Donald Knuth - Developing Metafont and TeX (62/97)

To listen to more of Donald Knuth’s stories, go to the playlist:    • Donald Knuth (Computer scientist)  

Donald Knuth (b. 1938), American computing pioneer, is known for his greatly influential multi-volume work, 'The Art of Computer Programming', his novel 'Surreal Numbers', his invention of TeX and METAFONT electronic publishing tools and his quirky sense of humour. [Listener: Dikran Karagueuzian; date recorded: 2006]

TRANSCRIPT: Somebody suggested that I call it Metafont, because it... it wasn't my... yeah... it wasn't something I... I made up myself and I can't... I think it was maybe Bob Filman, but I can't remember for sure. But as... as we were looking at these things up at the lab, there was lots of visitors and people talking to me all the time, and that... that was a nice name because... because of the parameters. I'm not designing just one font of type, but I'm designing a font... font that varies with the... with the specifications, with... you know, it varies substantially with... with different specifications, and the... the... and so there's lots of like words... like metamathematics means going a little beyond mathematics or metaphysics... metaphysics really meant after physics in... in Greek, but... but the word Metafont turned out to be kind of appropriate. Now, well... it's something like morphing, in a way that you can go from one shape to another, but - that's why I'm thinking metamorphosis here - and... but the... the... my first programs for fonts were... each... each letter was drawn by a specific computer program written in computer language. I had to... to change that so that... so that I... I could devise a special language just for fonts, so that it was easier to make changes and... and to make the design. Instead of writing... instead of writing computer code in a the language like ALGOL, I could write now in this Metafont language, and then I would have an interpreter that would... that would read the Metafont language and draw the shapes instead of the original way, which was calling subroutines that would... that would draw the shapes. So, it was a completely different kind of a language than had existed before and... but I came up with the first design of a Metafont language in 1979. Now, as the number of users grew from one person to 10 people, I needed... I needed to change the system because they had different, you know, they'd discovered different bugs, they had... they needed different... different facilities. Then it grew from 10 to 100, and again, I had to add more stuff to the... to the language. By this time it was called TeX instead of Tex because there was already a system called Tex, and the... Honeywell refused to give us permission to use that name, even though their system was an operating system, not a typesetting system. We couldn't do that. So, we said... we said well, this isn't Tex, this is TeX, this is Greek, Tau Epsilon Chi, it's not... it's not an English word at all, it just happens to look very much like the English word Tex. And well, we told all the users to pronounce it TeX, and, but their lawyers didn't think... didn't think that wasn't much... enough of a distinction, so we... we left it on paper that, you know, if they want to sue us that we... we didn't get an official registered copyright, we just... we just have it as a copyright, not a registered copyright. Metafont on the other hand, is a registered copyright. Nobody was complaining about that name. Well, I was going to say, as we get... as we get 10 times as many users, then we need... then I have to go back and retool the system. This can... this can be too long a story, so I'd... I'd better start summarizing. The first version of TeX - a prototype version - was called TeX 78, because it... it saw the light of day in 1978. Then there was Metafont 79. But these are quite different from the TeX and Metafont systems that we have now. These were the first versions.

I have another policy that I recommend, and that is if you're doing a first generation piece of software, get it working, do your best, try to get it... get it so that it's perfect and then get a lot of experience with it, then throw it away and start over. Scrap it and don't worry about being compatible with the old, because there are going to be many, many more users in the future than you ever had in the past, even though there might have been thousands of users in the past, there... there will be more, you know, tens of thousands in the future and they'll all thank you for having a better... a better system. I don't do it... the second one I keep though. And now there's so much invested in it that... that I'm... I'm sure that I would never want to use a system that, you know, I... I never want to change again to a... to a system that's incompatible with the one's... [...]

Read the full transcript at https://www.webofstories.com/play/don...