This commit is contained in:
Cat Flynn 2024-04-27 10:49:25 +01:00
parent efbce4ec40
commit 56fa0f9123
1 changed files with 1 additions and 1 deletions

View File

@ -171,7 +171,7 @@ You think as far as double-precision floating points can take you - which is, wa
The distance doesn't really matter anymore, you realise, since the precision is the same for an ellipse a metre across as it is for an astronomical unit. The distance doesn't really matter anymore, you realise, since the precision is the same for an ellipse a metre across as it is for an astronomical unit.
The limit now is double-precision time. The limit now is double-precision time.
Idly bouncing the ball in one hand, you work out what that means. Idly bouncing the ball in one hand, you work out what that means.
To be accurate to a single frame, you want be able to represent units of time as short as a frame, or 0.01 seconds. To be accurate to a single frame, you want to be able to represent units of time as short as a frame, or 0.01 seconds.
There are about ten million seconds in a year, or billion frames: looking good so far, that's only 10 significant figures, there are a few more to go yet. There are about ten million seconds in a year, or billion frames: looking good so far, that's only 10 significant figures, there are a few more to go yet.
In a thousand years, the smallest step we can represent is still less than a microsecond, so you keep going. In a thousand years, the smallest step we can represent is still less than a microsecond, so you keep going.
After a million years, the minimum increment finally creeps up to half a millisecond, which sounds about right. After a million years, the minimum increment finally creeps up to half a millisecond, which sounds about right.