Initial ideas were based on the idea of evolution and how it relates taxonomy in biology. The theory was that sites had evolved over time into different forms that suited their audiences. The “best” forms lived and prospered; audience numbers and revenue raised measured success. Successful ones spawned multiples, variants or developed into similar formats independently, much like evolution. The unsuccessful died.

If the Internet behaved like the biological world then perhaps we could categorise websites like the Linnean taxonomy in biology (Kingdom>Phylum>Class>Order>Family>Genus>Species). However, a branching categorisation system didn’t work. There were no divisible characteristics that create categories in Biology, so it turned out to be a bit of a false start. And websites don’t breed so we can’t draw a clear relationship between them.

So starting again, I thought that the task seemed daunting because of the idea that a complicated thing like the Internet needs to be looked at in an equally complicated way. But what biology lends is not a categorisation system, but a guiding principle – that very simple rules lead to great complexity. In biology the theory of natural selection leads to evolution, the ultimate outcome being the complex natural world we live in. The hard part about a simple premise is figuring it out.

The first relevant thing that sprang to mind about the Internet is that it must have a reason to exist. History says it was because people wanted to share academic papers by publishing online. The good part about that was that other people wanted to read them.

This is a task-based way of looking at things – which is the way that I work as an information architect. To create a website you would create a list of tasks that need to be possible. Things like downloading a PFD or finding some content. This is used to create structure and functionality. If tasks define websites, and websites are small structures appear within the larger form of the Internet, then perhaps its reasonable to think that its structure is defined by the tasks that people want to do within it.

With this in mind, I started considering if you could define the most fundamental tasks; the ones that shaped what we see online today. I came up with six and then added one with the help of Simon Carter, our Consumer Researcher.

• Find
• Learn
• Trade
• Share
• Store
• Talk
• Play

These task names refer to specific things that make up a restricted vocabulary where each item has a definition and examples. This vocabulary covers off things from a scientific/methodological perspective – everything has a definition. It also ensures that we can talk to each other in a language that we all understand.

This restricted vocabulary also ignores a certain philosophical debate (for those of you who are unfamiliar with French post structuralist linguistic theory, please ignore the following). The debate is about the difference between the sign and the signified and whether or not meaning is being constantly deferred in an attempt to define it. I say, that’s fine and it’s a fun debate, but it doesn’t really affect our ability to make a definition in the sphere of working in the Internet biz.

The list also allows categories to be blurry. They are fuzzy as opposed to mutually exclusive. Strict categorisations that we keep trying to apply in things like biology and library categorisations keep failing because there are exceptions, or new things come along that won’t easily fit into a scheme. So in this case its best not to try. That’s a bit heretical in terms of “traditional” information architecture, but the rise of faceted categorisation and tagging systems show us how exclusivity, (at least in terms of categorising content) is flawed. A fuzzy categorisation allows us to see that websites serve multiple tasks and allows us to create a task matrix – which is how the following was developed.

The matrix is based on the idea of being able to develop a rating or ranking system for websites. It would be based on audience surveys with a particular question-set about a website and tasks. The results mapped to the diagram show us a “task-space” for specific websites. This would let us:

• Make comparisons between sites.
• See differences between what audiences want to do and what we let them to do (when we make websites or applications for them).
• See if an audiences’ wish to do certain tasks matches with existing digital properties (like Flickr or YouTube or Facebook) – in doing this it might be possible to locate the best destinations to create experiences for them.

The next step is to test our model and see how it works, much like the application of the scientific method – to put the hypothesis and try to disprove it. Figuring out whom to ask what questions is the next tricky part; although Simon does assure me that he has a theory about how we can make a system work.