For inspiration, I think we can look at other areas in which structure has recently enjoyed a period of greater recognition, and see if we can borrow some of their insights into behaviors of the structures they study. From anthropology to sociology to biology to robotics, we see people beginning to study not only structures of people, animals, robots, etc., but the behaviors that arise from them when they are part of collectives. These behaviors are often complex and even counter-intuitive. For example, not only is an ant colony not an ant, but it's difficult for us to understand how the behavior of an ant colony derives from the behaviors of ants. I think hypermedia researchers could begin to address the "behavior" part of the data-structure-behavior triad by taking inspiration from other fields that look at the computational behavior of hypertexts.

I hope that we can build good systems with hypermedia, not just shiny ones. A good system is one that works like the user works. We are always very ready to adapt to the machine instead of building a machine that adapts to us. When I read Bush or Engelbart or the other early pioneers of hypermedia, I see a vision for exactly this type of system – one that adapts to us. Frankly, we've made little progress on this front.

The problem lies mainly with our inability to break the conditioning we have undergone, the conditioning that has led us to believe that machines work a certain way, and we as users must adapt to that way. This conditioning is not simply due to computers, though they certainly play their part. In this time, in which new technologies are being invented and brought to user bases very rapidly, we are constantly confronted with "beta versions" of technologies that are not yet polished and do not adapt to us.

Of course, the Web is now a ubiquitous technology. Many people see this as a great victory for hypermedia. I do not. I think the Web is a fantastic tool that has changed the way we work and even the way we think. But it does not seem a great leap forward in terms of structure-oriented systems. The Web is, first and foremost, a content delivery vehicle. Structure in the Web is an afterthought, existing almost solely in the mind of the user (effected by the artifacts of navigation.) This is not a technological argument – it revolves around the way the Web is actually used. The weak link is still the conditioned human user. Some of the "killer apps" of the Web, like Google, as useful as they are, are exactly the type of application Bush wanted to move past (or to augment, however you choose to interpret Bush's writings.)

So, for me, the last 60-odd years have been a slow and laborious journey toward realizing this brilliant vision of knowledge management and information organization systems geared to work as people work. There still seems a long way to go. However, I do not see this as a bad thing. It keeps hypermedia researchers in business! Also, the most rewarding days are ahead of us.

The second answer is to say: "I would dis-invent (insert technology here), because it has caused such great evil in the world." Common technologies to name here might be the atom bomb or automatic weapons, or even something like alarm clocks. (Think how much happier people would be if they would just get enough sleep!) The problem with this is that it is very hard to see all the implications of dis-inventing a technology, just as it is hard to see the implications of inventing one. I remember the effect of a microwave oven installed at an office in which I worked many years ago on the social life of the office. After the arrival of the microwave, people didn't go out to eat as often, meaning we spent more time together in the lunch room. This was a very fortunate unforeseen consequence of the microwave.

So, if I ever had this opportunity, I'd probably not take it, for fear of really messing things up. But, since I won't ever have this opportunity, I can take my best guess and say that I'd dis-invent cars. I bought a car when I moved back to the US, after a decade of not having one while living in Europe. I'd be happy to give mine up again, but with the huge number of cars in the US, there's no impetus to create replacements for cars. Only the mighty dis-inventing magic wand could fix this problem!

For me, hypermedia was the first exposure I had to computer science research that worked from the human back toward the machine. It is certainly not the only such field, but it is a good example of one. I find that explaining the Bush Memex versus traditional IR systems a good way of introducing this concept to students.

I think that hypermedia is more than just HCI. For example, association-based systems do not stand in opposition only to automated IR systems. They offer an alternative to the way in which information has been organized for centuries. When faced with placing books within a library, we need some external system to facilitate this process. We can place them alphabetically, or by LOC subject number, or by color and size. In all of these examples, we have to use a "one size fits all" a priori organization scheme. But human brains don't work this way (or not exclusively, anyway.) Bush's association-based Memex is an attempt to break the tyranny of exclusive, a priori organization. I think this as grand a vision as I've ever heard inside computer science.