CompuServe Thread

#PD Compiler?

3 messages in this thread
#44249From: Richard BielakApr 17, 1989 12:19 AM
Martin, I've been fooling around with C++. You are right, it requires somewhat different approach to thinking about programs. With C++ its not that bad, since C++ is close enough to conventional languages. Actually, C++ has some things in it I wish Modula-2 had. They don't seem that difficult to implement. I think that OOLs will be more important in the future. In fact Wirth's new language (it's called OBERON) incorporates some object-oriented ideas. This shift to OOLs, is kind of like going to structured-programming from flow charts. I mean, people who program in PASCAL/M2/C don't draw flow charts anymore do they?
#44260From: M2S/Phil CampApr 17, 1989 8:10 AM
Oberon has very little more than M2, and I like that. The type extension proposed by Wirth should not be *that* difficult to implement in M2. I am a little worried about the garbage collection though (I'll have to read up on techniques to allow the compiler to do most of that at compile-time). It seems there's a portable Oberon compiler scheduled to be out in a few months. If so, look for an Amiga version in the future. I'm also interested in the extensions proposed to M2 by the ISO commitee working on it (array and records constants for example). I've also been looking at OO M2 (Objective M2 ?)
#44271From: Richard BielakApr 17, 1989 12:41 PM
From what I read about OBERON, I think that type extension can be really all handled at compile time, I don't see a need for garbage collection. I know that Wirth has someone working on a portable OBERON compiler, but none is available yet. There have been a couple extension to M2. There is Modula2+ at DEC somewhere, and Modula-3, and someone in Europe has a product that combines Modula-2 and Prolog! The things are I'd like to see in M2 is termination sections for modules (just like in OBERON), mutli-dimensional open arrays. Structured constants can be useful, but they get in a way of re-entrancy.