Archive for January 2006

IE7 beta2 preview

Just tried to install the new Internet Explorer 7 beta 2 preview, with the following result:

IE7 failing to install

Oh. I didn’t expect that. Maybe Windows 2000 is beginning to get just a little long in the tooth…

Card sorting: xSort in action

xSort is a card sorting application for OS X. Or “Card sorting, the Mac way”, as the xSort website likes to put it. I’ve known about xSort for a few months now, but finally used it in anger for the first time in a card sorting session this week in a project I’m working on for cxpartners.

In short, I was really impressed with it. It’s very simple to set up a test session – you simply give the project a title and say whether you want the sort to be open or closed. You can then define the ‘cards’ and add details of the people who’ll be doing the sorting. In the sorting session itself, you just hit ‘exercise mode’ and you’re away.

Setting the initial options in xSort
Setting the initial options in xSort.

The biggest concern I had was the move from paper to screen. Paper-based card sorting is a tried and tested technique in IA, and people like to physically touch and manipulate objects. However, I was genuinely surprised as to how easily the users took to xSort. All that was needed was a very brief tour of the functionality at the start and the occasional reminder during the test. No worries there.

Doing the card sort in exercise mode
Doing the actual sort in ‘exercise mode’.

Also, running the session was a breeze. With card-based testing, you’re up to your eyeballs in piles of cards, paper clips and staplers. You’re also worried in case you sneeze or drop a pile of sorted cards. And then there’s the filing you have to do to keep everything safe before you start the monumentally tedious job of analysing the cards. None of this is a problem with xSort. No mess at all – it’s all set up for you. It even analyses the results for you as you go along. You can also export the data to an XML file or produce your own reports.

Generatiion of a dendrogram in xSort
Displaying the results: a dendrogram in xSort.

The downside? Well, you’re limited by the size of screen you have. I was testing on an iBook using around 35 cards, which was pushing it a little. While you have, in theory, a limitless ability to scroll horizontally and vertically, in practice you want all your cards to be visible at the same time. A 19” or 20” screen would be essential if you were testing with more cards.

However, the biggest negative of my experience with xSort was when it crashed towards the end of the very first session when the subject tried to remove a leftover grouping card. Although my heart sank at the time, it actually wasn’t too much of a problem; it’s the thinking bit that takes the time in card sorting and it only actually takes a couple of minutes to start again and re-sort the cards once you’ve decided where things go. It helps to have two people to remember where the cards were too. This is not something that I’d like to have do again, though.

I’d love something like this to be available for Windows. I’m happy to use a Mac, but 90% of my work is still done with Windows. There are some options: there’s uzCardSort from uZilla, which is a card sorting application that runs in Mozilla, and also Card Sword which is an open-source Java app. Neither of these even remotely come close to the elegance and simplicity of xSort for card sorting.

Maybe it’s no coincidence that there were two user experience designers behind xSort and only one programmer. I think I’ll just have to do card sorting on my iBook from now on…

Armando Iannucci on interactivity

While browsing through some old box files, I came across a Times article that I’d cut out by comedy god Armando Iannucci (of The Day Today and I’m Alan Partridge fame). Luckily, the article is still online. It was the subtitle of the piece that made me save it in the first place: Interactivity is a superficial sham leading only to hunger and emptiness.

The article’s main point is a political one but, as ever, it’s full of wit and insight as well. He even makes reference to interaction design:

Meanwhile, in the Natural History Museum I had my revelatory moment. I, my wife and children were all suddenly struck dumb by the futility of it all. Not the stuffed giraffes and elephants, which were terrific and, as far as we were concerned, suffered an extremely informative death. No, it was the relentless, raucous, lapel-grabbing interactivity of much of the newer displays that was so off-putting. You couldn’t walk a centimetre without a bright neon sign urging you to push a button to see how many farmers Mount Etna fried a year. Turn a corner, and a yellow motorised bumblebee the size of a fit gundog flapped down on metal chains and insisted that you find out how much honey was produced in Guatemala by shouting the word “buzz” up its fundament.

The sum result was a quiet disappointment on the faces of the children doing all the pushing and popping. The thwarping display cases invited participation in a process, only to lead you to the conclusion that the process was not worth pursuing. A large talking ant-hill urged you to spit on a glowing red saucepan to find out whether ants masticate grass. That’s exciting. So you spat on the saucepan, and a small sign lit up, flashing, “Yes. They do.” That’s dull.

Well worth a read (as is his book, Facts and Fancies).

Web design feedfest

I’ve put together an OPML file of all the web design feeds that I find really useful. You can download it from the Resources section of the site.

Words influenced by pictures

A recent research article in Cambridge Assessment’s Research Matters by Victoria Crisp and Ezekiel Sweiry suggests that a picture can ruin a thousand words (PDF, 4625K) when it comes to exam questions. The authors found that the information contained in images or diagrams can influence students’ interpretation of exam questions.

In a way, I guess this isn’t that surprising. I’ve encountered something similar in think-aloud user testing of websites, particularly in sites where the information scent is poor. In the absence of clear signposts, users look for whatever information they can extract from the interface, and this includes clues from images. Images are usually visually dominant on a page (and they’re much more interesting to look at than all that boring text!) and therefore have a potential priming effect on user’s subsequent interpretation of information.

This also has some important implications for the development of e-learning materials. It’s a common (and understandable) strategy in the production of e-learning to avoid the dreaded “wall of text” syndrome by using images wherever possible. This is often justified as a good learning strategy by reference to different learning styles and multiple intelligences. However, often these images are decorative (in the sense that, although they relate to the subject of the learning, they are essentially information-free). For a ‘visual learner’, is it possible that decorative images could actually serve to distract the learner from the main point of the learning?

More web standards articles than you can shake a stick at

A fabulous Web Standards Link Bonanza from SimpleBits.

Oh, and if you’re interested here’s the etymology of shake a stick at.

Next Page →


I'm Stuart Church, a user experience consultant with Pure Usability in Bristol, UK. Sensorydrive is my personal blog and covers user experience design, information architecture, product design, psychology, research methods, perception and pretty much anything else that takes my fancy! You can find out a bit more about me if you want...