On May 4, Richard Erlacher wrote: I have to agree with Richard here. Back in 1986/87, I worked on the Navier-Stokes Supercomputer project at Princeton University...the prototype processors for that machine were wire-wrapped (most of the boards by me! 8-)) and signals ran between 20-30MHz. We had no problems at all. We used twisted-pair wiring for some of the clock lines as I recall, but the vast majority of the wiring was simple point-to-point wirewrap with no special routing considerations of any kind. Don't knock wire-wrapping...it's fast, easy, works very well, and properly done and cared for, it will last for years. And if you're a weirdo like me (and I know some of you are) it can even be FUN! 8-) As an interesting side note, that supercomputer (the project is still underway from what I'm told, undergoing various refinements to keep pace with advancing technology) is built on Unibus-form-factor boards, using PDP11/24-style expansion chassis. It was a cheap and easy way to get nice, compact chassis with high-quality power supplies and very well-known (and large) board geometries. -Dave McGuire