Notes from Valentine’s Day

Here are the said notes from Knuth’s lecture on Valentine’s Day 2008. There was no specific topic that the Don spoke on. It was more of a Q/A session than a lecture. However there were few short questions and long insightful answers

(A day after the lecture, I did a full transcript which I seem to have misplaced. What follows is a reconstruction from the short-hand notes that I took during the Don’s talk. Due to the peculiar way in which Long-Term memory is reinforced in the human brain, we are capable of making grossly false recollections and delude ourselves into beliving it to be the gospel truth. I have tried best to stick to the words that I took down with no superfluous explanation. But be warned about the limitations of the cerebral cortex. The gist of the ideas and most of the nouns, verbs and adjectives were mostly by the original speaker. But the correctness of pronouns, infinitives, sentence ordering and in general a guarantee of authentic tone are in no way implied.)

Q: What was the most important significant event in CS history that was surprising to you for (a) having happened. (b) having not yet happened.

A: “I am a mathematician not an economist. I don’t believe in the model of history where you take the ‘Top X’ events and call it significant. All progress in CS or Mathematics has been incremental. It has not been big breakthroughs”. Knuth then talked about the evolution of Operating Systems – how the idea of Time Sharing Systems evolved from McCarthy’s IBM 704 to Multics to Unix to what it is today.

Q: What was the most important breakthrough that you did not think would happen and did indeed not happen ?

A: “Understanding how human cognition works”. Holistic principles behind human abilities like vision – which is incredibly difficult for a computer to generalize in non-pixel terms. Knuth spoke something about “seven cycles” taken to understand the meaning of something. Google or me fail to grok the nature of these “cycles”. Maybe he was using a colloquial expression in place of a rigorous neurological term. The Don also suggested that we embark on this quest by studying Ants and the nature of their intelligence.

Q: What is the most difficult usability or productivity issue in the computer systems of 2008 ?

A: The Don answered this differently. He talked about how most userland programs were still 32-bit and how operating systems and processors had moved on to 64-bits and made the point that in this scenario, half of the bits in the TLB were being wasted. He also talked about mistakes in his books and his $2.56 reward. There was this humorous part when he said, for the first time, he got a non-SPAM email from someone named “Candy” – who discovered a Tex bug for $ 32.768

Somehow, Knuth drifted to Linkers and Loaders and said in an other lifetime he would eliminate the concept of a runtime linker and would assemble programs as fast as the linker at runtime. Didn’t understand why he wanted to do this.

Q: Is Complexity good or bad ?

A: Knuth had a gag about how Dan Ingalls thinks Tex is bad. “The abstraction of complexity is the difference between mathematicians and computer scientists. Mathematicians seek patterns – They need a unifying principle to tackle complexity. Whereas computer scientists can work with non-unifying principles.”

Q: <Wasn’t clearly audible>

A: To learn how not to design a programming language, read Chapter 11 of “The Science of Programming” and implement the language B01 described in it. There was the interesting tidbit that Knuth worked as a Consultant with Burroughs – At which time he was oscillating between a career in Math vs. Computer Science. He explained how he had trouble with CS until he learnt the theory of programming languages. At this point, Knuth made the statement of the day: “Life is a Binary Search” – Meaning you keep oscillating between extremes until you find your niche in life.

People should be able to enjoy programming. Learn a new technique and implement it. Programming should have formalism and informal structures together – Literate programming Ideas should be a part of mainstream programming languages.

Q: What happens when human knowledge keeps on expanding ?

A: Knuth thinks the future will not be single specialization. It will be specialization in multiple fields – CS for physics or Maths for chemical analysis etc. Applied CS would be the future way of learning. As a sidenote, he joked about the target audience of his book – 98% (of what ?) does not care about my book or what I do and I don’t care about that 98%. My work is for the remaining 2%.

Q: What happens if P=NP ?

A: That depends on what power of P we are talking about. Whether P=NP at exponents as low as 2,5,10 or is it 10^4 or more ? It is possible that P exists but that we cannot prove it. One of the dangers of computer science is if it relies on math theorems whose effects as misunderstood.

He then mentioned about the God Lecture Series, and said “Computer Science is beautiful. But it is not everything in life.”

Q: What about the Parallel Processing future ?

A: “I am a sequential programming person. I can only think sequential algorithms”. It is just a matter of perception. “I think between recursion and iteration, Recursion is easier to understand”. I have no skills for parallel algorithms !

Q: So Parallel Programming is so hard that even Don Knuth has given up on it

A: There is one thing I can do really well and I am spending my life on doing it well.

The Don finished his speech saying he would be headed home to have a quiet Valentine’s Day Champagne with his wife.

I had to encode 2 hours of harmonic waves in the human audible frequency into 2 pages of ASCII characters, some 2 months after the event had taken place – Information theoretically, this process had to leak bits like a sieve. Nevertheless, hope that somebody finds it worthwhile.

UPDATE: For those of you who wanted to know which was the question that I asked, None of the above. Me and Friend who had dialled in, asked

We: “Which CS luminary whom you haven’t met and would like to meet ?”

DEK: What time is it in India ?

We: 3:30 A.M

DEK: What are your names ?

We: Shantanu and Ananth

DEK: You are the guys I would like to meet then, working so late ! (in Jest). “I’ve met most if not all of them.”

Filed under: Computer Science, DEK