Future of the Web, The Consortium etc: Mission, visions of the future
This document starts out being an attempt to put into article form the content of many keynote speeches. Wrapping them into one may lead to many articles. Keynote speeches have been for example:
When I am asked to predict where the web will be in five or ten years, I refuse: Instead, I answer a different question, "Where would you like this all to lead?". I'll try to do this in this article, and I don't do it idly. The spread of the web has demonstrated many things, one of which is that if you keep suggesting a direction to a creative group of people, things happen.
To put things in perspective, I'll start by explaining where I wanted the web to lead when I first proposed it in 1989 or even when I wrote its predecessor in 1980. We'll see how some of these hopes have come dramatically to fruition, and some are still waiting to be achieved.
I, and many of the early people to pick up the Web, had three hopes for the new Information Space. Or at least, I can divide those hopes into three categories. Firstly, the hope was for an incredibly empowering the individual. The first "wow" to hit someone new to the Web is the power it gives a person to be able to tap into such a vast world of information. Originally, at the same time there was to be another power, and at least as important: the power of self-expression in hypertext. I use the term "hypertext" loosely here: I mean the power to make links, to organize in one's own way the rich world out there; to communicate to others with the extra power of hypertext reference.
Now not many people currently (1997) have an intuitive hypertext editor on their desktops, and few have a piece of web space into which they can easily drop new creations. But the original prototype "WorldWideWeb" program allowed anyone to write hypertext just as anyone could read it. Currently, the only form of hypertext editing people use all the time is the "bookmark list". So the personal empowerment hope has been realized half way, and the results have still been as we have all seen remarkable.
The second part of the dream was that we would not only be able to write hypertext, but a group of people would be able to work together on a hypertext which defined their group understanding. If you look at the work of a group of a people, whether it is a team designing a space shuttle or a clothing salesperson with a customer, they are combining their knowledge to a point where they have amassed enough common understanding to achieve their goal. A huge amount of human endeavour can be viewed in this light. [...] The frustrating thing to me in the 1980s was that in offices and labs, much of the stuff of decisions, the reasons why things had been done, the dependences which existed between people, and between parts of companies or systems, were all written down at some time on a computer, but were effectively lost. The minutes of meetings would be typed into a Macintosh, printed off and distributed in paper copies. It would later be quite impractical for someone reconsidering a decision to actually find the reasons for it, even though the hard part (typing it in) had been done, and it was one some disk somewhere on a machine probably on a network of some sort. So, part two of the dream is that, by giving people a very easy way of putting new information into the Web, and a really easy way of linking it to places where it is relevant, that much of the misunderstanding which we spend so much time acting and reenacting would be prevented. I imagined that new people joining groups or projects would be able to get up to speed all be themselves, starting with their new task, and always being able to ask "why?", and "how?". To a certain extend we are seeing the benefits of this now. But we have yet to see huge shifts in the scales of tasks we can accomplish. Maybe we will manage to make management a thing we do largely within teams with computer assistance. Maybe we will be able to make more powerful democratic processes based on the exposure of arguments and reasons. Who knows - we are just starting, and we yet to develop the tools we will need.
These two dreams, of personal power, and of a transformation in the way we can work and play together, involve people communicating though the web with each other. The computers and networks disappear to become, very properly, part of the invisible infrastructure, leaving a people face to face with the information and each other. Indeed, a major step in the web was that in hiding the details of processors, disks, commands, operating systems and networks, it allowed "normal" people to use computers to handle information in a way they had not been able to before. But in the third part of the dream, the computer reappears. This time, the computer, or more properly software agents running on computers, appear if not our peers as visible servants.
If you want to buy a car now, you an put a web page somewhere on the web, explaning what you want, and you can wait for people to contact you. Alternaively, you can browse the web or ask a search engine to find you a page in which someone is selling a car. In either case, the buyer and seller have to find each other. The tool they use is often a search engine, and we all know that these are currently hopeless. Let me be clear what I mean. The engines are not hopeless - they do a very good hjob, considering that it is virtually impossible for a computer program to figure out what a web page is about. This is why, when looking for a car for sale, you end up looking fo web pages which have certain words in, maybe "car" (not a good choice!) maybe "price", maybe "low milage".
Now imagine a web in which you can, using a template, make a decalration of what you are looking to buy or to sell, in such a way that a computer program can understand it. There are two paths here: one, that computer programs will become so smart that they will understand typical web pages (unlikely), and others that we use some sort of interface, like a form, to answer very specific questions, and the answers are stored in a well defined computer language. Let's persue the second possbility. In this case, an agent program, perhaps sitting next to the we indexer of a big search engine, can look out for a mach between a tender for purchase and an offer of sale. Now the web has gained a certain smartness independent of the people. This is a really simple case, in which an agent matches a proposed solution and a problem. We can imagine computer analysis being much more complex. Suppose that all the dependencies with in a company were available for review by an authorised sofware agent: the documents, designs, reporting structures, financial paths. Computers can be tirless in looking at large problems. Could we use them to help us manage ourselves? There is always a phase change in a company as it grows past that point wher you know everyone. Suddenly, noone has total knowldge, structure arises, incentives become intertwined, and what was a human group in some way becomes a machine. The problem of scaling human effort has no easy answers. Maybe a world of human beings working with in an environment peppered with software agents will be able to crack problems which have always elluded us.
The slim user guide for my unplublished "Enquire Within" program in 1980 suggested that machine analysis might by a great payoff of an information space of linked information. But looking at the current Web today, its coporations, is enthusiasts, its consumers, its weirdos, the thought of analysis suggests mainly targetted marketing and mailing list generation, and concerns of privacy. How are we to chart a path from the web we see today, in ots stone age, to a future exiting enough for us not to be able to understand it? In fact, many of the pieces are, if not in place, close by. There is a lot of work to be done, technically and socially, but nothing seems imposisble. The worst part of it is, having defined an idea of where one would like to be, to find the sequence of steps which actual people and companies have a real incentive to make, which will get us there.
Looking at the developments which are proposed, or in progress under the auspices of the World Wide Web Consortium gives us clues to some of these pieces. I will try to show some scenarios in which they can lead, when put togther, to the dramtically new world we enviage.
By 1992, my role was changing. I had had to be the evalngelist for a whole idea of global hypertext, persuading people that it would not be too complex or too confusing or too difficult, and that a little effort in writing browsers for the major platforms and gateways into existing data would pay off.
In 1993, in my office in CERN in Geneva, I had a call from reception, saying that a group of four people from the USA had dropped in to see me. We got together in a conference room. They explained that they were charged with formulating a large coimputer compay's reoganizaion around the Internet and the Web. They were concerned, they said, that the whole Web seemed to hang on specifications which were sitting on some disk around here. Who would ensure that these specifications would eveolve? Who would ensure that they would remain open interoperable standards out of proprietary control? How could they be involved?
Discussions with that company and many others that year made it clear that the primary need was for a neutral meeting place in which the various companies involved in the exploding Web from all sides could get together and make it a solid base for a huge new business. Most people seemed to think that the new business was going to be a software business, but I pointed out that if it worked, the economic impact would be mainly not just the technology providers, or even those industries who sold information services directly over the web, but on a whole economy which would operate more efficiently in an information space without boundaries. The Consortium model as used by the X Consortium was generally seen to be an excellent starting point, though clearly the Web was a different case. Happily, Al Vezza at MIT/LCS, who was instrumental in helping Bob Scheifler set up the X Consortium, was prepared to do it again. Michael Dertouzos, who headed LCS, was keen not only on the W3C idea, but also, as a European, went along with my hopes that W3C would be an International rather than American body, with initially one foot in the US and one in Europe.
In September 1994, the W3C was born. Over the first three years, its membership has increased from 0 to 200 companies and is still rising. During that time it has changed to accomodate vastly differeing needs of different groups and different areas of technology.
The Consortium is open, in that any organization can join, so long as they sign the membership agreement (which is on the web) and pay the fee ($5k per year or $50k/year for companies grossing over $50 million). Participation in W3C activities is open to all members, generally subject to a commitment to a certian level fo effort by each working group member. A process document defines how new activities are chartered and mandated by the members. One fundamental aspect of W3C is that its activities have such a varied combination of technical, social, market and political pressures that great flexibility and careful thought it required to apply the appropriate tools (Workshops, Interest groups, Coordination groups and Working groups, sample code distribution) to each new activity.
The consortium's work is divided into three domains, which inter-relate.
The Architecture domain, which deals with everything "under the hood" to do with the fundamental structure and properties of the web, and its efficient and reliable operation, covers such aspects as HTTP, and URIs, and now work on new languages for structured information.
The "User Interface" domain, by contrast, specifically addresses the way the person and the computer interact to put the person into intimate interaction with the information space. This includes new versions of HTML, style sheets, and graphics. In this area also is the work on Internationalization ("I18n") to bring web access under the same terms to those of all cultures, scripts and languages.
The third domain, "Technology and Society", responds to the fact that we are building the new society with the help of the Web, and the society we build is directly influenced by, and directly influences, the protocols the technologies design. In ensureing that social concepts are represented by technlolgy, this domain covers for example labelling and endiosement, security, payment, intellectual prperty rights (IPR) and privacy. Specific projects, such as P3 in privacy and DSig for Digial Signature, address timely steps in this development.