Path: utzoo!attcan!uunet!ogicse!decwrl!megatest!djones
From: djones@megatest.UUCP (Dave Jones)
Newsgroups: comp.lang.misc
Subject: Re: Intermediate Code
Message-ID: <12602@goofy.megatest.UUCP>
Date: 14 Apr 90 01:14:07 GMT
References: <1102@calmasd.Prime.COM>
Distribution: comp.lang.misc
Organization: Megatest Corporation, San Jose, Ca
Lines: 20

From article <1102@calmasd.Prime.COM>, by cpp@calmasd.Prime.COM (Chuck Peterson):
> 
>         I'm interested in designing an interpreter for a procedural
>         language with object oriented features.  (No flames, please.
>         I'm just describing what I'm doing.)
>
>         What is going on in the world of intermediate code these days?

Everybody else seems to be using C as their intermediate code. Not a bad
idea, that. If you want to execute interactively entered code, as in
Sabre C, for example, you'll have a bit further to go, in that you will
need a fast compiler/dynamic-linker.

Directly interpretable code (p-code) usually suffers from machine-
dependence: byte-order, floating-point format, alignment conventions and
restrictions. So I would say, stay away from it if you can. I speak from
experience on this! I thought I was safe because I was more or less
committed to one vendor for a target-machine. But that vendor then came
out with a new architecture, bythefully imposing absurd alignment
restrictions on data, breaking my interpreter horribly. Aaaaaaarghhh!