[plt-scheme] PLT4 expectations

At Tue, 11 Mar 2008 13:59:33 -0400, Doug Orleans wrote:
> Matthew Flatt writes:
>  > At Tue, 11 Mar 2008 01:01:08 -0400, Doug Orleans wrote:
>  > > Matthew Flatt writes:
>  > >  > At Mon, 10 Mar 2008 14:34:49 +0000, "Paulo J. Matos" wrote:
>  > >  > > The idea of PLT4 creating by default immutable pairs and the so, 
> means
>  > >  > > that it also opens the possibility for far more and better
>  > >  > > optimizations in the compiler, right?
>  > >  > 
>  > >  > I doubt that there are many new optimization opportunities that matter
>  > >  > in practice.
>  > > 
>  > > Well, you could make "list?" be O(1) instead of O(n), at a cost of one
>  > > bit per cons cell.
>  > 
>  > Yes, `list?' is currently amortized constant time.
> 
> That's great!  I instinctively avoid "list?" (and <list> specializers
> in Swindle), but now I will stop being afraid.
> 
> How about "length" and "last"?  Or would those be too expensive in space?

Yes, I think that would be too expensive.

Matthew