>One might argue that a VFP p-code "interpreter" also converts p-code to a sequence of "machine code" routines that get run at runtime.

I'll only argue that p-code causes "calls" to machine code routines.

>As far as I am concerned, it doesn't matter whether the thing is "compiled" or "interpreted" as long as the system yields acceptable performance/resource requirement. VFP does that.

At some point, yes. But if one is going to engage in any type of productive dialog with people outside their immediate circle without having to continually hammer out definitions beforehand, then one needs a commonly understood and accepted set of "patterns". I believe we "had" those.