This website is now archived. To find out what BERG did next, go to

Making Senses revisited

Adaptive Path kindly invited me to their offices this morning, where I muddled through my Making Senses talk, on using the human senses as inspiration for next-generation Web browser functionality.

Optic flow

Revisiting the slides, and the conversation afterward, has shown me how to state the argument more directly:

  1. As far as interaction on the computer desktop and the Web goes, navigational and spatial metaphors dominate. On a micro level we talk about direct manipulation of files via icons: Dragging, moving, opening and so on. On a macro level, we have addresses, visiting, and sitemaps.
  2. When a person has navigated to something, they can know what it is because of the navigation itself. For instance, you know you’re in London because you followed all the signposts to London.
  3. In a world of cheap sensors, many, many display surfaces, and high connectivity, we are presented with information without that navigational context. Furthermore, in areas which have traditionally used the navigational metaphor (mainly the Web), navigation might not be the most appropriate approach to reading the news, buying books, or hanging out in chatrooms. Yet still we approach Web design armed with this metaphor.
  4. It’s as important that a thing can be instantly appreciated for what it is, as that it can be navigated to. ‘Instantly appreciate’ means comprehend pre-consciously, just as we instantly appreciate a chair as a chair when looking at it, without having to deliberately deduce the meaning of the pattern of light on our retina.
  5. As a guide to what qualities we should be able to instantly appreciate, we can use human and animal senses to show what features we need to recognise of things in the environment. Sensing these features is sufficient to let us intelligently interact, without navigating.
  6. To summarise these features, we need to be able to detect: Structure, focus and periphery, rhythms of activity, summaries, how this particular thing is situated in the larger environment (and more). The Web browser, as our sensory organs online, should do this job, instead of leaving it to the websites themselves.

Applying the sensory model to Web design triggers a few ideas:

  • The high-level structure of all sites should be represented by the browser in a consistent way, not by each site differently.
  • Regular patterns in browsing (such as the sites visited daily or weekly) should be supported by the browser.
  • Using the extracted keywords of a web page as its ‘scent,’ hyperlinks should indicate how their odour strengthens or detracts from the smell of the current browsing trail.

There are more ideas, but that’s what the presentation discusses and illustrates.

Incidentally, the image at the top of this post is from J. J. Gibson’s The Ecological Approach to Visual Perception, which talks about how we see continuously and actively as we move through space. I’d like to consider browsing the Web in the same light.

One Comment or Trackback

  • 1. minntc said on 8 December 2006...

    Regarding Gibson’s image: “I’d like to consider browsing the Web in the same light.”

    I agree. In fact, the first time I saw this image in your blog post, I assumed it was in reference to the nature of the Internet. I’ve actually drawn images similar to this (and seen them drawn, though I can not provide a source) when explaining how communication over the Internet, p2p, etc. works.

    Instead of following roads, signs, paths towards London around the surface of our spherical Internet, we can simply transport, [hyper]jump, hyperlink to London. Only by monitoring multiple jumps and tracking the path taken can you reinject a sense of journey. Trouble is, no matter how you orient all the sites on the Internet, your browsing history is going to look like a rat’s nest, with a few branches or paths showing frequent use. I think your discussion of senses serves to provide a way to reorient your organization of sites on this spherical Internet so a more linear path (or sets of paths) comes to life.

    By taking your “map” and overlaying it with mine, or another’s who hits many of the same stops, you can build clusters (neighborhoods?) to interrelate. I see as an attempt at correllating these maps, but it’s rather difficult to do in a tag cloud. Too verbose, not visual or easily navigable. If you move from one person’s links to a tag cloud, you get lost in noise immediately. (Sure it can make for enteraining browsing, but not necessarily relevant.) I think you touched on this correllation when you mentioned in your notes “reacting to what someone else is looking it”, which you said you’d return to but did not. This, of course, is what marketers are trying to do every day with research groups, studies, etc.

    Interesting talk you gave, from the sound of it.

Comments for this post are now closed.

Recent entries from
Matt Webb

Popular Tags