BWK on "How to succeed in language design without really trying"

A talk by Brian Kernighan at The University of Nottingham.

Describes all the usual suspects: AWK, EQN, PIC. Even AMPL. I was wondering which languages he had in mind when he mentioned that some of his creations were total flops. I'd love to learn from those!

The talk is a fun way to spend an hour, and the audio would be good for commuters. For real aficionados I would recommend reading Jon Bentley's articles on the design of these languages (as well as CHEM and others) instead.

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.

labor saving devices

Light paraphrases of talk highlights:

* AWK: These kinds of problems should be 1 liners.

* AMPL: Linear programming is simple in concept but there's a huge gap of tedious, mechanical programming translating word problems into solvers.

* EQN: Lots of people with a wide range of training can unambiguously "speak" math and write down what is spoken as a formatted equation on a chalkboard but in practice [at that time] a professional typesetter must manually make this tedious translation.

* Bell Labs should buy a printer that costs the same as 3 or 5 houses: With new tools like EQN the Labs can make it up with a reduced budget for human expert typesetters.

* pic: Doing a sample-position-lines graph of this voltage-over-time equation would consume a lot of hours of the time of a professional artist but with pic its just an extra line of code.

* functional programming: Usage lost in the noise. I don't think that way. Sometimes useful stuff is thrown over the wall.

* scripting languages: undisciplined/sloppy/practical/easy

* net value savings (labor reduction) depends on uptake (factors like familiarity, easy of learning) and niche (e.g. eqn lives and dies with troff)

Not a bad summary...

Not a bad summary...