This week's giveaway is in the Spring forum.
We're giving away four copies of REST with Spring (video course) and have Eugen Paraschiv on-line!
See this thread for details.
The moose likes Cattle Drive and the fly likes Assignment Java-7 (Sum) Big Moose Saloon
  Search | Java FAQ | Recent Topics | Flagged Topics | Hot Topics | Zero Replies
Register / Login

Win a copy of REST with Spring (video course) this week in the Spring forum!
JavaRanch » Java Forums » This Site » Cattle Drive
Bookmark "Assignment Java-7 (Sum)" Watch "Assignment Java-7 (Sum)" New topic

Assignment Java-7 (Sum)

James McLeod

Joined: Oct 16, 2008
Posts: 6
Can somebody point me to an explaination of why 0.1D is really 0.099999999999998590 and 1000 * 0.1 becomes 99.9999999999986? It's hard to accept a 'law' without a reason behind it.

James McLeod
fred rosenberger
lowercase baba

Joined: Oct 02, 2003
Posts: 11923

because you cannot represent all decimal values exactly to 100% precision. most numbers are correct to (something like) eight significant digits. You can't actually store 0.1, but you can store 0.099999999999998590, which is usually close enough.

The reason (I think) is that java stores number by adding up powers of 2. so to get something close to 0.1, you store 1/16+ 1/32 + 1/128.... etc. (note these are not nec. the ACTUAL sum, I'm just trying to illustrate it ).

you can get pretty close most of the time, but not exact.

There are only two hard things in computer science: cache invalidation, naming things, and off-by-one errors
Katrina Owen

Joined: Nov 03, 2006
Posts: 1367
It's actually not native to java, it's the very binary-ness (if I may invent words) of computers. The fact that they base everything off of 1s and 0s, means that they do their math in binary (base 2), which means that they have the same problems with dividing certain things by 10 as we do when we divide 10 by 3. In decimal based math that just doesn't compute very cleanly.
I agree. Here's the link:
subject: Assignment Java-7 (Sum)
It's not a secret anymore!