Re: [Gzz] OrthoCoordsys???, reverting

[Tuomas Lukka]

> But I'm still not entirely clear about why you think it would be more 
> efficient that way-- can you explain?

Quite simple: the asymptotic complexity goes from O(N log N) to O(N).
Because in the current, recursive model you go through all parents,
i.e. the whole tree, for each vob you're interpolating. Here the effort
is constant.

> >>And b), I don't buy into it being more efficient: it 
> >>creates one object per coordsys, which is precisely what I'm trying to 
> >>avoid in AWT. This is a place where I would like to see some serious 
> >>benchmarks first before changing anything ;-)
> >
> >Yes, actually I was thinking of storing ints instead of keys for 
> >efficiency.
> 
> Ok. -- Of course, then you'd have to use a custom hashtable, again.

Yes, but OTOH the interpolation wouldn't need to use hashCode() but just
integer comparison.

        Tuomas