>>No it doesn't. You posted no evidence to this, just speculation. It only proves that Refox is aware of what the tokens mean. Nothing else. In fact, I took a version of Refox that was produced before VFP 6.0. I took a simple fxp file produced by 6.0 containing nothing that wasn't in the language prior to 2.6 and it worked fine.
>
>Well then, the "tokens" ARE the "program". You've just confirmed that. No table offsets, library entry points, blah, blah.

In a sense. What I said initially was that they represent an offset into a table containing the address of the entry point in the runtime where the machine is. You may view this as "blah, blah". I don't. It is distinctly different than how a true interpreter works.