3 ms·
Absolutely. But Big O still comes up in practical stuff depending on what you're doing. You can usually get by without it, but sometimes when you need to proc
by jedberg 7y ago
Absolutely. But Big O still comes up in practical stuff depending on what you're doing. You can usually get by without it, but sometimes when you need to process a few 100 terabytes of log files, having a basic grasp on Big O comes in handy for reducing your processing time from days to hours by picking the right algorithm.
- RussianCow 7y agoI think the point the parent was trying to make is that the concept is way more important than the specific notation. Someone that doesn't know Big O could still tell you why one algorithm would be better, it just might be slightly less succinct: "The execution time of algorithm A scales up with the number of items, while algoritm B's is constant." vs "The execution time of algorithm A is O(n), while algorithm B's is O(1)."
- taqwerty 7y agoIn practice I've always found avoiding premature optimization and then profiling better than worrying about Big O.
- learc83 7y agoYou save an awful lot of time when you know have an idea of what's going to be slow up front. I've spent so much time trying to chase down bugs caused by junior programmers because something that worked fine with 100 rows in db, slows to a crawl when we move a feature out of beta and we suddenly have 1000 rows. If they had some idea that the query they were writing was O(n^3) they could have done a quick estimate and saved us all the trouble.