3 ms·
In addition to pjc50's answer: 5) Right now it's probably harder (in most cases) to write a correct formal spec than actually implementing that spec. Otherwise
by l0b0 6y ago
In addition to pjc50's answer:
5) Right now it's probably harder (in most cases) to write a correct formal spec than actually implementing that spec. Otherwise formal specs would be much more popular in general.
6) The proportion of software which is actually algorithmically interesting these days is tiny. I expect something like 99%-99.999% of lines of code are basically stuff that's been done a thousand times before.
- cloogshicer 6y agoThank you, your point 5) is extremely interesting. Something clicked when I read that. I guess what that means, right now, when we're writing Software, we're really only ever solving a small part of the whole problem. A lot of the problem space isn't even considered. (I'm saying this completely neutral btw, I don't think it's either good or bad. It's a trade-off and probably what allows us to have all the software we have now.)