Hacker News Reader

Comment by lanna | Hacker News Reader

Knuth is 80 and has been programming since at least the late 50s: https://en.wikipedia.org/wiki/Donald_Knuth#Education

In the 70's I had a co-worker, perhaps the best programmer in the department, that had gone to school with Knuth. He told me that one day while in college Knuth was using one of the available key-punch machines to punch his program on cards. My friend was ready to punch his program so he stood nearby to wait for Knuth to finish. Knuth, working on a big program, offered to Keypunch my friends program before finishing his own because my friend's program was shorter and Knuth could keypunch quite fast.

While watching over Knuth's shoulder, my friend noticed Knuth speeding up and slowing down at irregular intervals. Later he asked him about that and Knuth replied that he was fixing the bugs in my friend's Fortran as he punched it out.

(I wrote my first program in 1967)

Wow, that's a great story.

Some of Knuth's early programming exploits (from his undergraduate days) are described in “Stories about the B5000 And People Who Were There” by Richard Waychoff (see section III, “The Summer Of 1960 (Time Spent with don knuth)”): http://ed-thelen.org/comp-hist/B5000-AlgolRWaychoff.html#7

At the other end, he said recently that he still writes two or three programs a week, mostly in CWEB. (If you're not familiar with CWEB, it is his own invention: it's basically C, with preprocessors (ctangle and cweave) that enable literate programming.)

Some of his programs he puts up here: https://cs.stanford.edu/~knuth/programs.html — the most recent is from December 2017. (They are in .w format; if you don't have the patience to install CWEB and run cweave+tex on them, I've typeset them here: https://github.com/shreevatsa/knuth-literate-programs/tree/m... — last updated September; needs another update.)

I knew of WEB, which was pascal-derived. I guess CWEB is the C-derived equivalent?

Yes. WEB is Pascal+TeX, CWEB is C+TeX. CWEB is actually somewhat simpler, because for example the C language has preprocessor macros (#define) so the literate-programming system doesn't need to have its own macro system. It just has to rearrange (“tangle”) the input into the right order.

Thanks for typesetting the pdfs for other people to read!

I was going to claim that I've packaged CWEB for Homebrew so it's easy to get for MacOS users. I went to see if it still works and isn't moved to boneyard, and yes, it's there, but git log didn't show me in the commit history. Just to make sure I wasn't nuts I had to check my notes and found the original commit: https://github.com/Homebrew/homebrew-core/commit/77aa909188b... . Apparently there's been some history rewriting that I don't quite understand.

Oh great, I was just able to install CWEB with `brew install cweb`. Thanks for your work!

And actually `ctangle` and `cweave` already come with a TeX distribution like MacTeX. So the reason for providing a pre-typeset version is not so much that CWEB is too hard to install; it's just that people tend to not use it properly and either try to read the .w file directly, or (even worse) the generated .c file. :-)

> that enable literate programming

And if you've not actually dug into primary sources, you probably don't know what Knuth means by "literate programming" - what it's come to mean is a pretty small sliver of what was intended.

I thought it was pretty interesting that he no longer uses email. He did use it from (1975 - 1990), and he still works at Stanford.

https://www-cs-faculty.stanford.edu/~knuth/email.html

Well, except he does (or did as of that post) technically use email, just in the most wasteful and inefficient way possible:

> My secretary also prints out all nonspam email messages addressed to taocp@cs.stanford.edu or knuth-bug@cs.stanford.edu, so that I can reply with written comments when I have a chance

I had a client who would have his secretary print out all of his nonspam e-mails. He would hand-write his response on the print out, and then have his secretary fax it to us.

Eventually they decided to get more modern, and his secretary would scan in the response and send it as an e-mail attachment.

Being low or no tech is the ultimate status symbol! I remember the only billionaire I ever met didn't have a phone. He got memos from his assistants on paper and wrote his comments on them.

Not quite "no phone" but I know someone who was and is in senior technical roles at various well-known Silicon Valley companies. He only reluctantly got a (feature) phone--well after smartphones were the norm--when he had a kid and the nanny pretty much insisted.

Absolutely. I turned off my voicemail for awhile and forwarded my phone to the main number. It was liberating.

The part about scanning is how you responded to faxes in the old days. However the thing about making a quick note on paper and having a secretary who creates a nice reply is way more efficient than messing with mail directly.

Gosh, if only everyone had a secretary to expand a comment into a nice reply.

... to be read by the recipients secretary who then discards all the niceties and just tells the recipient "Acme said no on the Foo".

It sounds quite efficient for him. I would love to be able to do the same.

> in the most wasteful and inefficient way possible:

I mean, that's a drop compared to the paper waste in most companies. And it's not inefficient. Paper copies make annotations easier which makes replying easier and better quality.

If you can't think of a more wasteful and inefficient way to do that, you need to hang out more with government contractors. You ain't seen nothin' yet.

Also, this is unlikely to be more than a trickle. He talks about spending a day every three months responding to correspondence, as well as rather making a point out of refusing to respond to unsolicited emails. Chances are that his secretary also simply deals with a good deal of the email in the way, you know, a secretary does.

It is quite efficient for him: he spends a day reading email in “batch mode” every few months, and it saves him the cost of context-switching, and lets him focus on writing his books.

(Note that the only email he reads is bug reports: if you send something by email that is not a bug report, he says he ignores it, though I've been able to sneak in an odd thing or two (a comment on a TAOCP section) by including it along with bug reports… the catch of course is that to use this trick, you have to find an actual bug in the books.)

The emails are printed for him by his secretary on the back of one-sided printouts, so he's not wasting any “new” paper for the emails. He scrawls (with pencil) his replies on the printouts, and hands off the pile of outgoing mail to his secretary (who only comes in once a week or something), who contacts the recipients, asks them for their postal mail address, and sends his reply by post.

It works very well for him, and he doesn't have to waste time filtering email, typing into a computer, etc. He does everything with pen-and-paper (including writing programs like TeX and METAFONT over several months, without typing them up into a computer until the very end), so it just matches his most comfortable working environment.

He updated his web site quite recently though:

https://www-cs-faculty.stanford.edu/~knuth/news.html

"2018 began for me with an absolutely incredible 80th birthday celebration called Knuth80, held in the delightful city of Piteå in northern Sweden. ..."

So maybe he's just using that email-via-printouts thing as a ruse to avoid having to answer to emails from "important people" expediently? I would totally pull that move if I could.

I like how the old school legends prefer analog. Reminds me of Dijkstra's preference for typewriters.

On a similar theme, Tony Hoare might be in with a shout - active from approximately 1958 to now. He can still be found occasionally in the Microsoft Research building in Cambridge, UK.

Reply on news.ycombinator.com