THE SECOND COMING — A MANIFESTO

Conversation : TECHNOLOGY

THE SECOND COMING — A MANIFESTO

By David Gelernter [12.31.99]
Introduction by:
David Gelernter

 

Everything is up for grabs. Everything will change. There is a magnificent sweep of intellectual landscape right in front of us.

Introduction
By John Brockman

David Gelernter .....

"...prophesied the rise of the World Wide Web. He understood the idea half a decade before it happened." (John Markoff)

"...is a treasure in the world of computer science...the most articulate and thoughtful of the great living practitioners" (Jaron Lanier)

"...is one of the pioneers in getting many computers to work together and cooperate on solving a single problem, which is the future of computing." (Danny Hillis)

"...is one of the most brilliant and visionary computer scientists of our time." (Bill Joy)

Yale computer scientist David Gelernter entered the public mind one morning in January '92 when The New York Sunday Times ran his picture on the front page of the business section; it filled nearly the whole page. The text of the accompanying story occupied almost another whole page inside.

In 1991 Gelernter had published a book for technologists (an extended research paper) called Mirror Worlds, claiming in effect that one day, there would be something like the Web. As well as forecasting the Web, the book, according to the people who built these systems, also helped lay the basis for the internet programming language "Java" and Sun Microsystems' "Jini."

Gelernter's earlier work on his parallel programming language "Linda" (which allows you to distribute a computer program across a multitude of processors and thus break down problems into a multitude of parts in order to solve them more quickly) and "tuple spaces" underlies such modern-day systems as Sun's JavaSpaces, IBM's T-Spaces, a Lucent company's new "InfernoSpaces" and many other descendants worldwide.

By mid-'92 this set of ideas had taken hold and was exerting a strong influence . By 1993 the Internet was growing fast, and the Web was about to be launched. Gelernter's research group at Yale was an acknowledged world leader in network software and more important, it was known for "The Vision Thing", for the big picture.

In June '93 everything stopped for Gelernter when he was critically injured by a terrorist mailbomb. He was out of action for the rest of '93 and most of '94 as the Web took off, the Internet become an international phenomenon and his aggressive forecasts started to come true. Gelernter endured numerous surgeries through 95, and then a long recuperation period.

Now Gelernter is back. In this audacious manifesto, "The Second Coming", he writes: "Everything is up for grabs. Everything will change. There is a magnificent sweep of intellectual landscape right in front of us.""

-JB

DAVID GELERNTER, Professor of Computer Science at Yale University and adjunct fellow at the Manhattan Institute, is a leading figure in the third generation of Artificial Intelligence scientists, known for his programming language called "Linda" that made it possible to link computers together to work on a single problem. He has since emerged as one of the seminal thinkers in the field known as parallel, or distributed, computing.

He is the author of Mirror Worlds (1991), The Muse In The Machine (1994), 1939: The Lost World Of The Fair (1995), And Drawiing A Life: Surviving The Unabomber (1998).

David Gelernter's Edge Bio Page

THE REALITY CLUB: Stewart Brand, David Ditzel, John C. Dvorak, Feeman Dyson, George Dyson, Douglas Rushkoff, Rod Brooks, Lee Smolin, Jaron Lanier, David Farber, Danny Hillis, Vinod Kholsa, John McCarthy on "The Second Coming - A Manifesto" by David Gelernter

HE SECOND COMING — A MANIFESTO

Any Microsecond Now

Computing will be transformed. It's not just that our problems are big, they are big and obvious. It's not just that the solutions are simple, they are simple and right under our noses. It's not just that hardware is more advanced than software; the last big operating-systems breakthrough was the Macintosh, sixteen years ago, and today's hottest item is Linux, which is a version of Unix, which was new in 1976. Users react to the hard truth that commerical software applications tend to be badly-designed, badly-made, incomprehensible and obsolete by blaming themselves ("Computers for Morons," "Operating Systems for Livestock"), and meanwhile, money surges through our communal imagination like beer from burst barrels. Billions. Naturally the atmosphere is a little strange; change is coming, soon.

Everything Old Is New Again

1. No matter how certain its eventual coming, an event whose exact time and form of arrival are unknown vanishes when we picture the future. We tend not to believe in the next big war or economic swing; we certainly don't believe in the next big software revolution.

2. Because we don't believe in technological change (we only say we do), we accept bad computer products with a shrug; we work around them, make the best of them and (like fatalistic sixteenth-century French peasants) barely even notice their defects — instead of demanding that they be fixed and changed.

3. Everything is up for grabs. Everything will change. There is a magnificent sweep of intellectual landscape right in front of us.

4. The Orwell law of the future: any new technology that can be tried willbe. Like Adam Smith's invisible hand (leading capitalist economies toward ever-increasing wealth), Orwell's Law is an empirical fact of life.

Ripe Ready and hanging by a thread

5. We know that big developments are inevitable in the software world — if only because nothing in that world corresponds to a "book." You can see a book whole from the outside. You know in advance how a book is laid out — where the contents or the index will be — and how to "operate" one. As you work through it, you always know where you stand: how far you have gone and how much is left. "Book" can be a physical object or a text — an abstraction with many interchangeable physical embodiments. These properties don't hold for file systems or web sites. You can't see or judge one from the outside, anticipate the lay-out, tell where you stand as you work your way through.

Whenever we are organizing information, the book is too powerful an idea to do without in some form or other.

6. Miniaturization was the big theme in the first age of computers: rising power, falling prices, computers for everybody. Theme of the Second Age now approaching: computing transcends computers. Information travels through a sea of anonymous, interchangeable computers like a breeze through tall grass. A dekstop computer is a scooped-out hole in the beach where information from the Cybersphere wells up like seawater.

7. "The network is the computer" — yes; but we're less interested in computers all the time. The real topic in astronomy is the cosmos, not telescopes. The real topic in computing is the Cybersphere and the cyberstructures in it, not the computers we use as telescopes and tuners.

8. The software systems we depend on most today are operating systems (Unix, the Macintosh OS, Windows et. al.) and browsers (Internet Explorer, Netscape Communicator...). Operating systems are connectors that fasten users to computers; they attach to the computer at one end, the user at the other. Browsers fasten users to remote computers, to "servers" on the internet.

Today's operating systems and browsers are obsolete because people no longer want to be connected to computers — near ones OR remote ones. (They probably never did). They want to be connected to information. In the future, people are connected to cyberbodies; cyberbodies drift in the computational cosmos — also known as the Swarm, the Cybersphere.

From The Prim Pristine Net To The Omnipresent Swarm

9. The computing future is based on "cyberbodies" — self-contained, neatly-ordered, beautifully-laid-out collections of information, like immaculate giant gardens.

10. You will walk up to any "tuner" (a computer at home, work or the supermarket, or a TV, a telephone, any kind of electronic device) and slip in a "calling card," which identifes a cyberbody. The tuner tunes it in. The cyberbody arrives and settles in like a bluebird perching on a branch.

11. Your whole electronic life will be stored in a cyberbody. You can summon it to any tuner at any time.

12. By slipping it your calling card, you customize any electronic device you touch; for as long as it holds your card, the machine knows your habits and preferences better than you know them yourself.

13. Any well-designed next-generation electronic gadget will come with a ``Disable Omniscience'' button.

14. The important challenge in computing today is to spend computing power, not horde it.

16. The future is dense with computers. They will hang around everywhere in lush growths like Spanish moss. They will swarm like locusts. But a swarm is not merely a big crowd. The individuals in the swarm lose their identities. The computers that make up this global swarm will blend together into the seamless substance of the Cybersphere. Within the swarm, individual computers will be as anonymous as molecules of air.

17. A cyberbody can be replicated or distributed over many computers; can inhabit many computers at the same time. If the Cybersphere's computers are tiles in a paved courtyard, a cyberbody is a cloud's drifting shadow covering many tiles simultaneously.

18. But the Net will change radically before it dies. When you deal with a remote web site, you largely bypass the power of your desktop in favor of the far-off power of a web server. Using your powerful desktop computer as a mere channel to reach web sites, reaching through and beyond it instead of using it, is like renting a Hyundai and keeing your Porsche in the garage. Like executing programs out of disk storage instead of main memory and cache. The Web makes the desktop impotent.

19. The power of desktop machines is a magnet that will reverse today's "everything onto the Web!" trend. Desktop power will inevitably drag information out of remote servers onto desktops.

20. If a million people use a Web site simultaneously, doesn't that mean that we must have a heavy-duty remote server to keep them all happy? No; we could move the site onto a million desktops and use the internet for coordination. The "site" is like a military unit in the field, the general moving with his troops (or like a hockey team in constant swarming motion). (We used essentially this technique to build the first tuple space implementations. They seemed to depend on a shared server, but the server was an illusion; there was no server, just a swarm of clients.) Could Amazon.com be an itinerant horde instead of a fixed Central Command Post? Yes.

Stranger Than Fiction: Computers Today

21. The windows-menus-mouse "desktop" interface, invented by Xerox and Apple and now universal, was a brilliant invention and is now obsolete. It wastes screen-space on meaningless images, fails to provide adequate clues to what is inside the files represented by those blurry little images, forces users to choose icons for the desktop when the system could choose them better itself, and keeps users jockeying windows (like parking attendants rearranging cars in a pint-sized Manhattan lot) in a losing battle for an unimpeded view of the workspace — which is, ultimately, unattainable. No such unimpeded view exists.

22. Icons and "collapsed views" seem new but we have met them before. Any book has a "collapsed" or "iconified" view, namely its spine. An icon conveys far less information that the average book spine — and is much smaller. should it be much smaller? Might a horizontal stack of "book spines" onscreen be more useful than a clutter of icons?

23. The computer mouse was a brilliant invention, but we can see today that it is a bad design. Like any device that must be moved and placed precisely, it ought to provide tactile feedback; it doesn't.

24. Metaphors have a profound effect on computing. The desktop metaphor traps us in a "broad" instead of "deep" arrangement of information that is fundamentally wrong for computer screens. Compared to a standard page of words, an actual desktop is big and a computer screen is small. A desktop is easily extended (use drawers, other desks, tables, the floor); a computer screen is not.

25. Apple could have described its interface as a pure "information landscape," with no connection to a desktop; we invented this landscape (they might have explained) the way a landscape architect or amusement park designer invents a landscape. We invented an ideal space for seeing and managing computerized information. Our landscape is imaginary, but you can still enter and move around it. The computer screen is the window of your vehicle, the face-shield of your diving-helmet.

26. Under the desktop metaphor, the screen IS the interface — the interface is a square foot or two of glowing colors on a glass panel. In the landscape metaphor, the screen is just a viewing pane. When you lookthrough it, you see the actual interface lying beyond.

Problems On The Surface And Under The Surface

27. Modern computing is based on an analogy between computers and file cabinets that is fundamentally wrong and affects nearly every move we make. (We store "files" on disks, write "records," organize files into "folders" — file-cabinet language.) Computers are fundamentally unlike file cabinets because they can take action.

28. Metaphors have a profound effect on computing: the file-cabinet metaphor traps us in a "passive" instead of "active" view of information management that is fundamentally wrong for computers.

29. The rigid file and directory system you are stuck with on your Mac or PC was designed by programmers for programmers — and is still a good system for programmers. It is no good for non-programmers. It never was, and was never intended to be.

30. If you have three pet dogs, give them names. If you have 10,000 head of cattle, don't bother. Nowadays the idea of giving a name to every file on your computer is ridiculous.

31. Our standard policy on file names has far-reaching consequences: doesn't merely force us to make up names where no name is called for; also imposes strong limits on our handling of an important class of documents — ones that arrive from the outside world. A newly-arrived email message (for example) can't stand on its own as a separate document — can't show up alongside other files in searches, sit by itself on the desktop, be opened or printed independently; it has no name, so it must be buried on arrival inside some existing file (the mail file) that does have a name. The same holds for incoming photos and faxes, Web bookmarks, scanned images...

32. You shouldn't have to put files in directories. The directories should reach out and take them. If a file belongs in six directories, all six should reach out and grab it automatically, simultaneously.

33. A file should be allowed to have no name, one name or many names. Many files should be allowed to share one name. A file should be allowed to be in no directory, one directory, or many directories. Many files should be allowed to share one directory. Of these eight possibilities, only three are legal and the other five are banned — for no good reason.

Streams Of Time

34. In the beginning, computers dealt mainly in numbers and words. Today they deal mainly with pictures. In a new period now emerging, they will deal mainly with tangible time — time made visible and concrete. Chronologies and timelines tend to be awkward in the off-computer world of paper, but they are natural online.

35. Computers make alphabetical order obsolete.

36. File cabinets and human minds are information-storage systems. We could model computerized information-storage on the mind instead of the file cabinet if we wanted to.

37. Elements stored in a mind do not have names and are not organized into folders; are retrieved not by name or folder but by contents. (Hear a voice, think of a face: you've retrieved a memory that contains the voice as one component.) You can see everything in your memory from the standpoint of past, present and future. Using a file cabinet, you classify information when you put it in; minds classify information when it is taken out. (Yesterday afternoon at four you stood with Natasha on Fifth Avenue in the rain — as you might recall when you are thinking about "Fifth Avenue," "rain," "Natasha" or many other things. But you attached no such labels to the memory when you acquired it. The classification happened retrospectively.)

38. A "lifestream" organizes information not as a file cabinet does but roughly as a mind does.

39. A lifestream is a sequence of all kinds of documents — all the electronic documents, digital photos, applications, Web bookmarks, rolodex cards, email messages and every other digital information chunk in your life — arranged from oldest to youngest, constantly growing as new documents arrive, easy to browse and search, with a past, present and future, appearing on your screen as a receding parade of index cards. Documents have no names and there are no directories; you retrieve elements by content: "Fifth Avenue" yields a sub-stream of every document that mentions Fifth Avenue.

40. A stream flows because time flows, and the stream is a concrete representation of time. The "now" line divides past from future. If you have a meeting at 10AM tomorow, you put a reminder document in the future of your stream, at 10AM tomorrow. It flows steadily towards now. When nowequals 10AM tomorrow, the reminder leaps over the now line and flows into the past. When you look at the future of your stream you see your plans and appointments, flowing steadily out of the future into the present, then the past.

41. You manage a lifestream using two basic controls, put and focus, which correspond roughly to acquiring a new memory and remembering an old one.

42. To send email, you put a document on someone else's stream. To add a note to your calendar, you put a document in the future of your own stream. To continue work on an old document, put a copy at the head of your stream. Sending email, updating the calendar, opening a document are three instances of the same operation (put a document on a stream).

43. A substream (for example the "Fifth Avenue" substream) is like a conventional directory — except that it builds itself, automatically; it traps new documents as they arrive; one document can be in many substreams; and a substream has the same structure as the main stream — a past, present and future; steady flow.

In The Age Of Tangible Time

44. The point of lifestreams isn't to shift from one software structure to another but to shift the whole premise of computerized information: to stop building glorified file cabinets and start building (simplified, abstract) artificial minds; and to store our electronic lives inside.

45. A lifestream can replace the desktop and subsume the functions of the file system, email system and calendar system. You can store a movie, TV station, virtual museum, electronic store, course of instruction at any level, electronic auction or an institution's past, present and future (its archives, its current news and its future plans) in a lifestream. Many websites will be organized as lifestreams.

46. The lifestream (or some other system with the same properties) will become the most important information-organizing structure in computing — because even a rough imitation of the human mind is vastly more powerful than the most sophisticated file cabinet ever conceived.

47. Lifestreams (in preliminary form) are a successful commercial product today, but my predictions have nothing to do with this product. Ultimately the product may succeed or fail. The idea will succeed.

Living Timestreams

48. Lifestreams today are conventional information structures, stored at web sites and tuned-in using browsers. In the future they will be cyberbodies.

49. Today's operating systems connect users to computers. In the future we will deal directly with information, in the form of cyberbodies. Operating systems will connect cyberbodies to computers; will allow cyberbodies to dock on computers. Users won't deal with operating systems any more, and won't care about them. Your computer's operating system will make as much difference to you as the voltage level of a bit in memory.

50. A lifestream is a landscape you can navigate or fly over at any level. Flying towards the start of the stream is "time travel" into the past.

45. You can walk alongside a lifestream (browsing or searching) or you can jump in and be immersed in information.

51. A well-designed store or public building allows you to size up the whole space from outside, or as soon as you walk in — you see immediately how things are laid out and roughly how large and deep the space is. Today's typical web site is a failure because it is opaque. You ought to be able tosee immediately (not deduce or calculate) how the site is arranged, how big it is, how deep and how broad. It ought to be transparent. (For an example of a "transparent" web site, Mirror Worlds — figure 7.6.)

52. Movies, TV shows, virtual museums and all sorts of other cultural products from symphonies to baseball games will be stored in lifestreams. In other words: each cultural product will be delivered to you in the form of an artifical mind. You will deal with it not as you deal with an object but roughly as you do with a person.

Institutions Afloat In The Cybersphere

53. Your car, your school, your company and yourself are all one-track vehicles moving forward through time, and they will each leave a stream-shaped cyberbody (like an aircraft's contrail) behind them as they go. These vapor-trails of crystallized experience will represent our first concrete answer to a hard question: what is a company, a university, any sort of ongoing organization or institution, if its staff and customers and owners can all change, its buildings be bulldozed, its site relocated — what's left? What is it? The answer: a lifestream in cyberspace.

54. A software or service company equals the employees plus the company lifestream. Every employee has his own view of the communal stream. The company's web site is the publically-accessible substream of the main company stream. The company's lifestream is an electronic approximation of the company's memories, its communal mind.

50. Lifestreams don't yield the "paperless office." (The "paperless office" is a bad idea because paper is one of the most useful and valuable media ever invented.) But lifestreams can turn office paper into a temporary medium — for use, not storage. "On paper" is a good place for information you want to use; a bad place for information you want to store. In the stream-based office, for each newly-created or -received paper document: scan it into the stream and throw it away. When you need a paper document: find it in the stream; print it out; use it; if you wrote on the paper while using it, scan it back in; throw it ou

55. Software can solve hard problems in two ways: by algorithm or by making connections — by delivering the problem to exactly the right human problem-solver. The second technique is just as powerful as the first, but so far we have ignored it.

The Second Coming Of The Computer

56. Lifestreams and microcosms are the two most important cyberbody types; they relate to each other as a single musical line relates to a single chord. The stream is a "moment in space," the microcosm a moment in time.

57. Nowadays we use a scanner to transfer a document's electronic image into a computer. Soon, the scanner will become a Cybersphere port of entry, an all-purpose in-box. Put any object in the in-box and the system develops an accurate 3D physical transcription, and drops the transcription into the cool dark well of cyberspace. So the Cybersphere starts to take on just a hint of the textural richness of real life.

We'll know the system is working when a butterfly wanders into the in-box and (a few wingbeats later) flutters out — and in that brief interval the system has transcribed the creature's appearance and analyzed its way of moving, and the real butterfly leaves a shadow-butterfly behind. Some time soon afterward you'll be examining some tedious electronic document and a cyber-butterfly will appear at the bottom left corner of your screen (maybe a Hamearis lucina) and pause there, briefly hiding the text (and showing its neatly-folded rusty-chocolate wings like Victorian paisley, with orange eyespots) — and moments later will have crossed the screen and be gone.

But What Does It All Matter?

58. If you have plenty of money, the best consequence (so they say) is that you no longer need to think about money. In the future we will have plenty of technology — and the best consequence will be that we will no longer have to think about technology.

We will return with gratitude and relief to the topics that actually count.

Reality Club Discussion

lee_smolin's picture
Lee Smolin
Physicist, Perimeter Institute; Author, Einstein's Unfinished Revolution

David Gelernter has a wonderful imagination and I am a bit afraid to contradict him, as he has obviously spent much more time thinking about the future of computing than I have. I am intrigued by many of the things he proposes. But let me say a word in defense of the present Macintosh system. I do suspect that some computer scientists have forgotten just how revolutionary and useful the Mac operating system is, and may be underestimating the longevity of this particular technology.

It is true that the Macintosh operating system is based on the old fashioned metaphor of a desktop and filing cabinet. But I find that metaphor very useful. I do think of my computer as a very efficient and useful filing cabinet. I like the fact that the files have names and that I can search for them efficiently in several different ways. I like the hierarchical structure of directories. I like the fact that email is different from ordinary files, and I am happy that it only takes a few key strokes to turn an email into a file if I need it to be one, or vise versa.

I also like the limited area of the desktop on my powerbook screen. At work I have a Silicon Graphics which works a bit more like David wants: one can have many different desktops for different purposes and each can be much bigger than the screen, even though that is many times the size of the screen on my powerbook. But I find that I don't use any of these added features. It is too hard to remember how to use them, and I find that when I try to I often loose windows and icons which are off the screen. What is good about the desktop is that it is so limited. I can have piles of windows open at once, but I know where they all are. When there are too many I know I have to close some, which forces me to do a bit of cleaning up. It is like having to clean up ones desk when it overflows. Only unlike my real desk, which I can simply ignore, I do have to deal with my desktop and clean it up from time to time to keep working. I find this very useful as it enforces a minimal level of organization in my work habits.

What David is describing is a computer which would work more like my own mind. But I am not sure I need a computer of this kind. Perhaps I do, I've never had one. But I do already have quite a good associative memory. My guess is that its limitations are built in, as there is an inevitable compromise between the vividness of memory and associations and alertness to the present. I would not want going to my computer to work to be like opening a box of old letters and photographs or facing the task of throwing away old magazines that I never got to read. With a computer like this I might never get anything done. More than anything what I like about my computer is that it does not offer me any information that I don't ask for.

What has gotten so distasteful about going on line is the imposition of unwanted information. The web was a lot more useful before pages began to be crowded with advertising and unwanted information. The sites I use mostly are the ones that offer the least possibilities for diversion from what I am seeking. If randomness and unpredictability were built into the experience of computing it would cease to become a useful tool for me. Not enough has been said about the way that one site can change the working habits of a whole profession, by changing the way we communicate with each other. This is true of the xxx.lanl.gov site, which is now the universal tool for publication in physics and math. It is tightly and rigidly structured, and that is what makes it so useful. It is an extremely good filing cabinet, so good that it replaces many filing cabinets in thousands of offices all over the world.

I also don't like the metaphor of organizing my interface with the computer in terms of the flow of real time. Another very good aspect of my computer is that it provides the illusion that time can be frozen. I can work on several projects at once, and each one is exactly where I left it when I go back to it. In the context of a very busy life, full of travel and unexpected demands and developments, my computer provides an oasis in which time advances in each window only when I pay attention to it.

So I don't need a computer to enhance my imagination or associative memory. I need a computer that counteracts the effects of my own too active imagination and too busy schedule. Because of this I know that a computer that works the way my powerbook does is something I will always need. And what makes my powerbook so useful is the fact that it works so differently than I do. The fact that all the files have names and locations in a hierarchical system is part of what makes it so useful. When I want to find a paper I wrote three years ago on quantum geometry I want to be able to pull up that file right away, not every file I wrote in the last five years about some aspect of quantum geometry. Every once in a while I loose something and it might be good to have a search machine that worked associatively. But not very often.

I do agree with a lot of what David says. I can imagine lots of improvements on the present Mac operating system. Some of the things he suggests would be very useful. And of course the idea of a kind of cyber-agent who represents me in cyberspace is intriguing and perhaps useful. But I have the sense that David's manifesto is a bit like the predictions I read as a child that by the 21st century cars would have evolved wings and we would all be flying to work. The technology of cars has improved a bit since then, but the basic experience of driving is almost exactly the same. Personally I don't cherish that experience so I prefer living in places where one can get almost everywhere by public transportation. Here in London at the beginning of the 21st century the only people who helicopter to work regularly are a few wealthy businessmen and a few members of the royal family.

jaron_lanier's picture
Jaron Lanier
Computer Scientist; Musician; Author, Who Owns The Future?

I'm so delighted that David is still fighting the good fight, an idealist after all these years. Greed and even satisfied wealth have proven to be agents of distraction to all too many cyberdreamers. It's becoming ever more rare to find a young student with even half of David's quotient of fire in his/her soul about the potential for beauty and meaning in digital tools.

So, while I will offer some criticisms below, I hope they will be read as friendly and supportive.

David falls into a common trap that has snagged many a visionary over the years. He thinks about ideal Platonic computers instead of real computers. A billion Platonic computers support a seamless virtual space in which programs fly about unconcerned with which real computer might be visited at a given moment. A billion real computers, in contrast, require a ten million human beings to run helpdesks, many thousands more to fight lawsuits over software compatibility, and a few hundred more to track malicious viruses that invade the automated virus tracking software that never quite worked.

Real computers, unlike ideal computers, are the first machines that require an infinite rather than a finite amount of human labor for their maintenance. Real computers are less likely to allow us to forget them than any other gadget in the history of invention.

Furthermore, in order for a Platonic computer to appear, human good will and good taste will have to precede it. There will have to be no Bill Gates who forces technological sensibility into a retrograde motion in order to gain power.

In order for a Platonic computer to appear, humans will have to understand how to write large programs that interface with the real world in such a way that they are both modifiable, secure, immune to becoming the bearers of future legacy headaches, and amenable to decent user interface design. We simply don't know how to write such programs yet. I expect us to learn to do it someday, in the same way I expect us to be able to build anti-gravity devices someday. I am idealistic, but not for progress in any relevant timeframe.

Moore's Law simply doesn't apply to software as it does to hardware. Software uses every opportunity to get worse instead of better. More memory means more bloat. More users means more incentives not to change, which means more legacy highwire Band-Aids. Software is like culture, starting out fresh and becoming decadent.

Having said all that, I love David's vision. Reading it inspired me to dig up a bit of my old ranting about what virtual reality software should look like. As it happens, I was hoping for something very much like Lifestreams back in the mid-80s.*

(*See http://www.advanced.org/jaron/vrint.html)

As I re-read this old material now, about fifteen years later, it seems a little naive. Surely I didn't think I'd play back virtualized memories as if they were on tape, fast-forwarding and reversing. That works for a single movie, but is no more possible for a lifetime than naming all those 10,000 cows. How would I break memories into atomic units so that they could be summarized or re-ordered? Would I just see a little bit of each room I entered? Maybe rooms aren't the right divider markers for memories. I'd have to impose some ontology onto my memories in order to be able to reduce them enough to search through them and manipulate them. I can't deal with my memories in an unreduced form, because I don't have the time. (This is the temporal version of the old Borges story about the map as big as the country it represents.) The fact that my memories must be automatically reduced in order to be usable brings up another problem area in David's vision.

Although it isn't immediately apparent, there's a an implicit reliance on Artificial Intelligence in David's manifesto. Somehow the cybertraces that one leaves as one flits about the cyberuniverse, carefree like a butterfly, must at some point be parsed (according to that magic ontology) to be useable in the future. Either there's a sweatshop of third world workers going over the life experiences of every wired citizen of the industrialized world, or there are computer algorithms doing the job.

Maybe by now my colleagues on this list are sick of my unyielding stance on AI, but I must repeat once again that Artificial Intelligence just stinks. It's a phony effect. You can't get something for nothing; the computer can't add wisdom to the mix. Or if you believe it can, I feel you've reduced yourself in a deep way, morally and esthetically. Think of the Turing Test: How can the judge know if the computer has gotten smart or if the person has gotten stupid? How can you know if those omniscient credit rating algorithms are brilliant, or if you're being an idiot by borrowing money when you don't need to in order to feed the algorithm with data?

Once again, I feel a tension between the ideal and the real. I am sold on the Lifestreams vision, on David's whole package, but I think the experience of using it will be extremely labor intensive, for me and for everybody.

And utterly worth all the trouble.

I must reject the final paragraph of the manifesto, which imagines an aspect of life more meaningful than technology, which we will be free to pursue when we can forget about technology. This reminds of Marx's vision of what should happen after the revolution. He imagined we'd be reading the classics and practicing archery! Idealists always believe there's some more meaningful, less dreary plane of existence that can be found in this life. All we have to do is fix this hunking mess in front of us and we'll get there.

A lovely belief to hold!

david_farber's picture
David Farber
Chief Technologist, Federal Communications Commission

Gelernter's manifesto is certainly well written. It is flowery and eloquently stated. However, why is there always a "however", it introduces new terms but not that many new ideas that have not been often expressed.

We are at the edge of a real dramatic change in technology. For the past decade we have evolved from a view that the network is just a way of connecting computers together to the current view that the network is the action to the view often stated (by me and others) that no one cares about the network but only what they can access and interact with ¦ information and people.

We are about to replace our old slow electro-optical communications systems with all optical end to end systems. This technology offers an enormous increase in bits per second. One stand of fiber can carry more bits per second than the entire current national backbone. This will cause a dramatic change in every thing we have now. We will have to re-think our network protocols, the architecture of our computers and just what we mean by a computer and software. Old ideas will soon go the way of the big mainframe operating systems and computers.

Back to the manifesto. It blends well into this rethinking process that the new technology will force. It would be unfortunate if the result of this re conceptualization ended up with the same old appearance and world model to users. The manifesto is a major step in making sure that does not happen. Let's just realize that the ideas are not new - they reflect the ideas of many people over many years. Now we need an industrial structure that allows these ideas to be developed and marketed!

w_daniel_hillis's picture
W. Daniel Hillis
Physicist, Computer Scientist, Co-Founder, Applied Invention.; Author, The Pattern on the Stone

David Gelernter is basically right: current generation computer interfaces are not very good. (Since we are all among friends here, we can say it: they suck). The ubiquitous windows desktop is a classical example of "early lock in", like the Qwerty keyboard and strange conventions for English spelling. These are both generally acknowledged as unfortunate accidents of history. They are non-optimal, but not quite bad enough to be worth changing. In fact, the standard computer interface in incorporates both of these awful interfaces, yet interestingly, Gelernter does not suggest changing them.

Are we at the point where the desk top computer interface will be thrown out and replaced with something better? Is the computer desktop like the Roman alphabet, which we have learned to live with in spite of its quirks, or is like the Roman system of numerals, which we have pretty much abandoned? As much as I like the idea like the idea of starting with a clean slate, I think it is more like the alphabet than the numerals, and it is more likely that the desk top interface will be improved than abandoned. Most of the specific improvements that Geletner suggests, like content addressing, time-linking and multiple names, can be and are being incorporated into standard interfaces. It won't be elegant, but it will work.

So does this mean that we are doomed to a millennium of Windows 2xxx? I doubt it. As Scott McNealy is fond of pointing out, current PC operating systems are unwieldy "hair balls" of accumulated history. Eventually, someone will start from scratch and build something better. But I would be surprised if they start by throwing out the part that most users are the most comfortable with, which is the metaphor physical document handing. The replacement, when it emerges, will win by doing a better job of the same thing.

Yet, there is also a second type of competition, which is not so much a replacement as an addition. Computers are useful for more than handling documents, and other interfaces will be developed for these other functions. These are interfaces more likely to nurture the emergence of radical new ideas. If David Geletner really wants to invent a new interface(and he would probably be good at it) he should forget about looking for a better way to handle documents, and start think about a computer that handles ideas.

stewart_brand's picture
Stewart Brand
Founder, the Whole Earth Catalog; Co-founder, The Well; Co-Founder, The Long Now Foundation, and Revive & Restore; Author, Whole Earth Discipline

It's a great screed, inspiring and generative. It is a frame of reference worth filling with reality.

For me, Gelertner's manifesto speaks to widespread growing aggravation with the current system and growing impatience with the burgeoning tech possibilities not being addressed at a deep enough level. "About time!" was my gut response.

The sequence is clear. From "the user is a luser" (early programmer joke) to "the user wins" to "the user rules" (eg. Napster) and "the user creates" (the Web) to, with Gelertner, "the user is the system."

The still unanswered question though is: How does this system fare over time? How does it keep from the self-obsolescing self-erasure endemic to current computer tech? How do the lifestream contrails keep their shape amid ferociously turbulent winds? Those winds are not extraneous to the system; they are how the system grows.

vinod_khosla's picture
Vinod Khosla
co-founder of Daisy Systems and founding Chief Executive Officer of Sun Microsystems

A brief scan leads to the impression that while "the second coming" is inevitable, like most technologies, the path to getting there often changes the end we get to. Transition strategies here will significantly impact the end state.

david_r_ditzel's picture
David R. Ditzel
President, CEO, ThruChip Communications

David Gelernter's Manifesto is a humbling document read, because it points out the generally unrecognized, but herein revealed truth that we are only at the beginning of understanding how the evolution of the internet is going to change our lives.

Gelernter is ahead of us all in peering through the fog that we call the future of technology.

john_mccarthy's picture
John Mccarthy
Professor of Computer Science at Stanford University

Comments on the Gelernter Manifesto

i. I found a lot wrong with the manifesto, so I'll begin with something I found usable in it. Gelernter grumbles in item 31 that since email messages aren't files they don't have names and can't stand on their own. I also find it a problem, and it occurred to me how to mitigate the problem in my own mail reader which is within my word processor.

Suppose I'm reading a message that I consider significant. Typing a single command inserts a reference to the appropriate page in the message file at the end of a special file of messages, puts in the time, and puts me where I can add an identifying comment. The entry for the email with the manifesto is "Sat Jun 17 12:48:28 2000 /u/jmc/RMAIL.S00==1906 Gelernter Manifesto", giving the time, the location of the message in the mail file and the name I gave the message.

If I later click on that line, I'll be reading the message again.

The purpose of messages having names of some sort is so that the receiver can retrieve a message later. I doubt that such a name can be automatically generated from the message itself, because the subject line, etc. are in the mental space of the sender, not the receiver. The receiver has to somehow give the message a name if he wants to be able to subsequently retrieve it in one step. In this case, I chose "Gelernter Manifesto".

It took 12 minutes to write and debug the message naming facility in the Xemacs editor. The MS-Word users I consulted told me that it would be very difficult to script MS-word and Windows email systems to do it.

ii. We all find ourselves repeating essentially the same tasks in using computers. Here's a slogan.

Anything a user can do himself, he should be able to make the computer do for him.

Fully realizing this slogan would be a big step, but even a little helps. It's called letting the user "customize" his environment. Point i above is a small example.

Unfortunately, the making of computer systems and software is dominated by the ideology of the omnipotent programmer (or web site designer) who knows how the user (regarded as a child) should think and reduces the user's control to pointing and clicking. This ideology has left even the most sophisticated users in a helpless position compared to where they were 40 years ago in the late 1950s.

Scripting languages were a start in the direction of giving the user more power, but the present ones aren't much good, and not even programmers use them much to make their own lives simpler. Scripting is particularly awkward for point and click use. Xemacs customization is reasonably convenient, but it isn't contiguous with Xemacs Lisp, a really good programming language.

Linux is a step in the right direction of giving the user control in that the source of the operating system is available to users, but I doubt that many users, change Linux for purely personal convenience.

Back to Gelernter

iii. Most of the Manifesto's metaphors, e.g. "beer from burst barrels" and "scooped out hole in the beach", aren't informative.

iv. In item 4, Gelernter offers

The Orwell law of the future: any new technology that CAN be tried WILL be. Like Adam Smith's invisible hand (leading capitalist economies toward ever increasing wealth), Orwell's Law is an empirical fact of life.

It isn't true, and I don't believe Orwell said it. In the preface to "1984", Orwell wrote that "1984" is a cautionary tale that he didn't expect to happen. In particular, "1984" has the tv that permitted Big Brother's minions to spy on the viewer. I don't think Orwell expect that to be tried, and it hasn't been.

Indeed the reverse is true. Most possible new technologies are never tried.

v. Gelernter, like many other commentators, is glib about the system software and its documentation being bad. Don Norman beat that drum, and Apple hired him to make things better. He and they didn't have much success. A more careful analysis of what causes difficulty and how to fix it is needed.

vi. The problem with file systems and any other tree structures is that tree structures aren't memorable. Someone else's tree structure, e.g. a telephone keypad tree, is often helpful the first time you use it, but it is a pain to go through the tree again and again to reach a particular leaf.

vii. I couldn't figure out what Cybersphere was supposed to mean except that it's grand, and I see that the other commentators didn't either. Computers haven't changed people's lives to the extent that telephones, radio, automobiles and air travel did early in the previous century. Paul Krugman is eloquent on this point in the NY Times for 2000 June 18. Human level artificial intelligence would revolutionize human life, but fewer people in AI are working in that direction than in the 1970s. Erik Mueller documents one aspect of this neglect in his 1999 article http://www.media.mit.edu/~mueller/papers/storyund.html.

viii. I think the idea of doing an Amazon search for a book on your own computer is a bad one, because the computations are trivial, whereas the file accesses to the Amazon database are substantial. To do it on your own computer would require downloading the whole Amazon catalog before you started your search.

ix. Re item 21 thru 26, I don't think changing "desktop" to "information landscape" would have made much difference. The problem of what you can do with a small screen will remain as long as we have small screens. Two foot by 3 foot flat screens with 200 bit per inch resolution will change computer use much more than another factor of 100 in processor speed. We also need the bathtub screen, the beach screen and the bed screen.

x. item 32. Directories reaching out for files is vague and suggests more AI than is currently available.

xi. There's something in "streams of time", but it's vague. One thing that is feasible is for an operating system to make a journal including all the user's key strokes and mouse clicks and identifiable more substantial operations. The journal should be available for the user to inspect, replay bits of, and to offer for expert inspection when something has gone wrong.

xii. I don't understand to the objection to names; they were invented long before computers. In item 37, Natasha and Fifth Avenue are names.

xiii. item 41. "To send email, you put a document on someone else's stream." That suggests that the recipient would read it right away or at least at a time determined by the sender. Present email sits till you get around to it, and that's better.

xiv. Paper will be needed until screens are better. I use paper just as Gelernter suggests. Print the document for reading and then throw it away. I'll do that even at the cost of losing the pretty red ink I've put on my printout of the Manifesto.

john_c_dvorak's picture
John C. Dvorak
Columnist; Broadcaster in areas of technology and computing

Finally, someone who knows what they're talking about and who isn't simply viewed as a embittered cynic tells it like it is regarding the notion of remote computing among other dumb ideas. Bill Gates will love reading this stuff. Hating it will be the Ellisons and McNealys of the world whose goal is to de-ball the personal computer and replace it with a thin client running eunuchs. I also like his slamming the dubious concept of a computer "Desktop" and trashing the idea of file folders and other computer commonplaces promoted by the charismatic Steve Jobs and copied lockstep by Gates and company. Unfortunately all the points in the manifesto are right but otiose. Trends and fads promoted by strength of personality whether it be Fascism, rap music, thong bikinis or the WIMP (windows icons mouse pointer) interface are not easy to reverse. It's the mechanism of trend reversal that needs study and comment. A laundry list of all that is wrong with computing today is an exercise in futility when hero worship and sheep-like behavior are the norm. This manifesto will amount to nothing in the end. A shame.

freeman_dyson's picture
Freeman Dyson
Physicist, Institute of Advanced Study; Author, Disturbing the Universe; Maker of Patterns

Thank you very much for sending the Gelernter manifesto, full of wonderful imagery and eloquence. Here are some brief comments.

Gelernter lays out a grand vision of cyberbodies and lifestreams inhabiting the cyberspace of the future. He brings his vision to life with images that every child can understand, the bluebird perching on a branch, the cloud's shadow drifting across the paved courtyard. There will be a place for humans, even for children, in his cyberspace. In his vision of the future, we shall no longer be parking cars in a pint-sized Manhattan parking-lot. We shall be flying free in cyberspace, leaving behind vapor trails of experience and memory for other humans to explore.

Fifty years ago we heard about a different vision of a possible future. We heard that the automobile would soon be obsolete, its mobility diminished by the constantly increasing density of traffic, its destructive effect on the environment no longer tolerable in a civilized society. We heard that the automobile would soon be replaced by the helicopter as the preferred vehicle for personal transportation. We would soon be living in a three dimensional world, with helipads replacing garages beside our homes. The reasons why that vision of a roadless civilization never materialized are obvious. Helicopters remained noisy, accident-prone and expensive, roads and automobiles turned out to be unexpectedly resilient. The vision was beautiful, but the tools to make it real were defective.

Gelernter's vision is also beautiful, and his scornful sweeping of existing computers and operating systems into the dustbin of history is persuasive. The chief question that his vision raises is, whether we shall have the tools to make it real. Gelernter disparages tools. He says, "The real topic in astronomy is the cosmos, not telescopes. The real topic in computing is the cybersphere and the cyberstructures in it, not the computers ... ''. I know more about astronomy than about computing. I can certify that he has a one-sided view of astronomy. Modern astronomy is dominated by tools. It is about telescopes and spacecraft as much as it is about the cosmos that these tools explore. Every time we introduce a new tool, we see a new cosmos. And I suspect that he has a one-sided view of computing. I suspect that cyberspace will also be dominated by tools, as far into the future as we can imagine. The topography of our future cyberspace will be determined more by new tools than by Gelernter's vision. Still, he has pointed the way for the next generation of tool builders to follow. We must hope that they will be more successful than the builders of helicopters fifty years ago. If the tool-builders can build tools to match his vision, then our children and grandchildren might see the Second Coming and live in the world of Gelernter's dreams.

george_dyson's picture
George Dyson
Science Historian; Author, Analogia

Let us hope that Gelernter's prophecies continue to be fulfilled. The sooner spines replace icons the better — would you rather work in a library where the books are shelved at eye-level or left lying face-up all over the floor??

For fifty years, digital computing has rested upon two invariant foundations: the program (as given by Turing) and the address matrix (as given by von Neumann and Bigelow). Who could have imagined, 50 years ago, that we would load millions of lines of 'machine-building' code just to check our mail, or that an international political organization would be charged with supervising the orderly assignment of unambiguous coordinates to every bit of memory connected to the net?

Only a third miracle — dirt-cheap, near-perfect microprocessing — allows a system as inherently intolerant of error and ambiguity to work as well as it does today. Gelernter is right: a revolution is overdue. And underway.

In molecular biology, addressing of data and execution of order codes is accomplished by reference to local templates, not by reference to some absolute or hierarchical system of numerical address. The instructions say "do x with the next copy of y that comes along" — without specifying which copy, or where. This ability to take general, organized advantage of local, haphazard processes is exactly the ability that (so far) has distinguished information processing in living organisms from information processing in digital computers. This is not to suggest an overthrow of the address matrix — which is with us to stay. But software that takes advantage of template-based addressing will rapidly gain the upper hand.

The other foundation, the program, is based on the fact that digital computers are able to solve most — but not all — problems that can be stated in finite, unambiguous terms. They may, however, take a very long time to produce an answer (in which case you build faster computers) or it may take a very long time to ask the question (in which case you hire more programmers). For fifty years, computers have been getting better and better at providing answers — but only to questions that programmers are able to ask.

I am not talking about non-computable problems. Despite the perennial attentions of philosophers, in the day-to-day world such problems remain scarce. There is, however, a third sector to the computational universe: the realm of questions whose answers are, in principle, computable, but that, in practice, we are unable to ask in unambiguous language that computers can understand. This is where brains beat computers. In the real world, most of the time, finding an answer is easier than defining the question. It's easier to draw something that looks like a cat than to describe what, exactly, makes something look like a cat. A child scribbles indiscriminately, and eventually something appears that happens to resemble a cat. A solution finds the problem, not the other way around. The world starts making sense, and the meaningless scribbles are left behind. This is the power of that Mirror World we now perceive as the Internet and the World Wide Web.

"An argument in favor of building a machine with initial randomness is that, if it is large enough, it will contain every network that will ever be required," advised cryptanalyst Irving J. Good, speaking at IBM in 1958. Even a relatively simple network contains solutions, waiting to be discovered, to problems that need not be explicitly defined. The network can — and will — answer questions that all the programmers in the world would never have time to ask.

douglas_rushkoff's picture
Douglas Rushkoff
Media Analyst; Documentary Writer; Author, Throwing Rocks at the Google Bus

David Gelernter's "The Second Coming" reminds me just how arbitrarily so many of our decisions about how to do computing and networking have been reached. Techniques for sharing super-computing resources or keeping lines of code ready for a compiler have, through their very legacies, become the architectural basis for humanity's shared information space.

It seems to me that the trick to seeing through today's interfaces — a way of envisioning information architecture that David does effortlessly — involves distinguishing between our modeling systems and the models they build. While memory, information, hardware, and software might need to conform to certain realities, the very opacity of our current operating systems (both technological and social) imply an immutability that just isn't real. The only obstacles to this unencumbered perception of memory, information, storage, and interaction are our own prejudices, formed either randomly or by long-obsolete priorities, and kept in place by market forces.

rodney_a_brooks's picture
Rodney A. Brooks
Panasonic Professor of Robotics (emeritus); Former Director, MIT Computer Science and Artificial Intelligence Lab (1997-2007); Founder, CTO, Robust.AI; Author, Flesh and Machines

David Gelernter is no doubt right on about the coming revolution, but as with all revolutions it is hard to predict the details of how it will play out. I suspect he is wrong on the details of cyberbodies and his lifestreams. The first because as framed it relies still on a physical icon to identify the body, and the second because it is just one metaphor that many will find inconvenient. In the following paragraphs I'll outline my own versions of what the revolution will bring in these two departments, and no doubt my visions will be as wrong or more than David's.

But first the actuality of the revolution. David's criticisms of our current computing environments are eloquently stated, and I think widely shared. A number of projects were started about a year ago, originally through a DARPA sponsored `Computing Expeditions' program. At CMU the expedition is called "Aura", at Berkely it is "Endeavour" (named for Cook's ship, and hence the spelling), at the University of Washington/Xerox Parc it is called "Portolano/Workscapes". At MIT, Michael Dertouzos, Anant Agarwal and I are leading "Project Oxygen" dedicated to pervasive human-centered computing. The common theme across all these projects is that human time and attention is the limiting factor in the future, not computation speed, bandwidth, or storage.

In the past the human has been forced to climb into the computer's world. First with binary, and holes punched in cards, and then later by physically approaching that "square foot or two of glowing colors on a glass panel", and being drawn into its virtual desktop with metaphors bogged down by copies of physical constraints in real offices. In MIT's Project Oxygen, a joint project of the Laboratory for Computer Science and the Artificial Intelligence Lab, we are trying to drag the computer out into the world of people. Computers are fast enough now to see and hear---and these are the principal modalities which we use to interact with other people. We are making our machines interact with people through these same modalities, using the perceptual capabilities of people rather than forcing them to rely on their cognitive abilities just to handle the interface. Cognitive capabilities should be reserved for the real things that people want to do.

Now for cyberbodies and lifestreams. By making computation people centric it should not matter whether I am in your office or mine, whether I pick up your PDA or mine, whether I pick up your cell phone or mine. Wherever I am the system should adapt to my identity whether I am carrying a "calling card" or not. It should adapt to me, not to yet another technological decoration that I need to carry around. And it should be automatic and secure as it does this. Just as people can tell my identity through vision and sound so too can our machines. Furthermore, as computation is cheap, much cheaper these days than special purpose circuitry (and wherever that is not true yet, it will soon be), there is no need for artifacts to have any particular identity. According to my needs at that instant, the machine in my hand should be able to morph from being a PDA to a cell phone, to an MP3/Napster player, just be changing the digital signal processing it is doing. Physics requires a little bit in the way of an aerial, but beyond that demodulation, etc., can be in software. And then the systems should handle bandwidth restrictions behind my back, performing vertical hand-off between protocols as invisibly as today's cell phones perform horizontal hand-off between cells.

Lifestreams are one sort of metaphor. We will not be subject to the tyranny of a single metaphor as we are subject today to the desktop metaphor which Gelernter so masterfully scorns. For a lot of my everyday work I will prefer a metaphor of a personal assistant. I tell it something, and it takes care of the details, watching over me and only interceding when it sees that I need help, pulling in all the necessary information from wherever it is located, perhaps cached ahead of time in anticipation of my needs. After working with me for many years my human personal assistant knows so many details of my life and interactions that I can entrust her to handle many of interactions with the world, without me ever providing any supervision. I will want a similar relationship with my computation. Others might prefer a geographical metaphor, zooming around through a virtual world, while a few might like the lifestreams metaphor. Once a few of these metaphors get invented and tried out, there will be a deluge of new metaphors as the young hackers attack the interface problem with a vengeance.