Decoupling time from physics.

Mon, 02/02/2009 - 21:40 — chrisdew

It's possible to decouple time from physics, and it's especially easy in Haskell.

The (higher order) function below (lazily) accepts a list of events and yields a function which can give the historical or predicted displacement at any given time.

Note: The events do not have to be in chronological order - this provides a way to deal with lag in networked games.

The lazy nature of Haskell means that if the list of events is infinite (e.g. data from a network socket) then the generated function will give improved accuracy of position (for a given time) as more events are received.

newtype EvTime = T Double
newtype Impulse = I Double
newtype Displacement = D Double
  deriving (Eq,Show)
data Event = Ev EvTime Impulse
 
-- I can't yet figure out how to define or derive (+) for Displacement
plus (D x) (D x') = D (x + x')
 
-- 'calc' returns a function which calculates displacement at a given time,
-- provided a list of events are provided
calc :: [Event] -> (EvTime -> Displacement)
calc [] = (\ _ -> D 0.0) 
calc (Ev (T et) (I i):es) = f
  where
    f (T t) = if t > et
                then -- physics assumes unit mass
                  D ((t - et) * i) `plus` calc es (T t)
                else
                  calc es (T t)

Here's a simple demonstration...

*Main> :load Eventual.hs
[1 of 1] Compiling Main             ( Eventual.hs, interpreted )
Ok, modules loaded: Main.
*Main> let l = [Ev (T 1.0) (I 1.0), Ev (T 4.0) (I 0.5), Ev (T 2.0) (I (-2.0))]
*Main> let f = calc l
*Main> f (T 0.0)
D 0.0
*Main> f (T 0.5)
D 0.0
*Main> f (T 1.0)
D 0.0
*Main> f (T 1.5)
D 0.5
*Main> f (T 2.0)
D 1.0
*Main> f (T 2.5)
D 0.5
*Main> f (T 3.0)
D 0.0
*Main> f (T 3.5)
D (-0.5)
*Main> f (T 4.0)
D (-1.0)
*Main> f (T 4.5)
D (-1.25)
*Main> f (T 5.0)
D (-1.5)

chrisdew's blog
Login or register to post comments
1823 reads

Comments

Mon, 04/26/2010 - 20:06 — lemerian

affordable homes in

Sun, 02/21/2010 - 01:58 — edhardy11

Abercrombie

High Waist Straight-leg pants outfit stunted jacket, showing slender eye-catching position, pertinent mixture of units Cheap Abercrombie and Fitch Outerwears Cheap Abercrombie and Fitch Outerwears discount Abercrombie and Fitch Outerwears discount Abercrombie and Fitch Outerwears Abercrombie and Fitch Outerwears sale Abercrombie and Fitch Outerwears sale buy Abercrombie and Fitch Outerwears buy Abercrombie and Fitch Outerwears Cheap Abercrombie Fitch Outerwears Cheap Abercrombie Fitch Outerwears discount Abercrombie Fitch Outerwears discount Abercrombie Fitch Outerwears Abercrombie Fitch Outerwears sale Abercrombie Fitch Outerwears sale buy Abercrombie Fitch Outerwears buy Abercrombie Fitch Outerwears Cheap Abercrombie Outerwears Cheap Abercrombie Outerwears discount Abercrombie Outerwears discount Abercrombie Outerwears Abercrombie Outerwears sale Abercrombie Outerwears sale buy Abercrombie Outerwears at Break; DSquared jeans on The patches, holes, stains a Less.

Sun, 03/22/2009 - 15:24 — chrisg

define or derive (+) for Displacement

I discovered you can do this by adding:

{-# LANGUAGE GeneralizedNewtypeDeriving #-}

and then auto derive Num:

newtype Displacement = D Double
  deriving (Eq,Show,Num)

And then the line:

D ((t - et) * i) `plus` calc es (T t)

Becomes:

D ((t - et) * i) + calc es (T t)

Sat, 04/04/2009 - 15:51 — chrisdew

Thanks

Thanks, that's just what I was looking for.

final cog

Navigation

User login