NerdTV #7: Dan Drake (I, Cringely . NerdTV . Transcript | PBS):
(...) Bob: And Xanadu.
Dan: Right.
Bob: Now how the heck did that come about?
Dan: Well it was an interesting product-project. You don't know whether-how many orders of magnitude it is away from being technically feasible, until you've tried seriously to do it. The answer was three.
Bob: Three orders of magnitude?
Dan: At one point, that was the estimate. That was when it was being worked on seriously. A bunch of very bright people got together in their spare time and put together a lot of very bright ideas. They tended to try to keep-Ted had the basic idea and we liked him.
Dan: Well he and I, probably not. He was one of the two people credited with inventing the whole idea of links, hypertexts.
Bob: Hypertexts.
Dan: Which he did because he had an increasing number of storage lockers across the country filled with his papers they had collected, and he wanted some-when he first heard of computers, he thought, "Oh boy, now I can organize all of that in the computer," and he found out-well they wrote a book about all of that. So he was brilliant with the ideas and that appealed to us. And there were this group of people who had started doing serious technical talk about it and they talked about all of the problems.
Okay, Xanadu was hypertext, as it ought to have been. You could link to an arbitrary section of text in any other document. Everything had a back link, which means that if you wanted to-if somebody found an error in this financial document, or let's say, a technical document you put out, you didn't have to find all the people who had that error. You'd change it and if they were pointing to that piece of text, they get the new text. They can also get the old text. And every link is signed. And so when in this giant world database, somebody-there's a crackpot, you just don't see his stuff, because he's adjusted your stuff not to see, or you could assign weights to different signatures and see the stuff that made sense. And maybe, once in a while, at random, pick up some random person that you don't know. I'm talking here probably from my conception of it and it would be described differently by the people in it.
But the point is it was a whole technology of links that were signed, were bi-directional that could point to an arbitrary piece of text, which you could have included in your document just by its being there. Possibilities of the data being duplicated or being replicated and distributed, so you're not going to Timbuktu to pick up a copy, and copy writing with micro-money charges for every action to this. So it had all of that stuff. And they worked it out. All the things I had thought of in daydream bull sessions, they had worked out in detail over the years.
Bob: It sounds a little like the World Wide Web.
Dan: It's what the worldwide web would have been if they had failed to invent it for another ten years, right. Yeah, and so we were working on this thing and the web was just about to come into noticeable size, about to be invented when we took up this. Trouble is, they were heavy techies who wanted to get things right, and did not have the "you must ship something" imperative that the old hacking types did. They designed, and designed and designed and had a great trouble getting any-some crummy thing to ship. That was part of the problem. The other problem was sheer-oh, and then the project ran into-suddenly they realized object-oriented programming was it and had to-invented a really weird system for writing it in, I think, in C++ and-no, writing it in small talk and translating it to C++, this didn't help the speed of the project.
Anyway, it was impossible to get anything to ship. And as I said, a brilliant idea, we tried to make it work and it just was not able to get off the ground, as much for technical things as for the critical things I've said here about design.
Bob: Yeah. Well it seems to me that it was six to ten years ahead of the hardware it was going to run on it.
Dan: Yeah, that was part of the problem. And really, there was a report at one time that they had working stuff and how they would overcome the three orders of magnitude, too slow it was. And well there wasn't time to do that. We were, I think, too slow. We, meaning the old regime of the company were really too slow about just cutting it loose and saying sorry. Really, the new people who took over when the new CEO came in did so, although this caused a lot of resentment. They're cutting loose this brilliant technical thing. It was a reasonable business decision.
Bob: Sure.