Quote:
Originally Posted by tompe
Except that the page forward is so slow that this method is not working. I actually notice page number when I read and remember some of them.
|
Ok, add one stap more
Quote:
I was somewere in the middle, total size is 700 pages, so it must be around 350. Hmmmm...I already read this part, so it must be 50 higher...Aha...that is new. So it must be 25 lower. Aha...read that part. I'n within 25 pages now, on the avarage 12. Page forward from here
|
Do
not page forward from her, but increase with 12 and you are on an average of 6 pages from the correct one.
Page forward slow? It takes 0.75 second to switch a page. Add 0.2 seconds for the reaction time of the average human beeing. So you can turn a page about every second. Not much slower tha a p-book. But it means that even if you have to page forward 12 pages, you will find your page in about 15 seconds.
And even if you saw a page number a couple of minutes ago, you stil have the same procedure. Only you won't start at my 'educated' guess of 350 but at the 362 you have seen. That one is almost certainly to low, so you'll have to increase with 50 and follow the same procedure.
The trick with this binary search is that it really doesn't matter were you start. If, in my example, I didn't had a clue were to start, and made the stupid choice to begin at 690, the next step would have been 345, followed by 520 followed bij 430, 380 and so on. The numbers don't have to be exact, the only thing you have to do is split the difference until you are close enough. The difference between the last known number methode and the worst case is probably two guesses. The difference between the last known number and an educated guess is
maximum one itteration and probably no difference at all.