Yes/No/Cancel

A while ago I wrote some posts about the user interfaces of ticket machines in Germany (article 1, article 2). Meanwhile I am told that they have been improved considerably: the ‘Fast purchase’ route is now considerably faster, requiring a minimum of only 4 or 5 clicks to buy a standard ticket (compare that to 16 clicks previously!). The way they have done that is to skip the whole timetable thing; instead you only select whether or not you want to take the fast trains (which has an effect on the price). That’s a very good start, since it optimises the common case: people who routinely buy the same ticket and know exactly what they need. And for those with unusual requirements, there’s still the long route with its multitude of different options to choose from.

Despite these changes, plenty of usability challenges remain. For example, my friend told me that he didn’t realise when using the machine when he had reached the payment screen: he could have just inserted his card, but instead found himself looking around for the “next” button to press. There was just some small and non-obvious bit of text on screen explaining that you were now ready to pay.

In fact the usability problems of German train ticket machines are still so pronounced that the national rail company (DB) is now offering courses to teach people how to use them. (See the scanned newspaper article, taken from Aalener Nachrichten/Schwäbische Zeitung, Tuesday 18th December 2007. Sorry that it’s more than a month old, I’ve not had much time to blog recently.)

This article is somehow slightly scary and hilarious at the same time, in the way how the train staff systematically blame the users for their inability to use the system, rather than seeking the blame with the system itself. Hilarious because it’s so stereotypical, and scary because such a big organisation can get away with it without people putting up a fuss and explaining that this is just not acceptable.

Some highlights from the article:

Moschner [the course instructor] says that the new ticket machines have a more visible display and also accept cash besides credit and debit cards. “Are they just as cumbersome as the old ones?” an over-70-year-old lady enquires. The course instructor remains calm: “They are not cumbersome.”

Hmm. Complete denial of the existence of problems. Two more quotes indicate that there is a fundamental misunderstanding of user behaviour going on:

“Read what it says there. It is important.”

“The ticket machine really does tell you what it wants, you just have to look.”

Why should I be trying to find out what the ticket machine wants? It should be trying to find out what I want! Also, I shouldn’t have to read every word on the screen. That’s simply not what people do. People don’t even read whether doors are labelled ‘PUSH’ or ‘PULL’ before trying one or the other. People just press random buttons in the hope of getting somewhere quickly, and the system should be designed to cope with this sort of behaviour. Anything else is just unrealistic and designed for robots rather than humans.

Fortunately this course is a positive initiative, probably with a thought along the lines of “well, if we can’t get the design right, at least we can teach people how to use the broken design”. I guess that’s a valid approach to the problem. And hey, 8 people turned up to that course, maybe that’s 8 fewer people who get frustrated with the machines. Sounds a bit like a drop in an ocean to me though.

Netlife Research, a usability consultancy from Norway, has come up with a neat humorous way of pointing out some design elements which can help improve usability (or rather, design errors which can render a product pointless for most of its potential users). They have put these handy hints together in the form of a calendar — the Bad Usability Calendar. You can download it and print out a copy for yourself. Such as I have done, see the photo (which includes a gratuitous reference to potted plants). Thanks to Johannes for pointing it out to me.

Amusing though the exercise is, it proves once again how hard it can be to practise what you preach. The Bad Usability Calendar website, although apparently designed by usability experts, doesn’t actually work. I was going to enter their prize draw for an Amazon voucher, for which a required step is to provide your address. This address appears to be used to look up your longitude and latitude, so that they can plot a little pin on a map. I tried four or five variations of my UK address, but unfortunately none of them was accepted by the site. It failed with a badly written error message which gave me no clear indication as to how I would have to construct the address so that it would be accepted. Well, I would have even been perfectly happy to find Cambridge on the map myself and stick a virtual pin into it, but no, that wasn’t a foreseen option. Moreover, the option for uploading a photo didn’t appear to work either — and there wasn’t any error message at all, just no picture. So unfortunately I was excluded from the prize draw.

At least I’m glad to see that nobody else from the UK has managed to place a pin yet. You know, bad usability always makes you feel stupid (even if it’s clearly somebody else’s fault, not your own), so it’s a bit of a consolation that nobody else has figured out how to make the site accept a UK address. Now I am wondering whether something like that could be incorporated into other products. Some means by which users can see that they are not the only ones who are grappling with a dysfunctional product, maybe by social network or something like that. Not that it makes the product any better; it just makes its users feel slightly better.

Edit (29 January 2008): Meanwhile the issue has been sorted out and we have been placed on the map manually. The address search also seems to be working now, and we’ve even been joined by another UK pin in London. Netlife handled the matter very quickly and nicely — thanks!

I just got back from a garden centre, where I had ventured on the exciting mission of getting some indoor potted plants to make our office nicer. So there I was, amidst a cloud of greenery, not having the foggiest clue what plants may be suitable for our office space (with good artifical light but not much natural light, and more importantly, with somebody taking care of them whose track record with plants hasn’t exactly been glamorous).

I couldn’t spot any shop assistants nearby whom I could quiz. There were little labels on the pots which detailed the required conditions, but I wasn’t sure I could believe them — they looked very generic, most were very similar, and sometimes two pots with plants of the same species had different labels, contradicting each other. They looked very much as though they had just been distributed willy-nilly without any regard for the actual needs of the plant.

Enter the mobile web. Fortunately many of the plants were labelled with the name of their species (except for those which were helpfully labelled “foliage plant”), and I had my phone with me. Casting those names into Wikipedia on my mobile quickly gave me a good idea which plants were likely to survive my “care”. Now the mobile web is happy because it has solved somebody’s problem; the garden centre is happy because they were able to sell me plants without even needing to employ a shop assistant; and I’m happy because Ept Computing is now a nicer place to work. We’ve even got, believe it or not, a corporate watering can. (Somehow this reminds me of Douglas Adams and towels.)

It’s nice to practise what you preach and show that mobile web access really is useful in everyday life.

Over the Christmas holidays I was reading Joel on Software, the book summarising some of the most interesting material from Joel Spolsky’s blog. (The book is worth reading, although I did find it quite a shame that it was pretty much verbatim the blog contents pressed on paper for easier reading. It would have been nicer if the writing style had been changed from the slightly rambling, disconnected style of blogs to a more coherent style expected from a book. But the stuff Joel talks about is definitely worth reading for software engineers, in whatever form.)

The article which I found most interesting is his “Strategy Letter V” (page 281), which is also available on the web. It explains why, in his opinion, so many large companies are investing in open source software.

On the surface, open source seems a strange model for a business — why should a company spend a lot of time and money developing software, and then simply give it away? The claim that they have suddenly given up on capitalism isn’t exactly convincing. The claim that it’s cheaper from them to get free code contributions from teenagers than to write it themselves… not so sure about that one either.

Joel gives the first answer which I actually find convincing. He explains open source investment in economic terms, through so-called complements. For example, flights to Venice and accomodation in Venice are complements of each other: customers need both in order to get a holiday in Venice, but they are sold by completely different companies. And if flights to Venice get cheaper, more people want to go there, so there is higher demand on accomodation, so prices of hotel rooms in Venice go up. And vice versa. This economic effect of complements has been observed in many different markets.

So, if A and B are complements of each other, and the price of A goes down, then the price of B will go up. So, if you are a company selling B, and you are clever, you will try to push the price of A down as far as possible, even commodify it. That way, you can sell B for a higher price and you’ll be better off.

And now if you look around who is investing in open source software, you’ll notice that often the software released in this free manner is actually a complement of what that company is trying to sell. For instance:

• Google want to sell advertising on mobile web sites. Mobile web browsers and mobile operating systems are complements of mobile web sites, so Google make Android and release it freely in order to drive down the price of these complements.
• Our friends at Collabora are paid by Nokia to work on an open source platform for Nokia’s internet tablets. Nokia sell phone hardware, and the operating system is a complement of the hardware, so it makes sense for Nokia to commoditise it. Moreover, third-party applications are a complement of the hardware, so by opening the platform to the wide variety of freely available Linux software, Nokia increases the value of its hardware even more.

Once you think about it this way, it’s amazing how the economics begin to make sense!

Looking back at 2007, I can’t help but think that the iPhone was probably the most important and influential technology phenomenon of the year. It’s been talked about so much that I’m actually starting to get sick of it. But that doesn’t change the fact that it has had a significant impact, particularly on the way the mobile web works.

One of the most striking developments that ensued are the moves by several popular web sites to provide versions of their offering which are specifically tailored to people visiting their site from an iPhone. Amongst others, I found:

• Amazon.com
• Facebook
• Ta-Da Lists
• Meebo
• Netvibes

(Most of these sites show their iPhone look only to web browsers which identify themselves as Safari Mobile. To test them in a desktop web browser, see this article.)

iPhone specific development is fashionable, it seems. Some of those sites actually borrow and incorporate iPhone design elements (such as the style of lists and tabs, animation and icons), further blurring the distinction between web sites and applications. A List Apart, a widely respected resource for web developers, has even published a long article on how to develop iPhone-specific web sites (Part I, Part II). Apple themselves also offer in-depth information. iPhoneApplicationList.com maintains an extensive list of iPhone-optimised web applications.

Christopher Schmitt argues that making websites accessible for people with disabilities would expand a website’s reach far more than making an iPhone-specific site does. And Scott Gilbertson of WIRED thinks that the current situation is very similar to the old days when people were designing web sites specifically for Internet Explorer 4 (which was ahead of Netscape at the time). Jeff Croft suggests the opposite, saying that device-specific application development is going to happen anyway, and it doesn’t really matter whether it uses web technologies or not.

iPhone market share

What is the market share of iPhone internet browsing actually like? Hard to say, because it depends so much on what you measure. In terms of sold devices, the iPhone is performing strongly, but it is still a small proportion of the overall handset market. There are at least 100 million mobile devices with modern browsers (Opera Mobile, Nokia’s S60 browser, both of which are capable of displaying standard desktop web pages) compared to an estimated 2 million iPhones. Phones with WAP/XHTML browsers or adaptation browsers (such as Opera Mini) are a lot more numerous still. So in terms of the number of devices, writing iPhone-specific sites (as opposed to Opera-specific sites, for instance) really doesn’t make much sense.

Hitslink/Net Applications’ operating system statistics show that in December 2007, iPhone and iPod Touch users accounted for 0.14% of web page views, more than all other mobile platforms put together (Windows Mobile: 0.06%, S60: 0.02%). The Register reported this too. However, Ray Anderson, CEO of Bango, points out that these figures are misleading:

Hitslink produced this report by using data from their analytics customers. These people operate HTML/PC websites. They say: “You simply paste a small piece of HTML code on each page you wish to track statistics on”.

The 300 million or so mobile phone browser users (say 50 million S60) can’t or don’t browse HTML sites. They browse mobile friendly sites (WAP or XHTML or iMode) which will not have this HTML code in.

Considering that organizations like Bango, Admob, Peperonity, Vodafone report mobile browser traffic in the billions of pages per day, most of thse being S40 or S60 its clear that by ignoring non-HTML sites these stats are misleading and mistaken.

– Ray Anderson, in an email to Mobile Monday London mailing list, 5 Dec 2007

In other words, all that Hitslink’s statistics say is that iPhone users are more likely to visit more desktop web sites, but it doesn’t say anything at all about usage of sites which are specifically designed for mobile use. The conclusion I draw from this is that an iPhone user is much more likely to visit desktop sites than Windows Mobile or S60 users, and therefore non-iPhone users either mostly use sites designed for mobile, or don’t use the web much at all. This may have a variety of reasons — due to its large screen, desktop web sites are more usable on the iPhone than they are on devices with smaller screens; Safari Mobile has pretty neat zooming capabilities; and maybe iPhone users simply approach the device with a more web-oriented attitude, because after all it is more of an internet tablet than a phone.

If iPhone users actually prefer desktop-style websites, because they work quite nicely on the large screen, it really doesn’t make much sense to design a specific iPhone version of a site. On the other hand, if it’s that increased usability of the iPhone web experience which drives web usage per person to be many times higher than on S60 and Windows Mobile devices… then there’s a very strong case in favour of designing device-specific sites.

Is iPhone-specific design just a case of companies wanting to look cool by having an iPhone-optimised site and jumping on the bandwagon? Is this just a fashion which will go away again as quickly as it came? Or do such companies actually derive significant benefits from iPhone users? I’d be interested to hear your comments.

There are a number of web sites out there which provide specifically optimised versions for the iPhone. I was curious to test them (and to look at their source code to see what they are doing), but don’t have an iPhone myself. Many sites will only give a visitor the iPhone version of their site if the web browser identifies itself as Safari Mobile. How do you get it?

The solution is the “user agent” — a string sent by the web browser to the server as part of every request. It contains the name and version of the browser software you are using, the operating system, and a few other bits and pieces. It’s a very useful piece of information to website administrators, who can use it to compile anonymous statistics about the people who visit their site.

Many people consider it to be bad practice to serve different versions of a site depending on the user agent, but it happens often enough anyway. And that’s exactly what is going on here. Fortunately there are tools which will let you modify the user agent, so you can see what you would get if you were using some other software. This is sometimes called “masquerading” as another browser. The technique described here is for Firefox, but it’s possible to do the same thing with other browsers too.

Download the User Agent Switcher add-on for Firefox, and restart Firefox. In the menu, go to Tools -> User Agent Switcher -> Options -> Options. Add a new user agent, with description “iPhone”, and the following entry in the user agent field:

Mozilla/5.0 (iPhone; U; CPU like Mac OS X; en) AppleWebKit/420.1 (KHTML, like Gecko) Version/3.0 Mobile/3B48b Safari/419.3

The remaining fields (app version etc.) can stay empty. Now you can click Tools -> User Agent Switcher -> iPhone, and your browser instantly “becomes” an iPhone. If the site uses features which are not available in Firefox, it will not render correctly, but at least the site should serve you the same content as it would do to an iPhone. (The user agent above is taken from a real iPhone; there are probably many others which work too, but that one has worked for me.)

One big caveat: you shouldn’t really be doing this! Use it only briefly for testing a site, then reset the user agent to the Firefox default. Otherwise you’ll end up sending the iPhone user agent to all other web sites you visit too, and that isn’t good for anybody. You may up being locked out of certain web sites or getting the wrong version, and administrators of web sites will hate you because you mess up their statistics.

So please, please reset the user agent to the default when you’ve finished testing.