[tpm] Complexity and language

arocker at Vex.Net arocker at Vex.Net
Sat Jun 30 11:09:55 PDT 2012


Thanks to the audience on Thursday for helping me with the subject.

If you weren't there, we were trying to come up with an orderly taxonomy
of complexity in software development, to match the sophistication of
language features appropriately to the problem. Solutions should not be
more complex than the problems they attempt to address.

Assembly languages are quite easy to learn, because there's not much to
them. On the other hand, doing anything useful with them is hard work.
(Anything computable is theoretically computable with a Turing machine,
but generally you wouldn't want to do it.) Richer languages are harder to
learn, because there's more to them. Perl 6 is so insanely rich that most
people won't ever need to learn its full depth.

A tentative ranking of system complexity, (from simplest to most), was:

Toy problems
Simple computer management (job control)
Basic text  generation to simple Web output
Data processing (listing and summing input)
Text processing
Systems programming (compilers & operating systems

If anybody has reasoned views on this, I'd be happy to hear them.



More information about the toronto-pm mailing list