Re: The melting ice technology (2): levels

bertrand@eiffel.com (Bertrand Meyer.) writes:
|> It would thus be incorrect to say that Personal Eiffel for Windows is an
|> interpreter. It is definitely a compiler - in fact, exactly the same
|> compiler as Professional Eiffel (for Unix, for Windows or for other
|> supported platforms).


Come on. The preceeding description of Professionial Eiffel's large grain
recompilation and small-grain re-interpretation is very interesting, and
it sounds like a hybrid of recompilation and interpretation techniques.
But to reach the above conclusion, you have stretched definitions of
`compile' and `interpret' well beyond common usage to the point of
sophistry. You have defined them such that all systems are both compilers
and interpreters, and thus made them meaningless.


But I think to the vast majority of people, `compiled' has a quite useful,
if not rigorous, definition: is my code translated directly to machine
code or is translated and interpreted by some virtual machine? Because if
it purely interpreted, it will typically be one or more orders of
magnitude slower than the equivalent code directly translated to machine
code.


>From your description, Personal Eiffel is an interpreter with an extensive
compiled library: the code actually written by the user runs at
interpreter speeds. So it is an interpreter. Period.


Now if typical user applications spend a vast majority of their time
operations implemented by the compiled library, then you could argue that
the fact that Personal Eiffel is an interpreter has no effect on
performance. But Personal Eiffel would still be an interpreter.


        Peter Moore
        peter@objy.com
--

Return to the comp.compilers page.
Search the comp.compilers archives again.