Hacker News Reader

Comment by Tade0 | Hacker News Reader

This appears to be a case of something that I've known as "the BBBB problem" - should we optimise algorithms that will rarely be used in a way that makes computation time explode?

I mean, who removes history entries by selecting thousands of them manually? Okay, this guy, but who else?

I don't remember who named it like that, but the associated anecdote was about typing "BBBB" somewhere in Excel(I think it was the function search input), and how unlikely it was to ever happen.

Personally I've pasted the wrong thing into search bars more than once and they exhibit a range of behaviours.

View on HN

I use the example as the opposite of a canary in a coalmine, a good omen, some way to divine good attitude and thus, good things to come. If I wanted to point to areas where Mozilla have clearly wasted effort that would be easy enough, the comments and submissions to HN are littered with them. The point is that there are basic functions in the browser that are ignored while these other - shall we call them debacles, because that's what they tend to end up as - are released and pilloried or ignored.

Hence, to see something like this be fixed would be a sign that developer time is going into the core product. It could be anything else more fundamental than whatever I've seen announced by Mozilla lately.

> I mean, who removes history entries by selecting thousands of them manually? Okay, this guy, but who else?

I’ve often done it to clear out useless repetitive history entries that are just clogging up results when typing in the address/search bar. Ideally it would simply not make entries for some kinds of redirects and I could define rules of URLs to exclude, but in the absence of that I might just go through once or twice a year and delete a few of the biggest repetitive patterns. And how slow it’s been for the last decade has been rather frustrating at times. I just tried deleting a couple of thousand (I use Nightly, where a fix apparently landed last week) and it took half a second, which is much more relaxing than the ten or twenty seconds of full-blocking CPU thrashing that it used to be.

"who removes history entries by selecting thousands of them manually? Okay, this guy, but who else?" Me too. When I read the first guy's comment I was like "Goddam it's not just me"

I do understand what you mean, and in general I agree, but in this specific example making it faster doesn't require "optimizing" anything! If I am understanding correctly the "naive" algorithm here is the fast one, so it takes more effort to make it slow than it does to make it fast.

Reply on news.ycombinator.com