Ricardo Signes wrote: > "I imagine that we can invert the > parser so that..." is a very useful vision to present. Except it turned out to be a flat-out failure. :-)