>I don't agree. If the source code has been compiled into some tokenized sequence of data that is used to execute by an engine, it is not interpreted, it is compiled into some p-code. Interpreted is generally defined as being source code that is not 'modified' in any way before execution (just like VBScript and JavaScript). So in no way do I think VFP, Java, .NET, etc. is interpreted and I don't think most people would think so either. But all this is just terminology, verbage, nothing really important.

We're trying to pin down definitions here. All my sources (including the Web), have no problem with source code that is "modified" as it applies to interpreters. I'm simply curious where this "generally defined" and "what most people think" is coming from.

I suspect that if I were to ask a CS101 student what a "compiler" vs an "interpreter" was, I would get a simple answer/definition.

Frankly, I see nothing magic about "p-code". I don't think anyone would have a problem with doing "something" with virgin "text" before passing it to a run-time interpreter in order to eliminate the most obvious "coding errors". And since "band-width" has been a central issue with the Web, it only seems natural to "compress" this text in some sort of way (ie. bytecode) before passing it around. None of these actions automatically disqualifies "interpretation". (Considering the bloat that XML contributes, I see XML "bytecode" in our future ... and a corresponding "interpreter").

Why this is an issue at all (IMO) is because of the hay that has been made re: "Patterns".

An "interpreter" is simply another pattern (IMO); discuss "interpreters", and folks (used to) know what you were talking about.

Now, simply for the sake of "marketing" (IMO), we don't have "interpreters" (or compilers) anymore ... we have "something" that no one seems able to pin down and contributes nothing to the discussion at hand. Expect that it's "better" (or so it's claimed).