This entry gives some information about your way of working, your relation to this site, and device ownership.

There’s one restriction on the cross-tabs: at least 100 people must have replied Yes to the first question in order to get a more-or-less representative response. If the total number is less I refuse to publish the cross-tab.

Incidentally, you can create cross-tabs yourself at the site. Play with the data if you feel like it.

Basic

Urtak allows you to create cross-tabs of two questions (and no more than two). My primary interest is to see whether my European and US respondents differed. In the mobile questions I also split up iPhone and Android users.

58% of you live in Europe and 30% in the US — the rest lives somewhere else. 42% of you has an iPhone as primary phone, against 35% Android.

The US has clearly more iPhone users, and Europe more Android users. Still, because more Europeans were surveyed, most iPhone users live in Europe.

(How to read: 56% of those using an iPhone as primary phone live in Europe. 41% of those who live in Europe use an iPhone as primary phone.)

Your way of working

A few questions about your way of working. Americans are more likely to use a Mac and at least two screens, while Europeans care more for Opera and for low-bandwidth sites. HTML5 and CSS3 are now used across the board (81%), and new features are routinely used without a fallback (69%).

72% uses jQuery in most of their projects, while 30% uses another library and 42% prefers native JS. These numbers do not add up. Do a few people use both jQuery and another library in every single project? I hope not. The native JS question asked for preferences, and not actual behaviour, so that might explain the overlap with the other categories.

The last two questions were submitted late in the poll, and haven’t garnered enough replies for cross-tabs.

QuirksMode.org

Then the part that interests me most: how do you interact with QuirksMode.org? What surprised me most is that only 43% follows me on Twitter. Since I used Twitter as the main marketing tool for this survey I expected that number to be far higher. Europeans and iPhone users are more likely to follow me.

Android users are more likely to visit this site at least once per week, while US residents and iPhone users are marginally more likely to donate.

Device ownership

Finally, what devices do people own? Remember, iPhone and Android are those whose primary phone is iPhone or Android, not those who own such a device. Regardless, some people are confused. 2% of those whose primary phone is an Android don’t actually own an Android phone. For iPhone this is 1%. ??!?! (Unless a few people use the device, but do not own it; for instance because their employer bought it for them. Possible, I suppose.)

The geographical differences are clear: Europeans are far more likely to own a Symbian, MeeGo, or S40 phone (figures, seeing Nokia has hardly any market share in the US). Oddly, the same goes for Windows Phone, and I hadn’t expected that.

Americans are more likely to own a gaming console, a Kindle, a webOS device, or a Chromebook. The first and last mildly surprise me, the other two don’t.

iPhone users are more likely to own lots of phones, and generally score higher on all phone types but Symbian, webOS, and bada. Android users are more likely to own a Chromebook.

I have now published all the questions. Maybe I'll create some new interesting cross-tabs later.

My tour will end in Orlando on 16 and 17 April, where I’ll speak at the third Breaking Development conference, Mobilism’s American sister conference. The good Breaking Developers have given me the code ORPPK12 for a $100 discount on tickets, and if you register before end of next week you’ll save another $200. Not a bad discount for a top-notch conference.

That leaves me with almost two weeks in-between conferences. I’m fairly certain I’ll be in Toronto on the 10th, and I plan to visit San Francisco and the Valley at some time (duh), but other than that my itinerary is still unclear.

If you’d like me to swing past your city, let me know. Unfortunately I’m not super-wealthy, so I do expect you to pick up my hotel bill, as well as help me out a bit with domestic flights.

Last year I ended up going to San José, Austin, New York, and Albany. I have no idea what this year will bring, but that’s what makes this kind of touring so exciting.

Anyway, I’ll post an update when I have more news.

All in all support has increased slightly since the last time I tested them, although Safari desktop, Chrome, and BlackBerry have seen some decline. Safari and Chrome have mainly done away with badly or buggily implemented types — it’s clear they’re rewriting significant parts of this module.

Quiz

To understand how complicated these new features can be, consider the following quiz question:

I add an <input type="number" step="3"> to the page, and the user enters 11. What should the browser do?

You can try it for yourself at the test page. Don’t forget to submit the form. My Twitter followers gave they following replies:

* This total includes the votes for an error message.

So the desired outcome is unclear. The browsers’ behaviours have been documented on the desktop page.

Purpose

A quick repeat of the underlying purpose of the input types seems in order. In the end, the idea is to not send incorrect values to the server. Thus we can see three purposes:

1. The browser could perform a validation and show an error message and refuse to submit the form when an incorrect value is found.
2. The browser could also restrict the accepted input — for instance by not allowing letters in an <input type="number">.
3. Finally, the browser could also give the user a nice interface — for instance the various date pickers in Opera, BlackBerry, and Safari iOS.

The last purpose is a nice extra. If the browser offers an interface but doesn’t support either validation or input restriction, I judge its support Minimal. It misses the actual point, after all.

Findings

With that said, my findings were the following:

• readonly is still the only universally supported attribute or type.
• IE10’s support is decent, though not great.
• Opera and BlackBerry added support for color.
• Firefox on desktop has added datalists, but its interface is an autosuggest, and not a complete list of options, as IE and Opera show.
• Firefox Mobile, however, has a very nice interface (see screenshot) that I hope other mobile browsers will pick up.
• Firefox Mobile does not show any error messages, though, while Firefox desktop does.
• Chrome dropped support for the date types. That’s curious, because Chrome 10 supported them reasonably well. I can only assume a major re-factoring is going on.
• Safari desktop dropped support for dates, datalist, email, and url. Since they were hideously buggy previously, this makes sense. I assume they’ll reappear.
• Safari on iOS added support for dates, numbers, and ranges. The interface for the range slider is a bit confusing, though.
• Android 3.2, MeeGo and UC support autofocus, but they cleverly don’t pop up the software keyboard, which would be very annoying. As far as I’m concerned this should become a mobile standard.
• Still, poor Android can’t do anything right. Android 3.2 tried to add support for number and range, but the range slider is invisible, while the number type restricts user input only to actual numbers. No negative numbers; no exponents.
• And then BlackBerry, which in my previous round of testing was the best browser. The PlayBook 2 default browser dropped support for required, pattern, email, url, step, and month (though not the other date types). The old BlackBerry Torch (OS6) supported them all fine. I’m not sure what’s going on here, and I hope the PB 2 final version will restore these types.

For this edition we’re happy to welcome Lyza Danger Gardner, who started as a web developer and moved to mobile back in 2007, co-founding Cloud Four with Jason Grigsby, who’ll also speak at Mobilism.

And it’s in May in Amsterdam, when the city is at its most beautiful. So what’s not to like? Buy your ticket now!

This entry gives some information about the respondents themselves, about how they deal with IE6 and 7, and whether and how they test mobile and tablet devices.

Urtak allows you to create cross-tabs of two questions (and no more than two). For this entry my primary interest was to see whether my European and US respondents differed.

There’s one restriction on the cross-tabs: at least 100 people must have replied Yes to the first question in order to get a more-or-less representative response.

Incidentally, you can create cross-tabs yourself at the site. Play with the data if you feel like it.

About you

Here are the findings about you. The most interesting one is your gender: 95% of you is male. Weirdly 98% of my European respondents is male, against only 86% of my US respondents. That’s a large gap.

Overall, slightly more than one-third of the respondents works as a freelancer, and slightly more than one-third in a medium- to large-size web company. I assume the others work either in companies of 2-4 people. or in the web department of a non-website company.

Nearly half uses Chrome as their default browser (this fits the browser stat data from my site, though the fit is not perfect). About three out of four have their own, private projects next to their jobs. Four out of five consider themselves autodidact.

Differences between Europe and the US: Chrome is used more in the US, US respondents are more likely to have five years of experience under their belt, but Europeans are more likely to be autodidact.

About IE6 and IE7

How many of you still care for IE6 and IE7? Six out of ten of you still test your sites in IE6 and IE7, and about half of you charge your clients extra if you must support these browsers.

However, less than one out of five of you report that IE6 and IE7 together account for at least 1/3 of the traffic to your sites. This percentage is much lower among people who charge their clients extra.

Europeans are more likely to refuse to make their sites work with these older browsers.

About testing on mobile phones

Then we come to the questions that interest me most personally: on what kind of mobile phones and tablets do you test? I split up the replies according to geography, but also among iPhone- and Android users.

In general Europe has slightly fewer iPhone users than the US, but slightly more Android users.

Very generally speaking iPhone users are more attuned to the mobile Web, although in most cases the difference is slight. Only when it comes to testing on two different Android devices do Android users significantly outpace iPhone users — perhaps not surprisingly. Also, Android users use WAY more QR tags than others (23% vs 11% for iPhone).

iPhone users have more experience with Windows Phone 7 and the BlackBerry PlayBook. For some reason this surprised me, but it’s clear now that, at least among my readers, iPhone users are more mobile-savvy than Android users.

The difference between Europeans and Americans is larger. In general, Americans are more of the opinion that only iOS and Android will remain important (56% vs 42% in Europe), care less about testing OSs that are neither iOS nor Android (18% vs 28% in Europe), and test less in Opera Mini (14% vs 25% in Europe). Still, Americans test their sites more on mobile/tablet devices in general (77% vs 72% in Europe) and use more jQuery Mobile (17% vs 12% in Europe).

In any case it’s clear that geography is more important than OS preference for deciding what to test on. This makes sense — the European and US mobile browser market shares are quite different.

The Mobile First question came in late in the survey and too few people answered Yes, so I will not publish the cross-tabs.

To be continued

There are plenty more fun facts to be gleaned from the reader survey. In a future post I’ll give some more numbers and cross-tabs.

While I was at it I also gathered data from StatCounter about the browsing shares of these device classes.

Your donation for keeping this research up and running would be much appreciated.

• Desktop: traditional computers with a hardware keyboard and a non-touch screen. This includes laptops.
• Mobile: devices that fit in your pocket and have mobile connectivity.
• Tablet: devices that do not (easily) fit in your pocket and have a touchscreen but no hardware keyboard.
• Other: other devices. I admit that this is a bit of a cop-out. Right now it consists of the iPod Touch and Internet-enabled PlayStation and Nintendo devices. None of these fall in any of the other categories.

These four categories fall into three markets:

1. Mobile market: mobile. These devices are bought for their mobile connectivity.
   Devices on this market are subsidised by the mobile operators. Everybody needs a mobile device, even if they’re not interested in other kinds of devices or already own several of them.
2. Computing device market: desktop and tablet. These devices are bought for productivity and/or media consumption.
   It’s clear that tablets are becoming competitors to desktop computers. One purpose of this new line of research is figuring out how rapidly tablets are going to replace desktops for browsing purposes.
3. Other markets: other. These are devices that are bought primarily for a goal other than productivity or mobile connections. Examples include gaming consoles or music players. (OK, OK, this remains vague. People can easily buy both a gaming console and a music player.)

In general people buy a mobile device first, then a computing device, then one or more other devices. In the developing world, that usually means people only have a mobile device. Hence the immense change in browsing we’re witnessing.

The stats

With that said, here are the general browsing stats. I use hundredths of percentage points here because some shares are very low. I don’t really like doing that, because it suggests a precisions that StatCounter’s figures do not have, and I’m going to change my methodology later on.

Interestingly, it turns out that the growth of mobile browsing has collapsed in Q4. Where earlier mobile browsing grew with over a percentage point per quarter, in Q4 it grew very little. Let’s see if this is a trend or an outlier before we comment on it.

As to tablets, its rate of growth is itself growing. That’s nothing surprising: the tablet market is growing vigorously, and that is reflected in its browsing shares.

Tablet browsers

Which browsers do tablet users use? Nobody will be particularly surprised to hear that Safari for iPad leads with an immense percentage. Its share is going down, but that’s only because it has very little room to go up.

Keep in mind that we’re using very small numbers here, especially for BlackBerry. If the total share of tablet grows, I expect these numbers to starts to fluctuate.

StatCounter reports slightly less Android browsers than Android OSs. I assume most of these non-default Android browsers to be Opera. The Safari/iOS numbers match perfectly, so very, very few iPad users use a non-default browser.

Classes in 12 countries

Here are the four device classes in my twelve countries. The table clearly shows rather large differences between these countries. Unsurprisingly, developing countries have a high mobile share and a low share of the other three classes.

Findings:

• The highest-scoring country when it comes to tablets is the Netherlands. Yay!
• Tablets score surprisingly well in Nigeria: close to 1% of all devices. That’s a lot for a developing country.
• Tablets score surprisingly lousy in South Korea: it seems Koreans are not taken with the concept.
• Of course the iPad takes 75-95% of the tablet market. I expected this to be different in South Korea due to the Samsung Galaxy being a local product, but that doesn’t really seem to be the case.
• The bulk of the Other category is made up of the iPod Touch, as one would expect. The only exception is the UK, where the iPod Touch takes only about 40% of the Other category.
• Mexico scores very well in Other. That’s mainly due to the iPod Touch, which is used as much as the iPhone. In other countries iPod use is well below iPhone use.

Methodology

• StatCounter gives desktop and mobile stats, and also states the share of these two classes.
• Other: desktop PlayStation or Nintendo browsers, as well as mobile iPod Touch and Sony PSP browsers. I took this data from the desktop and mobile browser stats.
• Tablet: “desktop” devices running on iOS, Android, or BlackBerry. I took this data from the desktop OS stats.
• Mobile: all mobile devices that are not Other.
• Desktop: all desktop devices that ar not Tablet or Other.
• For the tablet browsers I compared the desktop OS (iOS, Android, BB) to the desktop browsers (Safari iPad, Android, BB). Any difference between the two was assigned to Opera (and other).

Unfortunately the story isn’t true, and I suspect it’s a good example of headline grabbing aimed at those who don’t have deep knowledge of mobile browser statistics; a group that seems to include the entire tech blogging caste.

Java ME is a middleware layer that’s supposed to make creating a Java-based app for all OSs supporting it easier. Java ME’s big problem is that it’s not a case of write once, run anywhere, but more like write once, and make it marginally easier to write it elsewhere, too. Still, the relative ease of writing apps for Java ME is not the issue here.

Java ME (or rather, an OS- and device-specific version of Java ME) comes pre-installed with Symbian, BlackBerry, Windows Mobile (and, I think, bada), as well as a host of proprietary feature phone operating systems from Nokia, Samsung, Sony Ericsson and the like.

So it could mean that the combined market share of these operating systems is growing. Unfortunately Net Market Share does not mean that: it counts Symbian and BlackBerry separately.

Besides, the combined tech press completely overlooks the fact that Net Market Share’s own numbers clearly show Java ME has lost market share last year. Java ME had 26.80% market share in February, and 21.27% in December. A complex calculation based on my secret and proprietary formulae shows that it lost 5.53% market share this year. Hardly a winner.

Finally, what Net Market Share measures are browser impressions on its member sites. So which browsers do all these Java ME users use? If you compare the OS stats to December’s browser stats it clearly emerges that the Java ME share coincides fairly exactly with the Opera Mini market share.

What is going on is that Opera Mini is slightly on the rise (a trend confirmed by the latest StatCounter numbers), and that Net Market Share counts most of these Opera Minis as running on Java ME.

Now it is notoriously difficult to read out the operating system from Opera Mini’s browser string, and it is true that Java ME is often mentioned. Clearly, Net Market Share has decided to cut some corners, forget about detecting the true OS underlying these Opera Minis, and call them all Java ME. (Compare this to StatCounter, which at least tries to get the true OS right, and says “Unknown” if that’s impossible.)

So all that Net Market Share’s numbers tell us is that Opera Mini is on the rise again. Good news for Opera, but it isn’t that much of a story. Java ME replacing Android sounds much better, and will grab headlines all over the world.

I wish tech reporters would actually study the numbers they report on instead of copying press releases.

Your donation for keeping this series up and running would be much appreciated.

Mobile

Opera once more overtakes Safari. It’s clear now that Android’s untrammeled growth has ended, and that the race for first position will continue to be between Opera and Safari.

Desktop

As was already known, Chrome has passed Firefox as the second browser. This change seems to be permanent: Firefox loses terrain every month, and Chrome wins it. Meanwhile IE is still going down, too.

Q4

Mobile

In the quarterly overview we can see the most important changes in the mobile browser market. The top two are still Safari and Opera, with now the one then the other leading.

Just below the top two there has been a change. Where at the start of the year it was a very close race between Android, Nokia, and BlackBerry, the three have drifted apart decisively, with rather large differences between them. I expect that this won’t change any time soon.

StatCounter’s detect was changed in two important ways during this year:

1. In early Q2 it introduced the Dolfin, Obigo, and Jasmine browsers as separate entries. I documented this change.
2. In early Q4 it changed its detection for UC, the Chinese proxy browser. Turns out roughly 4% of the entire browser market was counted as Nokia, but should have been counted as UC. This accounts for a large part of Nokia’s drop.

Desktop

The quarterly desktop table shows that the trends we already noticed have been going on for the entire year. IE reliably loses 2 points per quarter; Firefox rather less, but it also loses. Chrome is surging.

Year

Mobile

The yearly table reminds us how far Safari used to be in front of all other browsers, including Opera. It also shows Android’s surge, and BlackBerry’s rise and fall.

Desktop

The yearly desktop table shows Chrome’s surge and IE’s fall from grace clearly. It also shows 2010 was Firefox’s heyday.

My hypothesis is that Google’s acquisition of Motorola will make the other Android vendors look for other OSs to use. Google is now a player on the device market instead of just a software house, and Samsung, HTC, and the rest will start to think seriously about using another OS for their future devices.

Although this is still a very real problem for Android vendors, the lack of an alternative complicates matters considerably. Back in August I expected that webOS would be sold by HP in October at the latest to one Android vendor (privately I bet on HTC). Once that happened, other Android vendors would also start to cast around for alternatives.

The curious case of webOS

That did not happen. webOS will never be sold; instead it was open-sourced. This confused me mightily.

What happened? It turns out that HP asked $1.2 billion for webOS, and that’s way too much for an OS that has failed twice and has to be re-customised for new hardware once it’s been sold. Besides, it was reported that webOS was built on flawed software.

A third problem might be that HP demands too much influence on potential licensees, trying to position itself as a smartphone ecosystem curator like Apple and Google. There have been no reports on this (and it presupposes licensing instead of sale), but it’s something to keep in mind. Of course HP, not being a software house, has no chance of actually curating a succesful ecosystem, but that won’t stop it from trying.

On the other hand, Robert Cringely is positive, quoting especially webOS’s versatility. (Besides, he feels a truly open-sourced webOS would give HP the chance to build good hardware and compete with IBM and Oracle. A dinosaur competing with other dinosaurs will have zero impact on the smartphone market, so I ignore this angle.)

Update: See also Michael Mace's take on Palm: it did not have a redeeming killer feature that kept clients happy while the kinks were worked out of the system, and Palm's pockets weren't deep enough to keep it alive for a generation or two.

Other Android alternatives

The webOS debacle teaches us a valuable lesson: an OS is worthless without a device vendor making a major commitment to produce phones and push them to the operators’ sales channels.

Right now webOS does not have such an interested party, so its chances are bleak. On the other hand, this could change pretty quickly once a major vendor announces its intent to use the OS.

This device vendor intent is the most important feature as-yet lacking for two other potential Android replacements: Boot to Gecko and Tizen.

Mozilla people told me they’re working with a device vendor, but quite properly declined to name it. Samsung is Tizen’s co-chair, so one would assume it’s planning to take Tizen devices into production. Still, we’ll have to await more specific news before we can be certain that these OSs will become a significant factor for Android vendors.

I expect more news around Mobile World Congress, 27th of February to 1st of March. What we want to hear is that a vendor commits itself to one of these OSs.

The post-Android market

So what’s going to happen on the post-Android market? To be honest I don’t dare to make a prediction right now. Weasely, I know, but there you are. Let’s say I’ll return to this problem after MWC.

There’s no denying that Google has had an incredible stroke of luck when its insanity in buying Motorola was offset by HP’s insane handling of webOS. Android’s most direct competitor has disappeared in a cloud of vagueness, and the chance that Android vendors will defect from the flock has been diminished.

The best way of viewing the post-Android market is that all Android vendors are weighing their options, and that those options include Android. So if a vendor decides to continue with Android, that means it sees no better option right now.

That last occurrence would be bad for the smartphone market as a whole, I feel. It’s better to have healthy competition than to have one single OS occupy more than 50% of the market.

So let’s hope webOS, Boot to Gecko, and Tizen, all of which are HTML5-based, will be picked up by post-Android device vendors.

Mobilism 2011 was a blast; see also the videos. We’re set to repeat and extend this success in May. with seven world-class speakers, including Ethan Marcotte, Jeremy Keith, Horace Dediu, and Jason Grigsby, who’ll show you how the mobile web works and how you should adapt to it as a web designer or developer. What’s not to like? And even more speakers will be announced later.

See you at Mobilism.