🚀 go-pugleaf

In the referenced article, Howard Chu cited as evidence that it would be harder
than I suggested to make a 16 MHz ST the experience of Megabyte Computers,
whose 16MHz accelerator board never reached production.

I'm suggesting something different.  The story I heard about why the speedup
flopped was that the MSI-type logic to connect the cpu to the slower video 
circuitry was designed around a chip that was later found to be inadequate.
You make the change that video gets every FOURTH memory cycle instead of every
second memory cycle (the cpu still couldn't use them, since it's a design 
principle of the ST to use memory faster than the cpu, but maybe DMA perform-
ance could benefit from the newly available cycles).  You also make those 
custom chips in the box faster.  It's not an add-on, it's a new machine, but
the engineering changes would seem to be of a kind that could be done properly
by a small number of people in a short time (=affordable).

If anyone from Megabyte is reading this, I hope you're back at it after the
disappointment.