Slashdot Mirror
Knuth: All Questions Answered
sunhou writes: "The AMS published a lecture by Donald Knuth called All Questions Answered (pdf), where Knuth simply responded to questions from the audience. Topics ranged from errors in software ('I think Microsoft should say, "You'll get a check from Bill Gates every time you find an error"') to how he gets distracted by fonts on restaurant menus, to software patents. There were some really good questions (and responses)."
101. What do you think of the Slashdot effect?
TeX has always facinated me. Let's face it, it works. I believe there is more bugs than he is writing checks for, but that said they are seldom encountered by mere mortals. If you do normal stuff it just works.
There is nothing else like it. No commercial product, no non-commercial product. If you want to typeset mathematics, it's the only game in town. If you want to typeset anything, it's one of very few games in town. It's open source. It's multi-platform. It has a huge following, but gets no press.
It really is an amazing thing, and something that every open source project should aspire to....
http://russnelson.com/fea-knuth.pdf
Don't piss off The Angry Economist
By Dan Doernberg
December 7th, 1993
CLB: You have just-published books on both CWEB and the Stanford GraphBase, two areas of your own research. Let's start with CWEB, which integrates C and TeX to facilitate program documentation.
Knuth: The CWEB system is an add-on to C that makes programming better than any other method known in the world, by far. I simply have to be honest and say that it's the greatest thing that's there. The CWEB System of Structured Documentation is the definitive user manual and complete explanation, more than anybody really needs to know about CWEB.
CLB: You've said that CWEB gives an order of magnitude improvement in programmer productivity--- how so?
Knuth: Well, maybe not an order of magnitude, maybe only a factor of two. People who have used CWEB have noticed that they write better programs, that the programs are more portable, more easily debugged, more easily maintained... and they don't take as long to write.
CLB: Has CWEB been used just at Stanford, or in industry as well?
Knuth: It's being used around the world. We've had WEB, the original version (for Pascal) in a variety of systems, and then more and more people started getting infected by it. TeX was written in WEB. Silvio Levy did the conversion to CWEB in 1987. It was experimental for a long time, and now I'm just saying "The experiment worked!". CWEB is much better than WEB, because C is a much nicer language to work with for system programming and lots of other things. For anybody who really cares about programming, I have no idea why they would not prefer this to any other system.
CLB: Easy to use, runs fast, all that good stuff?
Knuth: Right, and it makes you happy after you finish writing a program!
CLB: Even if you write a bad program?!
Knuth:(Don's wife--Ed.) Almost... well... yeah! Jill will tell you, I come out of my office several times a week saying, "CWEB programming is such fun!" It's true, I just can't do enough of it.
The frame of mind that you're in when you're writing a CWEB program is that much better than the old attitude. You think of yourself as writing for a human being, explaining to a human being what a computer should do, instead of thinking of yourself as talking to the computer telling it what to do. You get your act together better when you're explaining it to another person. This approach helps even for a program that you're going to throw away after an hour. CWEB is a tool that I recommend using even if you're writing a program only for yourself, for your eyes only.
CLB: CWEB seems very close to the structured programming models of the 70s...
Knuth: Right, it's the next step. With structured programming, there were some people saying program top-down, and others saying program bottom-up. With WEB/CWEB you can do parts of it bottom-up and parts of it top-down, whatever you feel is right for the program, or for the part of the program you're in.
The structured programming methodology was great... but the way to really understand it is not as a cookbook of rules, but as a way to understand the relation between high-level and low-level views of a program. The way you do that is by viewing the program as a web, as a bunch of small pieces that are simple in themselves and that have simple connections to other small pieces. This way of understanding the complex whole in terms of simple small parts, and the connections between those parts, is supported by the WEB scheme.
You can create the parts in whatever order is psychologically best for you. Sometimes you can create them from the bottom up. Bottom-up means that you know somehow that you probably need a subroutine that will do something, so you write it now while you're ready, while you're psyched for it. With this bottom- up programming, your pencil gets more powerful every page, because on page nine you've developed more tools that you can use on page ten... your pencil is stronger.
With top-down programming you start at the beginning and say "I'm going to do this first and then this, and then this"... but then you have to spell out what those are--- you can wind up gasping for breath a hundred pages later when you finally figure out how you're actually going to do those things!
Top-down programming tends to look very nice for the first few pages and then it becomes a little hard to keep the threads going. Bottom-up programming also tends to look nice for a while, your pencil is more powerful, but that means you can also do more tricky stuff. If you mix the two in a good psychological way, then it works, even at the end.
(TeX: The Program--Ed.) I did this with TeX, a very large program: 500+ pages of code in the book . Throughout that entire program, all those lines of code, there was always one thing that had to be the next thing I did. I didn't really have much choice; each step was based on what I'd done so far. No methodology would teach me how to write a piece of software like that, if I followed it rigorously. But if I imagined myself explaining the program to a good competent programmer, all that this long program was, then there was just this one natural way to do it. The order in which the code appears in the book is the order in which I wrote it.
CLB: To what extent did you or do you follow the "holy war" debates about software engineering methodologies?
Knuth: I didn't follow every nuance of that work, but I was aware of the dominant ideas. I didn't know what the CASE tools were until many years after other people did. I think it was bad to make too much of a religion out of it. There was a lot of "political correctness" about how to program in those days.
There was a similar thing in the mathematics community in the 1920's, where people were saying that good mathematicians would have to prove theorems a certain way. You weren't supposed to use certain tools of proof that some people thought might lead you into paradoxes. It was like trying to do mathematics with a hand tied behind your back. Similarly, politically correct structured programming was keeping people from getting good programs done, when they knew perfectly well what they were doing, just because their approach didn't happen to fit with the current idea of correctness. Computer science is like every other field; it goes in waves of fashion. Some of the trends are good, but almost every good idea seems to get used in a different way than it should have been.
For example, take random number generators. People had no theory about how to generate random numbers for fifteen years. Then somebody proved one small result about a particular method: if you averaged the serial correlation over an entire period of a billion numbers, the average would be zero, which was good. All of a sudden, everybody switched over, they took out all their old routines and converted to this new method, because it was the only one that had any theory to it whatsoever. It turned out this was a horrible random number generator; the theory had not noticed that the average over the first half was +1 and over the second half was -1! All through history, people have taken ideas and misunderstood the limitations of them.
CLB: Which method was this?
Knuth: Well, it was called RANDU in most subroutine libraries. It's been pretty well purged by now; still, if anybody sees a subroutine named RANDU, get rid of it!
CLB: Did you integrate WEB with C because so many programmers today are using it, or do you personally like C and write with it?
Knuth: I think C has a lot of features that are very important. The way C handles pointers, for example, was a brilliant innovation; it solved a lot of problems that we had before in data structuring and made the programs look good afterwards. C isn't the perfect language, no language is, but I think it has a lot of virtues, and you can avoid the parts you don't like. I do like C as a language, especially because it blends in with the operating system (if you're using UNIX, for example).
All through my life, I've always used the programming language that blended best with the debugging system and operating system that I'm using. If I had a better debugger for language X, and if X went well with the operating system, I would be using that.
An extreme case occurred one year I worked in a lab where the operating system had been designed by Ned Irons. The system was for one of Cray's early machines, and Irons had also written a compiler language called IMP. IMP had a lot of horrible features. One, it was an extensible language, and everybody in the lab would keep extending it. A program that worked on Monday wouldn't work on Tuesday, and the first thing that you'd do if your program failed was to check whether the compiled code was OK. The second thing about IMP was that it was an extremely terse language. For example, where in PASCAL you would say "IF X > 0 THEN...", in IMP you say "X+=>". In other words, your program was very short. You felt like you were writing elegant programs, because there were only a few characters, but you couldn't read them the next day! Being very terse meant that you couldn't fathom this bunch of marks on the page...
CLB: I realize your current emphasis is on "literate programming", but were you ever whatsoever attracted to APL as a math-oriented language?
Knuth: That's another story. APL is for people who have problems to solve and don't care too much about efficiency; they want a nice elegant way to state the solution to their problem, but the solution that they come up with is not necessarily anything that a computer has an easy job doing. It's a problem specification language, but not a system programming language... there is an APL-WEB.
But I want to say more about IMP. The third thing against it was, if you made a mistake, the compiler would either get into an infinite loop, or it would stop on your first error and say "ERROR ERROR ERROR" and quit; you would have to figure out what the mistake was! It was not a great language or compiler.
However... it was still my language of choice, because it fit that operating system perfectly. The arrays would be named in a way that you could easily see in the debugger, and you could know where the storage was being allocated, you knew what was going on, and you could actually get your program running reliably, because IMP blended with the operating system. You couldn't do that with any of the other languages. You might be writing with a better language, but you would get your work done a couple of weeks later, instead of getting answers. I used IMP.
CLB: Was IMP being used at Stanford?
Knuth: It was at a research lab in Princeton. A year before I came to Stanford, I worked there on a classified cryptanalysis research project.
CLB: Please tell us about your other new book, The Stanford GraphBase.
Knuth: The GraphBase book is for two kinds of people. It has a research purpose; the people who are working on the study of new algorithms for combinatorial problems need a standard set of test data on which to compete with each other, and for benchmarks. As I was preparing Volume IV of The Art of Computer Programming, I decided that I would make all the examples and data that I'm using in that book available to everyone. There was a need for some standard benchmarks, and everything should be well arranged so that it is easy to use in thousands of ways. So... I now have a collection of thousands of standard data sets; anyone in Poland can have exactly the same data as anyone in California or China. It's very portable, and can be downloaded from the Internet.
The second purpose of the GraphBase hook is that it is an example of CWEB programming--- it's actually 32 examples of CWEB programming. They're short programs that illustrate the programming style that I prefer. The examples are like little essays, little short stories of computer programs, that are perhaps fun to read.
CLB: What is your current hardware and software environment?
Knuth: I use CWEB for my programming. I use the Emacs editor very heavily, and I use a great high-level language called METAPOST for drawing technical illustrations. This is a new language by John Hobby that is going to be released soon, I think. It's based on METAFONT. 75% of the code is mine from METAFONT, but it's fixed up so that it generates PostScript. I love it.
I also use Mathematica. The people at Maple are trying to convince me to switch over to Maple, another excellent system. At the moment, I like Mathematica because you don't have to type your multiplication signs; you can say "2X" instead of "2*X". Also, the Mathematica manual is exceptionally good.
CLB: You like Wolfram's writing style?
Knuth: Especially the index... you can easily find your way around that book. With the first edition, when I had a new problem to solve, I would look in the index and it would almost always refer me to the right page. There were three or four times when the word I tried wasn't there, and I penciled in where to look when I had this problem next time. In the second edition those had all been fixed, and I had not reported them to anybody.
CLB: Let me get your quick impressions on a few research areas, and whether you've read or done any work in them. The first is genetic algorithms. How do you feel about the general concept, that instead of the human determining the algorithm, you somewhat let the machine have at it...
Knuth: I plan to do a lot of experimenting on this as I get into Volume IV. There's genetic breeding, there's simulated annealing, there are other strategies that people have developed. I have a method in The Stanford GraphBase book that I call "stratified greed". These techniques are all competing for the same kind of problems, and I want to try a lot of examples; some of them might work better on one than the other, and I want to get a feel for this. Certain problems are naturals for neural nets... genetic algorithms are likely to do well on tasks related to language recognition, and people say also like predicting the stock market or something like that. Somehow the closer a problem is to nature, the more you expect the genetic algorithm to work, while the closer it is to number theory or something artificial, the more you expect some other kind of approach will help. It's hard to understand the way these methods scale up; on a small problem they might do terrifically, and then they might break down completely just when the problem gets a little bit bigger... or it might go the other way.
CLB: It sounds like you have several years of disciplined testing with your data sets ahead of you.
Knuth: The Stanford GraphBase gives me an unlimited supply of problems that I and other people can do. I read what other people have claimed about their methods, but I also try them all myself. The original work I do in The Art of Computer Programming is to take the methods of two different authors and analyze method A from the standpoint of author B, and method B from the standpoint of author A. They have only given their sides of it, so I try to fill in ....
CLB: What about object-oriented programming? Is it just a current buzzword, or does this approach appeal to you?
Knuth: I've always thought of programming in that way, but I haven't used languages that help enforce the discipline; I've always enforced the discipline myself in other languages. Programming languages can now catch you if you make a mistake, and they make it easier for you to hide information from one part of the program to another. In my own programs, with older languages, I wouldn't use what I wasn't supposed to use; I would have to discipline myself to follow these rules. I could, so I did. There weren't programs I couldn't write... but the new tools do help.
The problem that I have with them today is that... C++ is too complicated. At the moment, it's impossible for me to write portable code that I believe would work on lots of different systems, unless I avoid all exotic features. Whenever the C++ language designers had two competing ideas as to how they should solve some problem, they said "OK, we'll do them both". So the language is too baroque for my taste. But each user of C++ has a favorite subset, and that's fine. CWEB fully supports C++ as well as C.
CLB: What are your thoughts about chaos theory, fractals, those areas? Their indeterminateness seems a little discordant with the domains you've focused on in the past.
Knuth: I did some early work with fractals and so on, and I think it's a great new abstraction. People can build models that they wouldn't have thought of building before, that really match a lot of things in nature that have this character of looking the same when you change the scale. You know, if you magnify the coastline, it still looks like a coastline, and a lot of other things have this property. Nature has recursive algorithms that it uses to generate clouds and Swiss cheese and things like that. So now we have mathematical techniques for understanding such processes that go beyond the kind of differential equations that people used to have in previous centuries. Now we have a brand new tool to work with, but I'm not very intuitive about such methods. I know the limitations of my own intuition; I can solve some problems well, but I know other people are able to see something right away which takes me a long time... It's not my cup of tea.
CLB: To what extent have you ever followed developments in artificial intelligence? The third program you ever wrote was a tic-tac-toe program that learned from its errors, and Stanford has been one of the leading institutions for AI research...
Knuth: Well, AI interacts a lot with Volume IV; AI researchers use the combinatorial techniques that I'm studying, so there is a lot of literature there that is quite relevant. My job is to compare the AI literature with what came out of the electrical engineering community, and other disciplines; each community has had a slightly different way of approaching the problems. I'm trying to read these things and take out the jargon and unify the ideas. The hardest applications and most challenging problems, throughout many years of computer history, have been in artificial intelligence--- AI has been the most fruitful source of techniques in computer science. It led to many important advances, like data structures and list processing... artificial intelligence has been a great stimulation. Many of the best paradigms for debugging and for getting software going, all of the symbolic algebra systems that were built, early studies of computer graphics and computer vision, etc., all had very strong roots in artificial intelligence.
CLB: So you're not one of those who deprecates what was done in that area...
Knuth: No, no. What happened is that a lot of people believed that AI was going to be the panacea. It's like some company makes only a 15% profit, when the analysts were predicting 18%, and the stock drops. It was just the clash of expectations, to have inflated ideas that one paradigm would solve everything. It's probably true with all of the things that are flashy now; people will realize that they aren't the total answer. A lot of problems are so hard that we're never going to find a real great solution to them. People are disappointed when they don't find the Fountain of Youth...
CLB: If you were a soon-to-graduate college senior or Ph.D. and you didn't have any "baggage", what kind of research would you want to do? Or would you even choose research again?
Knuth: I think the most exciting computer research now is partly in robotics, and partly in applications to biochemistry. Robotics, for example, that's terrific. Making devices that actually move around and communicate with each other. Stanford has a big robotics lab now, and our plan is for a new building that will have a hundred robots walking the corridors, to stimulate the students. It'll be two or three years until we move in to the building. Just seeing robots there, you'll think of neat projects. These projects also suggest a lot of good mathematical and theoretical questions. And high level graphical tools, there's a tremendous amount of great stuff in that area too. Yeah, I'd love to do that... only one life, you know, but...
CLB: Why do you mention biochemistry?
Knuth: There's millions and millions of unsolved problems. Biology is so digital, and incredibly complicated, but incredibly useful. The trouble with biology is that, if you have to work as a biologist, it's boring. Your experiments take you three years and then, one night, the electricity goes off and all the things die! You start over. In computers we can create our own worlds. Biologists deserve a lot of credit for being able to slug it through.
It is hard for me to say confidently that, after fifty more years of explosive growth of computer science, there will still be a lot of fascinating unsolved problems at peoples' fingertips, that it won't be pretty much working on refinements of well-explored things. Maybe all of the simple stuff and the really great stuff has been discovered. It may not be true, but I can't predict an unending growth. I can't be as confident about computer science as I can about biology. Biology easily has 500 years of exciting problems to work on, it's at that level.
CLB: Use of the Internet is exploding right now, with everyone getting on...
Knuth: Some day we are going to try to figure out who is paying for it!
CLB: Do you currently use it? I know you did in the past.
Knuth: I spent fifteen years using electronic mail on the ARPANET and the Internet. Then, in January 1990, I stopped, because it was taking up too much of my time to sift through garbage. I don't have an email address. People trying to write me unsolicited email messages get a polite note saying "Professor Knuth has discontinued reading electronic mail; you can write to him at such and such an address."
It's impossible to shut email off! You send a message to somebody, and they send it back saying "Thank you", and you say "OK, thanks for thanking me..."
Email is wonderful for some people, absolutely necessary for their job, and they can do their work better. I like to say that for people whose role is to be on top of things, electronic mail is great. But my role is to be on the bottom of things. I look at ideas and think about them carefully and try to write them up... I move slowly through things that people have done and try to organize the material. But I don't know what is happening this month.
So now I don't read electronic mail, but I do use it occasionally. Say I'm taking a trip to Israel and I've got to make last minute arrangements. When I visit another university or research center for a few days, I have to send email from there. I've learned how to use the email facilities in Emacs, but I don't want to get good at it.
CLB: You have many interests outside of computing and mathematics--- music, religion, writing. Is music a creative outlet for you, a means of recreation, or a spiritual outlet?
Knuth: At the moment it's recreational. I like to have friends come to the house and play four-hands piano music. If I could do it every week, I would. I hope to live long enough so that after I've finished my life's work on The Art of Computer Programming, I might compose some music. Just a dream... it might be lousy music, of course.
CLB: You have written some compositions already, haven't you?
Knuth: Yeah, but it was mostly arrangements of other people's themes. I did write a short musical comedy when I was in college called "Nebbishland". Remember how Nebbishes were all the rage in the late 50s? "Nebbishland" was only about a ten minute skit, but it was all original music and lyrics.
CLB: Do you have the score somewhere in the attic?
Knuth: Yeah... no actually, I think I've lost it. I have only part of it. I'm hoping to come across it. I'm going through my files now and making a computer index of everything I have in the house.
CLB: Sounds like you don't have a paperless house!
Knuth: No!
CLB: Have you fiddled with MIDI computer technology for music, or have you purposely stayed away from it?
Knuth: I have fun with it. I bought a synthesizer for my son last Christmas, and I played it for hours and hours. I loved it. I had once played on a Kurzweil synthesizer years ago, at Marvin Minsky's house, a grand piano imitation. More recently, a friend went to England for three years and didn't want to bring his grand piano him, so he bought a Yamaha with six voices. When I visited his house, I had a tremendous time for three days going through all of the pieces I'd learned on the piano, playing them as if they were on vibraphone, or on a harpsichord, or some other voice. His "piano" has a harpsichord voice, but the keyboard is pressure-sensitive, so you can play loud and soft, which you can't do on a real harpsichord. These synthesizers are great.
CLB: When did you retire from Stanford?
Knuth: This year. I was on leave for two years until I could officially retire. Unofficially, I retired in 1990, on the same day I gave up email. I announced my plans three years earlier. I realized that my main goal in life was to finish The Art of Computer Programming; I had looked ahead and seen that it would take twenty years of work, full-time. If I continued doing everything else that I was doing, it was going to be forty or fifty years of work. I was just not getting anywhere, I was getting further and further behind. So I said, "Enough." Naturally, I hate to give up many of these other things that I like doing very much. But there are some things I didn't hate giving up, like writing proposals. I'm very happy to give up those!
CLB: You had to write proposals?? I assumed you were insulated from that somehow.
Knuth: You've got a great sense of humor! I don't have to do it anymore; but as a professor, in order to have decent equipment for my grad students, or to have visitors for active research programs, to publish reports, etc., I needed to find sponsors. It's a lot of work begging for money. The System Development Foundation said they'd give me a million dollars so that I could finish TeX and get back to The Art of Computer Programming.
CLB: Did you take them up on it?
Knuth: Sure, but it still took many, many years to finish TeX. I decided that the only way I would be able to finish The Art of Computer Programming is by going into full-time writing, and being a hermit, and telling people "No." It was hard to adjust the first couple of years. Now I feel real efficient, and the writing is going well. A nice steady state.
I give lectures at Stanford every month or so, when I'm in town, called "Computer Musings" . I plan to keep this up for twenty years, to give a talk on whatever I find interesting that month, on neat ideas I've picked up... I bring up problems that I can't solve, so that somebody will do it for me. Now, if I can't solve a problem in two hours, I've got to give it up and tell somebody else to work on it; otherwise, I'll get behind again. As I write the book, I've got to move from topic to topic, and my attention span is maybe three weeks on any particular topic.
CLB: You're best known for your writing and research; did you enjoy teaching and the interaction with students?
Knuth: We had the greatest students in the world. I can still get together with students through my lecture series, except I don't know their names anymore. That's a problem.
CLB: No student interns?
Knuth: Suppose I give a "Computer Musings" lecture, stating an open problem, and suppose that a student in the audience solves that problem, writes his thesis and finishes it in the next two weeks (maybe two and a half), and shows it to me. Then I'd still be interested in the topic, would still read it, and I'd be glad to sign his thesis... but that's the only way. 28 is the total number of Ph.D. students I've had graduate, and that's probably all that I will have... unless something happens at high speed through the "Computer Musings".
CLB: Real-time Ph.D.'s! What changes have you seen in the students coming into the computer science program over the years?
Knuth: There is a very profound change that I can't account for. In the 70s, the majority of our students were very interested in music. The first thing we'd ask them when they came in was "What instrument do you play?" We had lots of chamber groups and so on. Now almost none of the students are interested in music. I don't know if it's because a different kind of people are enrolling in computer science, or because it's true of all today's students, or what. If you ask computer science students now what their hobby is, the chances are most of them will say "Bicycling". I recently had one who played a harmonica, but there were almost no musicians in the group.
CLB: Any changes in the quality of the students?
Knuth: Not the quality... but they don't know as much about mathematics as they used to. We have to do more remedial stuff in college, even at a school like Stanford.
CLB: How about changes in the field itself... with so much progress and so many more people involved, is computer science today very different than it was earlier?
Knuth: Well, there's all the media and the visual things, that's a lot different than it was. There's also the competition; it's a great deal more difficult now than it was in my day. When I started, it was so easy to come up with something new compared to now, when you've got thousands and thousands of smart people all doing great stuff. There might have been ten great Ph.D. theses a year at one time; there's just no way to keep up with all the stuff now.
No matter what field of computer science you're in, everyone is finding it hard to keep up. Every field gets narrower and narrower, since nobody can cover all the territory anymore. Everybody can choose two small parts of computer science and learn those two parts; if one person knows parts A and B, another knows B and C, and another knows C and D, the field remains reasonably well connected, even as it expands.
CLB: Do you see yourself as one of the last of computer science's "Renaissance Men"?
Knuth: I'm not as broad as you might think--- I only work on one thing at a time. I guess I'm a quick study; I can become an instant expert in something. I've been collecting stuff for thirty years so that I can read the literature on each topic in "batch mode"--- not swapping lots of different topics in and out. I can absorb a subject locally and get good at it for a little while... but then don't ask me to do the thing I was doing a few months ago! Also, I have lots of people helping me correct my mistakes.
CLB: My last question, your least favorite to be asked... what is the current plan for completing all seven volumes of The Art of Computer Programming?
Knuth: I'm going to have fascicles of about 128 pages coming out twice a year. We're gathering four of them before we come out with the first two actually; we're going to keep some in the pipeline! Look for the first fascicles in 1995 or 1996; they will be beta-test versions of the real books. I'm thinking I can finish Volume IV (parts A, B, and C)in the year 2003, Volume V in 2008, then come out with new editions of Volume I, II, and III, then work on VI and VII... There will be a "Reader's Digest" version of volumes I through V.
CLB: What would your career, and life, have been like had you not announced the 7-volume set?
Knuth: Oh, I didn't announce it at first. I thought I was writing only one book. But if I hadn't done that, I suppose I still would have been doing a lot of writing. Somehow it seems that all the way through, I've enjoyed trying to explain things. When I was in high school, I was editor of the student paper; in college I edited a magazine. I've always been playing around with words.
Acrobat works fine in linux. I'm currently using the plugin in galeon and it displays fine. No need to use windows.
n uth.pdf
Here's another mirror
http://www.cs.columbia.edu/~spotter/fea-k
got this in an email a few days ago....
Richard M. Stallman, Linus Torvalds, and Donald E. Knuth
engage in a discussion on whose impact on the computerized world was the greatest.
Stallman: "God told me I have programmed the best editor in
the world!"
Torvalds: "Well, God told *me* that I have programmed the
best operating system in the world!"
Knuth: "Wait, wait - I never said that."
The AMS site is slashdotted--I did manage to download and view the PDF file and the above is a copy of the interview.
http://access.adobe.com/adv_form.html
just enter the url of said pdf and hit submit and voila good ol' html is returned
But I think that it's important to note that Donald Knuth, like many other brilliant men, is a Christian. Thus, it's unlikely that he would presume himself to be God.
:-)
It *is* funny, though.
Regards,
Anomaly
But Herr Heisenberg, how does the electron know when I'm looking?
Well, I would like to ask the slashdot editors to please post more Knuth news. I think that his contributions to Computer Science deserve to be more widely appreciated by everybody even remotely connected with Computer Science (and not only the theoretically inclined people).
Unfortunately, I recognize that not everybody reading Slashdot has a theoretical education in Computer Science (well, many people are only practically trained -- if such a thing even exists) and miss the elegant construction of algorithms that Donald Knuth does in his books, algorithms which are efficient both regarding space and time (things which I miss in most software being written today, sadly).
This is not to mention the care with which his books are written, from didactic, technical and typographical standpoints: a lesson on how to write well.
I guess that the problem I mentioned above about current programmers writing code which is not exactly space- and time-efficient is that they must think "it's not worth it" (or many haven't actually even thought about the subject). A pity indeed.
This is, unfortunately, one of the bad sides of the ease of current (integrated) programming environments (which doesn't man that they are bad): people which aren't exactly trained can program, their programs run, but in a sub-optimal way.
I also think that many programming environments are an incentive to trial-and-error programming ("recompiling the program is too easy -- don't even bother to think if we have to add 1 or subtract 1"). This, of course, leads to sloopy programming.
Anyway, back to Knuth, I would really love to see a Slashdot interview with him, as I would appreciate anything regarding him and computers.
How dare they use an openly documented, commonly used format.
Danny.
I have written over 900 book reviews
Then I happened to sit next to an engineer at a big radio dinner and speech thing on a holiday party cruise - we picked out all the fonts, figured out where the photocopier for the custom meny had drum faults versus schmutz on the glass, and walked over to the corner, where they started laying out the menus, and found the printed original to verify it was laser printed.
Ghod, the speeches were boring.
--
Evan
"$30 for the One True Ring. $10 each additional ring!" -- JRR "Bob" Tolkien
I have 800x600 laptop and it is pain in the arse to
read this 2 column PDF file on it. I have to scroll
up and down all the time.
The guy is smart, but his choice of format is more suitable for old, soon to be obsolette printable media not for the Internet.
It is boring - a bunch of electronically enhanced cars going around and around 70 times.
Knuth: There's one man who lives near Frank-
furt who would probably have more than $1,000
if he cashed all the checks I've sent him. There's a
man in Los Gatos, California, whom I've never met,
who cashes a check for $2.56 about once a month,
Seems to me Microsoft has been pulling a reverse knuth. They have $2.56 from you for every bug they create!
$15 billion is about 5,859,375,000 bugs..
Taking into account MS has been around for 20-25 years... that figure sounds about right!
Apparently Richard Feynman, on the last day of classes each semester, made the class an optional thing where students could come and ask questions on any topic except religion, politics, and the final exam. And Knuth followed his example.
Now that I'm teaching, I'm thinking of trying that. I can't decide if I really want to exclude religion and politics, though. I wonder if they excluded those topics to avoid offending people, or because they thought those topics are too subjective/personal, or if it was for some other reason?
I also wonder if anyone would come on that day; the last day of classes in the spring at Cornell is "Slope Day", where all the undergrads hang out on the hill by the main library and get drunk (the police basically look the other way, as long as people aren't getting hurt). It's truly a sight to behold.
I don't know if that was intentional, but it does convey some of the sense of blithely ignoring the consequences of what we do.
Knuth truly does represent much of what is good an interesting about our profession.
..." is still frighteningly true; other engineering professions do not often have a commonplace equivalent of a blue-screen or core-dump. There are occasional engineering failures, but none as widespread as programming errors.
For those who didn't read the article, or didn't come across this fact elsewhere, Knuth actually personally writes a cheque to anyone who finds errors in his books.
While the algorithms and theory that he wrote about in his classic texts are used by computer programmers worldwide every day, it's unfortunate that the kind of pride of workmanship that he personally demonstrates, doesn't seem to be the norm.
I've always felt like the programming profession was, and still is, a bit of a joke as far as standardized quality goes, as compared to other engineering disciplines. The old joke, "if builders built buildings the way programmers wrote programs
Maybe because we're still forging new ground so quickly that it can't be expected to have solid results. Still, for things as standardized, commonplace, and essential as operating systems, the design should be such that a blue screen is unheard of.
-me
Love many, trust a few, do harm to none.
According to Knuth the last character of the official name is actually a capital Greek "Chi," not a Latin 'X'. Thus most Americans will pronounce it "tech." However, the Greek letter usually written in English as "Chi" is actually pronounced as an 'H' in Greek. Thus TeX really *should* be pronounced with an "h" sound at the end. I'm not sure how you'd write this in English (as "teh" would make the 'h' silent). Of course this goes for all other Greek-derived words ("character" and so on) as well.
10 PRINT CHR$(205.5+RND(1)); : GOTO 10
- Donald Knuth
That last comment is one of my favourite computer-industry quotations.
You know, I went and installed Mandrake 8.1 on my crappy Celeron 500 this past January and I was pleasantly surprised with the eye-candy-riddled installer, etc. I installed the office workstation package group. When I got done, I wrote a letter, and soon discovered that TeX was not included! What the hell? I get the impression that the folks at Mandrake are too busy trying to make their product look like an MS-bundle system to include a decent word processor. I think the world would be a better place if folks would try to stop imitating the WYSIWYG giants (Corel, MS, SUN) and start evangelizing about TeX and LaTeX. I can't remember the last time I used a WYSIWYG word processor when I had a choice.
As a mathematician, I am forever grateful for Knuth and Lamport. They are the Gutenburgs of our time.
There was a /. story a year or so previous about some Knuth lectures where he talked about his faith. Also, a brief search turned up this.
Given that outfit, I don't think I could have refrained from asking him about the Dark Side of the Force.
If I were learning this stuff today, I'd probably learn Tex / LaTeX, but the stuff really is ugly. I'd rather have a good math package for html.
Bill Stewart
New Fast-Compression-only CPR http://preview.tinyurl.com/dy575ks
Should really be proud to work with a legend. ..
:-)
I have been coding in C for six years, without
seeing the internet, then I saw the picture of
Denis Ritchie, and I almost brusted in tears
ahem, I actually did.
I have read about Ritchie, Kernighan, Thomson,
Pike, and the rest of the bell guys (from books.)
Then I got into BBSing, and read source code writen by
Joy (and the rest of BSD), McKenzie (and few other GNUers.)
, Kees J Bot and the Minix hackers at Vrije or just around MINIX,
Bob Stout, Paul Hsieh, Terje Mathisen, Delorie, the DOS extender
scene lead by TRAN and many many others who have release public
domain Unix and DOS source code.
I have read almost every line of code, I could lay my hands on. To
the point I was able to reuse a huge table of hex #defines, I ripped from
a compression code in another project (I didn't know about perl, and
generating constants for table driven code was a bitch. I knew about
code that had the exact same values
Then I met Knuth. It was a new birth. It had the same effect on me, as
reading Abelson&Susman had, a bird's eye view of all that is there.
It took me 11 months to learn Chapter 1 (without the MIX specs!) but after
that, it was a revelation, and I am never the same man ever since.
MathML is intended to be such. Without some additional formatting work it wouldn't be a good typesetting language though, as (in the tradition of HTML) it merely specifies the mathematical formulae, leaving their exact rendering up to the browser (while for typesetting you probably want some more exact control).
10 PRINT CHR$(205.5+RND(1)); : GOTO 10
As an example of what can go wrong, look at your average TeX-written math/cs paper on your average PC screen. The font's too grainy and greeky to read at 75-100dpi, and it's probably in some two-column format that looks really nice printed on portrait-mode dead trees, but is horrendously annoying to read on a portrait-mode screen that can only display about half a paper page at bad resolution. Arrrgh!
Somebody once commented that there are better renditions of the TeX fonts for PCs - I think it was a TrueType implementation of CMR fonts or something, but it's been too long to remember the correct details.
Bill Stewart
New Fast-Compression-only CPR http://preview.tinyurl.com/dy575ks
I tend to agree, LaTeX/TeX is pretty horrible for computer-readable stuff. It's basically designed for typesetting printed pages, and is only really good at that. But many papers are available like that because the authors are writing them primarily for journals, and outputting to a PDF they put up on their website is the easiest way to publish it online.
So yes, for web stuff, I think MathML is the best choice. It's now supported by default in the latest Mozilla; I don't know about other browsers. Unfortunately it's not usable in a standard HTML document, but only in XHTML documents (XML/CSS basically), which makes it have a bit of a learning curve. But then again LaTeX certainly has a learning curve...
The main argument I've seen against MathML by mathematicians is that it's clumsier than LaTeX to use a lot, especially for those who are already used to LaTeX. Basically this is due to the nature of XML tags -- and pairs are going to be more clumsy to use a lot than LaTeX's \( and \) commands.
10 PRINT CHR$(205.5+RND(1)); : GOTO 10
The last half of the last sentence should read "-- and and pairs are..."
10 PRINT CHR$(205.5+RND(1)); : GOTO 10
Did you even read the link you posted?
He explains in perfectly understandable English exactly why he doesn't use email (note the lack of hyphen, which is also explained in the document you linked) - aside from the very occasional post to someone via his secretary. Try reading that document a few times, maybe you'll get it.
Communication via email may be appropriate for you. It's certainly appropriate for me, and probably for the vast majority of people who read and post to slashdot. But Knuth has made the decision that it's not appropriate for him anymore - note that the guy did use email for 15 years, from 1975 to 1990 - so he is perfectly aware of what he's missing out on (not much) and what he's gaining (a hell of a lot of extremely valuable time) by choosing to not use email.
Perhaps you should reconsider the way your system of respect works.
Pete.
PS. I have no idea what made you think that he's proud of not using email.
F1 isn't everyones cup of tea.
However, it's still as much about drivers as it ever was. With reasonable confidence, i can say that no amount of electronics will make you as good a driver as schumacher or montoya. After all, we're talking about drivers that can out-brake ABS, out perform traction control, etc etc. Electronic aids only go so far. And the cars dont steer themselves (yet).
That the cars happen to shift gears faster than a human ever could is a nice touch. Friday i rode in a car (BMW E46 M3) which had an electronic clutch system and the shifts were breathtaking.
My opinions are my own, and do not necessarily represent those of my employer.
I usually generate two outputs of my LaTeX documents: Postscript and PDF. The PDF version usually looks a bit better on screen than the PostScript version.
mind you: I generate the PDF version using pdflatex . I can't remember exactly, but I think I've seen a utility that converts DVI files to PDF and that this produced horrible output. use pdflatex .
because generate multiple output formats (PS, PDF and HTML) from the same LaTeX document, I usually use a package I wrote that contains a lot of convenient macros to make use of the different features in the different formats -- in addition to automating a lot of boring tasks.
I remember how delighted I was when I discovered pdflatex. once you work it into your repertoire you get all this cool stuff for free, like hyperlinks in PDF documents etc. I really recommend you give it a try.
Check out the PDFTeX web page for more information.
Also have a look at Matt Welsh' page about creating presentations in PDFLaTeX. He has some useful information on how to install TrueType fonts for use by PDFLaTeX.
-Bjørn
Comment removed based on user account deletion
It's Teh-Ten, damn it! - Steve Jobs
Last holidays, I was given a copy of Things a Computer Scientist Rarely Talks About, a transcript of a series of six lectures by Knuth on the subject of religion. (Why six? "Because I could only think of six funny things to say for the lectures... and that was the first one.")
It's fascinating. Much of the lectures he talked about his 3:16 project, and the rest was an open Q&A session. Even if you're a hardcore atheist, you should pick up a copy; there's no preaching, and the audience discussion is far more insightful than the crap you're paying subscriptions for here on slashdot.
(Random bit of trivia from the book: for the project, Knuth used his own Bible translations. One of the lectures is mostly a "Bible-translate.HOWTO" commentary.)
You cannot apply a technological solution to a sociological problem. (Edwards' Law)
I'm speaking of modern Greek, in which the Greek letter X, or "Chi," is supposed to sound like an "H". The hard G sound is created through two ways -- either a double-gamma dipthong or a gamma-kappa dipthong.
Note also that this most Greeks pronounce ancient Greek in this way. It is only non-Greeks who pronounce classical Greek in the way you describe, and this interpretation is not accepted by any Greek scholars I know of. There's really very little evidence either way for how classical Greek was pronounced, so most Greeks think that the fact that most classicists pronounce classical Greek with similar sounds as those found in German is simply a result of them being mostly Germans, not because this is actually the correct way to pronounce it.
Simply put, most classicists can't pronounce Greek, so they mispronounce it in their own language and decree that this is how it "used to sound".
10 PRINT CHR$(205.5+RND(1)); : GOTO 10
Yes, that's the German interpretation of classical Greek. As Germans were the leading classicists for most of the 20th century, it's accepted as the "standard" way of pronouncing ancient Greek, but it's really nothing more than mispronouncing Greek with a German accent. If you speak to any Greek person you won't find any pronounciations remotely similar to that (whether they're speaking modern or ancient Greek).
10 PRINT CHR$(205.5+RND(1)); : GOTO 10
I know a good fraction of my students care about my views, and I care about theirs. A few are there just because they need the credit to graduate, which is unfortunate, since their hearts aren't in it, and I don't get to interact with them as much.
:-)
I know my students respect me, perhaps largely because I respect them. (That works well in a lot of places, not just the classroom.) A group of my students from a previous semester even got together and wrote a letter to the chair of my dept. on my behalf about my teaching skills and about how much I care about them, which was nice. This is only my second year teaching, and as one student wrote on the course evaluation form last year, I'm not bitter and angry yet like the other faculty.
As for questions about my sexual activities, I think the students would feel more embarassed asking such a question than I would answering it. If they had the guts to ask it, I'm pretty sure I'd answer it.
That actually looks pretty useful.
10 PRINT CHR$(205.5+RND(1)); : GOTO 10
I'll certainly be happy to use better display hardware when it's affordable, whether on CRTs, laptops, or PDAs. But I'm finding this article perfectly readable on a grainy CRT, while the typical latex->dvi->ps->pdf is pretty illegible unless magnified enough to show a small window into the picture. But even with better displays, I expect my laptop to be landscape-mode, not portrait, and my PDA to have way too small a screen. NeWs postscript-based window system was wonderful, at least compared with X or MSWindows, but there are still jobs for an content-description language like HTML/MathML rather than a page-description language like Postscript.
Bill Stewart
New Fast-Compression-only CPR http://preview.tinyurl.com/dy575ks
At any rate, if you are confident in your assertion, I suggest you have a *lot* of reading to do.
i've never met him personally, but i really have alot of respect for him. although i dont use tex directly, i do use latex on a daily basis. after using word as an undergrad, i am very grateful for his work. i do agree it was a good joke.
-- john
> Adolph Hitler was Catholic.
Only by heritage, at most he paid lip service.
> The Vatican owns Great Britain and the United Nations.
Ridiculous assertion, can you back it up? Great Britain is at odds with the Vatican for five hundred years already, and the United Nations actively promote actions contrary to the Vatican like artificial birth control.
> The Catholic religion was started by the Romans.
‘Catholicus’ means ‘universal’ – it’s not a proper name, but simply the expression of the Roman church pretension to be universally dominant. Anyway Romanism is a mixture of Christianity, heresy and Roman political institutions and structures. Nothing is as simple, you should read more than fanatical anti-Romanist tracts.
> The Catholic Church is evil.
Man is evil, Romanism is a human institution, therefore Romanism is evil. Man was created in God’s image, Romanism was created by man, therefore Romanism still has some reflections of God’s image, distorted as it may be. The same holds true to any other religious organization, from Eastern misticism to localist congregational protestant churches, but in varying degrees.
Again, nothing is so simple, neither God nor Man.
Leandro Guimarães Faria Corcete DUTRA
DA, DBA, SysAdmin, Data Modeller
GNU Project, Debian GNU/Lin
> I think Microsoft should say, "You'll get a check from Bill Gates every time you find an error"
Obviously we’d need MS Windows and MS Office source code to be able to find errors – without that Microsoft will usually just push it to the hardware or other aplications or specific configuration or user errors in the system tested.
Leandro Guimarães Faria Corcete DUTRA
DA, DBA, SysAdmin, Data Modeller
GNU Project, Debian GNU/Lin
It's not a mirror, it's not a different interview. It is quite funny, however.
:wq