[BigJono]

I think maybe it's possible that given a very smart programmer with 1-5 years experience, the value of not having had time to learn the wrong way of doing things far outweighs the value of another 10 or 20 years experience. A lot of our 'best practices' seem curiously tailored toward keeping mediocre developers in check and producing consistent rather than extraordinary results.

Maybe our industry just grinds the greatness out of you by 25 or 30 unless you actively fight it.

[taneq]

I think if someone's better with 5 years' experience than 20 years' experience then they have a very low skill ceiling for the task at hand.

Part of that 20 years' experience, of course, is learning to ignore 'best practices' that actually aren't 'best' at all (or more likely, are 'best' for some situation you're not actually in).

[Swizec]

> Maybe our industry just grinds the greatness out of you by 25 or 30 unless you actively fight it.

Maybe people just get tired of churning out CRUD apps after 10 years and want to do something else. A lot of stuff loses its charm after you see behind the curtain.

[Fr0styMatt88]

There’s that saying, something about 10 years of experience vs 10 years of doing the same thing (‘1 year of experience 10 times’ or something like that).

I also remember someone bringing up a good point (it might have been here or something I watched, I can’t remember) that software engineers don’t actually get that many ‘attempts’ at doing a full software system from scratch. How often does the average software engineer get to design a system from requirements, build a system from scratch, take it through production and then into maintenance in that length of time?

[AQuantized]

What do you think is an example of such a "best practice"?

[BigJono]

I mean the classic Java > Lisp decision that Paul Graham is always talking about surely fits the bill.