RE: Adding a New "primitive" Method.

[Matthew Flatt <mflatt@cs.utah.edu>]

Quoting Brent Fulgham:
> Is the tight coupling of MrEd to MzScheme truly necessary?

No, not anymore, because we eventually figured out the right way to
implement threads, custodians, and thread-GCing in MzScheme.

While we were sill figuring out those elements, the tight coupling was
necessary to make MrEd work (using implementation strategies that
weren't quite right).

> Are there great performance gains or better memory use the
> way you implemented MrEd?

No. It was merely the development path that got us here.

I do expect MrEd to become a MzScheme extension (instead of a MzScheme
embedder) in the near future. But it's about #3 on the list of planned
improvements. More on that soon...

Matthew