[Haskell-cafe] Re: Optimization problem

[Ross Paterson]

On Sun, Sep 17, 2006 at 01:08:04PM +0200, apfelmus at quantentunnel.de wrote:
> Ross Paterson wrote:
> > It's interesting that these composed transformations don't seem to cost
> > too much.
> 
> That the composed transformations are indeed cheap is not necessarily
> disturbing.

I meant the composition of the transformations of the tree (update or
reverse insert) that Bertram does for each list element.  They're cheap
because they're bounded by the depth of the tree, and even cheaper if
you're probing some other part of the tree.