Re: Distance units

On 18/03/2005, at 10:00 AM, Andrew McDonald wrote:

> But it doesn't require metres per second. It's an assumption you're 
> making that that value is in
> metres per second - but it's actually not. It *happens* to be the 
> value needed to simulate standard
> Earth atmospheric properties, but it could equally well represent a 
> different atmosphere if regarded
> in feet per hour or whatever unit. It's ugly to have it in the spec as 
> a magic number, that's why I
> wanted it separated out as a user variable instead of having a scaling 
> factor thing. Of course it's
> ugly to have a default that works so well for metres per second on 
> Earth - that's why I suggested
> having 0 as a special value that signifies the Doppler effect should 
> be disabled. The default value
> for the speed of sound could be this, implying Doppler is disabled 
> unless the API user explicitly
> sets the speed of sound. This is exactly the same as how the 
> velocities default to 0 and are then
> set by the user to whatever value they want - and because the user is 
> setting both the object
> velocities and the speed of sound they'll be in the same units already.
I think having 0 represent infinity is a bit unnatural, especially 
since a value of 0 for speed of sound actually still works when you 
plug it into the Doppler equation (representing the case when the sound 
waves are standing still in the air and you are "hearing" them by 
moving through the air), giving a behaviour very different from 
infinitely fast sound.
I think more natural and logical (in my opinion) would be to either...
1.  Have a speed of sound parameter defaulted to 1.0 and a separate 
on/off switch to turn doppler on and off, or
2.  Use a vector scale parameter defaulted to 0.

Either of these would be nice (although I do have a slight preference 
for number 2 since it feels very natural to me and it has the ability 
to turn Doppler on and off in a nice and more mathematically sound 
way).