Re:Thinking too much can be bad for you (2012)
by magnusmundus on Tuesday February 09, 2021 @08:51PM
(#26084947)
> the ability to scope things down to what exactly what matters
In the last few years, I started shifting my focus from what variety of cases to handle with an elaborate solution, to what possible cases I might be blocking with a simple solution. Gratuitous YAGNI was an important intermediate step in that process. It might not be very FAANG-friendly (or it might not scale to four-digits-engineers regardless of brand), but both in terms of real life performance and peer feedback, it's yielded very good results.
While the first is prone to becoming a bottomless pit of SWE-self-pleasuring; the latter is actually very useful in coming up with something that's cheap, easy and maintainable, but still flexible enough -- which doesn't necessarily mean extensible! Ease-of-refactoring is another flavor of flexibility, and that's where I found this kind of thinking beating YAGNI alone.
One actual drawback I see is that it makes DRY difficult to achieve, but the more people I work with, the more I'm becoming disillusioned with that anyway. It's such an overly adaptable dogma to generate work that is often pointless, and becomes harmful so easily. I think we found so much comfort in "don't type the same thing twice" (though I still try to keep it to 3) that we don't stop to think about "don't maintain the same information twice" anymore. But this is 3AM rambling, so I'll stop here.
In the last few years, I started shifting my focus from what variety of cases to handle with an elaborate solution, to what possible cases I might be blocking with a simple solution. Gratuitous YAGNI was an important intermediate step in that process. It might not be very FAANG-friendly (or it might not scale to four-digits-engineers regardless of brand), but both in terms of real life performance and peer feedback, it's yielded very good results.
While the first is prone to becoming a bottomless pit of SWE-self-pleasuring; the latter is actually very useful in coming up with something that's cheap, easy and maintainable, but still flexible enough -- which doesn't necessarily mean extensible! Ease-of-refactoring is another flavor of flexibility, and that's where I found this kind of thinking beating YAGNI alone.
One actual drawback I see is that it makes DRY difficult to achieve, but the more people I work with, the more I'm becoming disillusioned with that anyway. It's such an overly adaptable dogma to generate work that is often pointless, and becomes harmful so easily. I think we found so much comfort in "don't type the same thing twice" (though I still try to keep it to 3) that we don't stop to think about "don't maintain the same information twice" anymore. But this is 3AM rambling, so I'll stop here.