unix also well represented at 17%. Even Mac OS X, which is usually far down survey lists for server OSs, is used by 5.5% of respondents.
Apache at 70% is the dominant web server, with IIS at 23%.

Over 90% sites are database driven, with the open source MySQL at 70% and PostrgeSQL at 10% together accounting for the significant majority of sites by respondents. Microsoft’s SQL Server at 22% and Oracle at 9% were the other widely used database systems.

When it comes to server side programming languages, PHP is the most commonly used, at 63%, with JavaScript at 55%, ASP.NET at 17% and Python at 15%. Despite its flavor of the month status, Ruby comes in at 14%, with Java at 12%, indicating that the language which came to prominence with the rise of the web is well and truly being challenged from all sides when it comes to web back end development.

Posted by Dion Almaer at 9:10 am

3.2 rating from 27 votes


Comments feed TrackBack URI

The remarkably high figure for server-side javascript is probably incorrect. The question was actually “which programming languages do you use?”, and javascript was one of the options. The context was that you were supposed to mention only those user on the server, but in practice most people probably checked off the javascript box without thinking.

Comment by Joeri — January 9, 2009

Yeah, that SSJS figure can’t possibly be right. 55% on SSJS, yet in the survey on server-side frameworks, there were no JS ones listed at all as being popular. I would love to see more people using SSJS, but something must have been screwy there.

Comment by kriszyp — January 9, 2009

I’m sorry but these figures are stupid. They seem vary bias and at certain points impossible. For example:

– Nearly half the respondents use Mac OS X Leopard, and over half use a non Windows Operating system.

So no-one uses Windows?

– On the back end, open source accounts for the majority of technologies used. Among server operating systems Linux at nearly 60% is used more than twice as often as Windows at 28%, with Unix also well represented at 17%. Even Mac OS X, which is usually far down survey lists for server OSs, is used by 5.5% of respondents.

Again, this adds up to 110.5?!?

– Over 90% sites are database driven, with the open source MySQL at 70% and PostrgeSQL at 10% together accounting for the significant majority of sites by respondents. Microsoft’s SQL Server at 22% and Oracle at 9% were the other widely used database systems.

70% + 10% + 22% + 9%?

Comment by LewisWright — January 9, 2009

Did they bother to ask any of the 1.2 Billion people in India what browser they use? I think there would be a slight skew at IE 6, as most of them are scared to update their pirated operating systems or just don’t have the bandwidth to update.

As much as we all like to think everyone owns a Mac and those ridiculous commercials make us think that the Mac has no problems, a 21% share in Safari seems far fetched.

How about we all share our web logs and come up with a “real” number?

Comment by axiom — January 9, 2009

Here’s the reason most of the results don’t match our common experience. 25% of respondents work for design firms. I know a lot of developers, and none of them work in a design firm. Maybe that’s why the first stat I read claims the close to 50% of the respondents use an apple os for development.

Comment by azappdeveloper — January 9, 2009

Sheesh there are some hard task masters out there :-)

Dion’s quoted the mainstream press friendly press release there – there’s a lot more substance in the report itself, and you can also see every question, all the answers, and download a CSV of the data used for the analysis.

The other thing to note is that this was in many ways a subjective survey. It asked for philosophies, practices and attitudes to web development, as well as directly about technology. It’s objectively very difficult to learn about these “soft” aspects of development.

To those who made reference to the nature of the sample, we are very up front about who was surveyed – it is definitely early adopter types – those who read blogs like Ajaxian, those who attend conferences and so on. That’s probably why 25% of respondents work in design firms – those sorts of developers are it would seem more likely to be the kind of people who keep themselves up to date with developments.

I definitely think there’s value in learning about the technologies and practices that early adopter developers use. The argument would be that the technologies and practices such people use prefigures what more mainstream developers will be using in coming years. That’s not to say that because a lot more of our respondents use Mac OS X than the market share figures would indicate, then one day soon nearly 50% of computer users will, but rather, it may indicate that the growth in Mac OS market share may well continue.

@axiom – there’s a very big difference between sharing logs (which is essentially what netcraft and other surveys do), and this survey. This survey asked developers what browsers, operating systems and so on they use. The figures for Mac OS are higher than I expected, but if you attend web developer conferences, you’ll anecdotally see that the significant majority of people with laptops have Mac laptops.

@LewisWright, why some of these numbers come in at over 100% is that in many cases, you could provide more than one answer. Clearly some developers use more than one database system for example. In other cases we asked for a single answer (for example primary browser) – so in those cases they should definitely add up to 100% (if you look at the survey results proper, not the made for TV adaptation :-)

Indeed @kriszyp these results are biased. All survey results are – they are in this case biased toward the kind of respondent who attends conferences, reads blogs and web sites focussed on best practice web design and development, speaks English well, and so on. We made all that quite clear in our report. There’s no way we are arguing that the results are representative of all web developers, let alone all computer users. The question is, knowing the nature of the respondents, is there value in knowing their responses to these questions. I’d argue yes – but YMMV :-)

BTW, as the report and data makes clear, of course people use Windows – it’s just that a majority don’t.

Regarding JavaScript on the server side – we too thought that this was high – I think the report concludes as @Joeri did that there was a misunderstanding with that question – it appeared in the “back end” section of the survey, but the question itself asked about programming languages used, and one of the possible answers was JavaScript.

Thanks for the thoughts


Comment by JohnAllsopp — January 9, 2009

This survey is terrible. Web designers are not “developers” and they can’t code their way out of a paper bag. HTML is not coding, it’s markup … deal with it.

Mac’s as development? That’s ridiculous! Even a large percentage of non-windows developers use windows for php, rails, etc….

I’ve worked for one of the web’s top design firms and quit because the work was too easy and boring. Design firms make pretty sites that are usually handed off to real developers. The code that design firms end up doing are contact us forms and other brain dead easy pages.

I’d like to know how many of these “developers” could even explain what a singleton pattern looks like.

Are the editors asleep?

Comment by DeviantSatan — January 9, 2009

Wow you folks are harsh! I thought this was a very interesting survey; yeah, these kinds of things are tricky to figure out, but I’d rather have some insight into what fellow developers are doing then none. Sheesh.

Comment by Brad Neuberg — January 10, 2009

@DeviantSatan: This was a survey of web professionals, not web developers, therefore quite a few designers were surveyed. If you want a survey of web developers, look at a different survey.
Btw, the forums have 167 threads mentioning the word “singleton”. Most of the people on there come from a design background, and quite a lot use macs. You’d be surprised how much emphasis there’s been in the flash community on good code design the past few years.
I like this survey. It’s nice to see a wider survey than just developers. There’s too wide a gap between designers and developers, and anything that bridges them is good.

Comment by Joeri — January 10, 2009

@DeviantSatan: “Mac’s as development? That’s ridiculous!” I’m sorry but YOU are ridiculous. Get with the times you ancient old caveman. A LOT of people use Mac for development, perhaps if you had any real experience in the industry you’d know this. Learn to develop professionally, noob.

Comment by Gavin — January 11, 2009

Gavin no need to get all defensive. He’s not talking about Mac as a useless development tool, but as such a higher percentage as it is supposed to have. You’ll find that the majority of lower-level programmers are still on Windows. After all, it makes sense. Programmers aren’t interested by the fancy effects or ease to use of Mac, they’d rather have something they’ve been using for years and can power-use. Plus all those blurry fonts would kill a programmer’s eyes ;)

Comment by LewisWright — January 12, 2009

If you take a look at the very first question, asking respondents how they characterize themselves, and those calling themselves “developers” outweighed “designers” by a significant margin.

If you look at the percentage who say they use JavaScript (over 90%), and have database driven sites (a similar figure), you can see that what the overwhelming majority of respondents is doing is a lot more than simply marking up HTML.

Now, there are those out there who would might scoff at such efforts as being worthy of the term “developer”, but I’m afraid it seems we’ll have to just get used to that.

The web is democratizing “programming”, “development”, whatever you want to call it in a similar way that the Commodore 64, Apple II, and similar computers did in the late 1970s, and early 1980s, and in a similar way that the web is democratizing other traditionally “elite” activities (from journalism to banking).

Personally, as a computer science graduate with nearly 20 years experience developing and marketing software, and the developer of quite a few web deployed “applications”, I think that that is fantastic. But YMMV :-)

Comment by JohnAllsopp — January 12, 2009

95% use javascript, and a large percentage of those responding also do unobtrusive javascript.

I call BS on that. Where is the proof of this?

I still see [paraphrased] all over the net, and all over everyone else’s code. It’s a rare thing in 2008 to see a site using a lot of unobtrusive javscript.

Comment by blinkingmarquee — January 12, 2009

my last comment didn’t come across right. meant to say, although everyone says they do, I still see:

[div id=”some_id” onmouseover=”some_func()”][/div]

almost everywhere. Unobtrusive JS is an exception to the rule.

Someone else mentioned there were quite a representation of web designers, not just developers, in the survey. Every web designer I work with (10) does not know javascript.

n=1 as always.

Even with my doubts on the data I still think the survey is a very good thing.

Comment by blinkingmarquee — January 13, 2009

From, they say survey audience come from:
– past conference ( attendees, and
– users of our software Style Master,
– contacted a number of web design and development blogs, sites, and experts
– etc.

This why this survey did not reflect web professional from the ‘third world’, like Indonesia, China, India, etc.

Comment by lokamaya — January 14, 2009

Actually, we did have developers from Peru, Ghana, and China, to name a few – but the point is that this is a self selecting group, within a subculture of “early adopter” web professionals, who to some extent may be giving the answer they think the survey wanted to hear.

Al of that is acknowledged in the report.

As much as anything, the goal of the survey was to identify philosophies and trends. It was subjective, not objective, and deliberately so.

Comment by JohnAllsopp — January 21, 2009

Hello Everyone.

Cost-effective online feedback solutions will provide deeper consumer insight; and can easily be created using a Web survey software solution.

Properly utilizing Web surveys and online survey software, can reduce the administrative costs and complexities; associated with tasks like: developing a customer satisfaction survey, and conducting focused market research.

As well, your Web survey efforts will improve customer satisfaction; and likely increase revenues and promote longevity.

Thank you for sharing your insight and knowledge.

Best wishes.

Comment by WebSurvey — September 24, 2009

Leave a comment

You must be logged in to post a comment.