Timeline for How to overcome inaccuracy in Java
Current License: CC BY-SA 3.0
11 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| May 23, 2017 at 12:08 | history | edited | URL Rewriter Bot |
replaced http://stackoverflow.com/ with https://stackoverflow.com/
|
|
| Mar 5, 2014 at 5:39 | vote | accept | Hungry Blue Dev | ||
| Mar 5, 2014 at 5:32 | answer | added | keshlam | timeline score: 2 | |
| Mar 5, 2014 at 5:29 | comment | added | Hungry Blue Dev | @user2864740 thanks, i'll see to it. | |
| Mar 5, 2014 at 5:29 | comment | added | user2864740 | The only immediate solution I can think of is to special-case the input before calling tan. If it's "close enough to PI/2" for your liking, simply return Inf/-Inf (or whatever you are expecting) instead of calling tan. If you need to do "trigonometrical correct" math, there may be a library that can can deal with the exact mathematical concepts. (I think "symbolic math" may be a useful search term). | |
| Mar 5, 2014 at 5:24 | comment | added | Hungry Blue Dev | @user2864740 Thanks, that tells me why... But can u tell me how to fix it? | |
| Mar 5, 2014 at 5:22 | comment | added | user2864740 | See tan(PI/2) for why you don't get Infinity. However, I don't know of a way to get tan to "accept values close to PI/2" as PI/2, and all the environments seem to produce a value. | |
| Mar 5, 2014 at 5:18 | comment | added | user2864740 |
theta isn't exactly half-pi, which is why there is no Infinity in the output - NaN/InF is not "stored as" 1.633123935319537E16, which is still a finite number.
|
|
| Mar 5, 2014 at 5:18 | answer | added | yshavit | timeline score: 4 | |
| Mar 5, 2014 at 5:14 | answer | added | Florent Bayle | timeline score: 3 | |
| Mar 5, 2014 at 5:07 | history | asked | Hungry Blue Dev | CC BY-SA 3.0 |