The BCS and IET this year honoured Donald Knuth in its Turing Lecture. Knuth is now professor emeritus at Stamford. His achievements are legendary. Computer scientists will recognise his name primarily from The Art of Computer Programming (TAOCP), of which he has just published the most recent volume in a lifetime series. Others may know him better from the TEX typesetting program; I count myself an early adopter though for me it gave way to WYSIWYG word processing when that arrived on my desktop.
Most honorands give a lecture. Knuth, citing Richard Feynman as his inspiration, offered a format akin to “An audience with …”, inviting questions on any topic. Well, more or less; he did block one or two! but with four hundred people onsite at the IET, and countless more watching the Internet relay, there were plenty to go on.
The format meant that I can’t report a strong view of the future of IT, or a retrospective on a very influential life, or anything like that. And you can watch the whole session online, or perhaps attend repeat events in Manchester, Cardiff and Glasgow. But a few nuggets.
Inspired by a fringe event at an Object Technology conference some years ago, and by the fact that Knuth is also a recognised organist, I asked about computing paradigms which have parallels in music. He’d just answered a previous question with a put-down on the subject of re-use (“we don’t re-use; we re-edit”) so I didn’t refer to the many examples of re-use in the music of baroque composers such as Bach or Handel but to concurrency in works for the organ (two hands, two feet, all potentially working independently). Don’s response mentioned combinatorial techniques (his main recent focus) and the use of form and constraints (such as sonata form, in music).
In the next answer, to a question about trends in programming languages, he asserted that there will always be complexity, there is no optimum language, and “It’s going to get worse”. He likes special purpose languages; but “the fundamentals are not subtle – I like whatever is supported by a good debugger”. And he prefers languages “close to the way the machine works”; I take issue there, since the world’s worst language was designed the way an IBM machine of the 1960s worked and the language is still around but machines aren’t like that any more.
Other questions covered Jevons Paradox; the Lithos font; and agile programming (which gave him a track to plug his other recent book, “Selected papers on Fun and Games, with a nice c-t ligature on the front cover title. As a typographer myself I may have been the only other person there to appreciate this..).
He talked about archiving historic code: how do you do this? We are better at preserving hardware than software. Some is only available now as program listings, and OCR isn’t accurate enough.
And why “The Art of Computer Programming” not “The Science …”? As he said: Science is what we understand well enough to explain to a computer. Art is everything else.
A couple of other impressions. He showed pages from his books, directly on screen, using a technology I thought had disappeared long since: the epidiascope. But this one was computer-driven via the main screen. Perhaps that mix of ancient and modern epitomises a man who, as the bio we were given says, hasn’t had an email address since 1990. Typography is an ancient skill, craft or art (or all three) and yet Don Knuth almost single-handed brought it into the computer age: the same mix of ancient and modern, and still evolving. A giant of the computer age.
• Donald Knuth: Wikipedia (accessed on 2 Feb 2010) with bibliography and other aspects not mentioned here
• An evening with Don Knuth – all questions answered, Turing Lecture 2011 (1 Feb 2011); click the link on the mugshot to open the replay
• Don Knuth’s website at Stanford
• There are many links for TEX: try Google!
• The Lithos font (it’s an OpenType font, so there are many potential sources, but this reference from Linotype shows a fairly complete family set)