3 ms·
> you should know beforehand. Don't you know what your software does? If you think you know beforehand without measuring it I suggest you do measure it now and
by alpaca128 3y ago
> you should know beforehand. Don't you know what your software does?
If you think you know beforehand without measuring it I suggest you do measure it now and then. Because if there's anything I've learned about performance it's how terrible one's intuition can be about performance.
> The performance when n is small usually doesn't matter.
Sure, when the problem is simple enough that there's only N and not also M and K.
- marcosdumay 3y agoOh, sure, I'm not saying you shouldn't measure. What I'm saying is that you should know beforehand (with much less precision), and on the are many decisions you have to make you should apply that knowledge, and not go with YAGNI every time without questioning it. YAGNI is not at all a good principle. It only works as a slight bias. If you use it as any stronger kind of rule, it's disastrous.