It's probably not very helpful, but at least I get the idea of a
concurrent redisplay planted into brains, where it can do it's evil work
:-).

>
> (However, before anyone gets their hopes and/or fears up, my code
> depends on disabling most of the regexp code, and the additional number
> of garbage-collected objects is so great that I concluded I'd wait for
> MPS to land before resuming work on it. One of the few distinct
> advantages of the current gap buffer approach is that it doesn't affect
> GC...)
>
> I know virtually nothing about redisplay.
>
> Pip

What I've written is pretty high-level, nothing to worry about.