4 November 02003

#2 Past Projections of Future User Interfaces

still image from Starfire video, copyright Quailwood designThe Starfire video prototype, produced by Bruce Tognazzini and colleagues at Sun Microsystems, was explicitly conceived as a kind of response to Apple's Knowledge Navigator (described in my posting a couple of days ago).

But Starfire set itself a harder challenge by focusing on a scenario exactly ten years in the future (Knowledge Navigator was set some unspecified date at least twenty years in its future). Nine of those ten years have now elapsed, so the chickens are on their way home to roost. Starfire shows an even clearer case study of social and economic factors having hard-to-predict but easy-to-underestimate influence in shaping the development of user interface technologies.

The Starfire video is not available online (or offline at the time of writing), but you can read the full, annotated script, see a short clip, and read an account of the thinking behind it. Tognazzini is an entertaining and punchy writer, by turns ebullient and self-deprecating, so if you have time I'd also recommend Tog on Software Design and the earlier Tog on Interface. Unfortunately my copy of Tog on Software Design is inaccessible in a packing case at the moment, so I can't refer to it!

In his case history of the production of Starfire, Tognazzini notes how some aspects of the Knowledge Navigator scenario — such as the anthropomorphic interface agent I railed against previously — 'may not be approachable for 100 years.' He stresses that the Starfire team continually questioned whether each object or action could be accomplished in ten years on a real computer, and they based their work on technologies that they could either build in 1994 or that were well launched in laboratories around the world at the time.

But it seems pretty clear that only a small proportion of the hardwire or the software in Starfire will be widely in use twelve months from now.

  • The large wrap-around screen, with gestural input and scanning technologies built into the desk surface, shown in the picture above, possibly seems further away now than it did in 1994.
  • The holographic-effect telepresence video-conferencing may be technically achievable but I don't know many organisations that rely so much on video-conferencing that they would bother with the cost of the extra telepresence fidelity.
  • Large screens for multimedia presentations are reasonably common, though their use is nowhere near as seamless as shown in Starfire.
  • Also the wireless laptop PCs with high definition screens — shown on the same page — are pretty close to the mark (being among one of the more conservative projections, based on products already widely used in 1994).
  • But the kind of visualisation of the 'heroine's' information space, containing 100 million documents, is almost as fanciful as that in Knowledge Navigator when you compare it to the hierarchical window-based directories and Google search results that most people use on a day-to-day basis.

What's perhaps most depressing is that some of the best visionary usability ideas were good enough to inspire small steps towards them — but those steps have since been abandoned. You could (if you were a little generous!) see Apple's eWorld as a move towards rich visualisation of information spaces, of the kind used by Starfire's heroine, but look what happened to that two years after its 1994 launch.

More substantially Tognazzini was a great evangelist for what he and others called 'document-centred' computing. He argued that the 'application-centred' approach was a historical design accident brought about by early PCs' limitations of RAM, which have been irrelevant for many years. In the document-centred approach you can take a single document or file and use whatever favourite application does the job (text, graphics, calculation, multimedia) that you want to do on that document. The advantages of this were well understood, and steps were made with Apple's OpenDoc technology, incorporated into its ClarisWorks software suite. Yet OpenDoc too was abandoned in 1997.

The problem with these improvements is that the real benefits come only when the technologies — rich visualisation, full document-centred computing, or seamless gestural interfaces — are in place, mature, and supported by the social and economic infrastructure of the marketplace. While they are young, new and different, they are vulnerable. Especially in the ICT marketplace of the last ten years where battles have been won more by big players throwing their weight around (or should that be "leveraging their installed customer base"?). This is a variant on the idea of what economists call a 'network effect,' for which I refer you to people like Phil Agre, who understand and write about such things better than me!

I'd like to know what Tog things, but I'm probably too shy to ask him.

Posted by David Jennings in section(s) Human-Computer Interaction on 4 November 02003 | TrackBack

David may be shy, but he is also highly perceptive.

I have been happy with a few changes we had envisioned. We showed a user using a web browser, something that did not yet exist when we conceived it (although we had played with prototypes of such at Apple and Sun). Indeed the web arrived faster than we expected.

Sun users do have the capability of having a single user space, accessible from around the world. This space is not populated by a million objects, nor is the network bandwidth available to support the kinds of navigation we envisioned. The public is being sold a bill of goods with today's so-called "wide bandwidth" connections. The large desktop we envisioned for Starfire could absorb as much as one terabyte per second of information while still delivering useful information to the user..

Let me address several points directly:

• Displays continue to lag hopelessly behind all other technologies. If memory were tracking display progress, we'd be using 4K memory chips right now, up from the 1K of the 1960s. I still expect we will hit the threshold where displays have sufficient density and high enough yields that we can begin to add scanning and other capabilities. Unfortunately, it will be a while.

• High-presence video conferencing is needed and could save companies millions in travel expenses. The tiny networks of today are holding it back. What we showed, by the way, was not holographic. It was a flat image on a curved surface. It could be accomplished quite handily with the new high-speed Internet currently being developed.

• Large screen compatibility has gone up tremendously. Today, unlike 12 years ago, when we made the film, you have a reasonable chance of connecting your laptop in 15 minutes or less. 12 years ago, there was less than a 50/50 chance it would happen at all.

• David is absolutely correct that the big players have used the last decade to cement their positions, to the detriment of their customers. Both the web and Linux have helped put the fear of God in them, but, today, Microsoft Office continues to be Microsoft Suite of Fractured, Disconnected Programs Developed at Different Times with No Admission to Outsiders, and it and all the other traditional applications likely will remain so for the foreseeable future. Apple's Open Doc, unfortunately, was almost ten years too late. By the time they decided to do this project I'd been pushing since the mid-'80s, the company's market share and influence had waned.


Posted by: Bruce Tognazzini on 29 December 02003 at 1:43 AM
Post a comment

Remember personal info?