Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!uwm.edu!cs.utexas.edu!swrinde!ucsd!ucrmath!proton!meson!nusbaum From: nusbaum@meson.uucp (R. James Nusbaum) Newsgroups: comp.lang.c++ Subject: Re: C++ Not Ready for Commercial Use Message-ID: <534@proton.UUCP> Date: 17 Oct 89 18:18:18 GMT References: <24.UUL1.3#913@acw.UUCP> <6590301@hplsla.HP.COM> Sender: news@proton.UUCP Reply-To: nusbaum%proton.uucp@ucrmath.ucr.edu (R. James Nusbaum) Organization: Radiation Research Lab, LLUMC, Loma Linda, CA Lines: 50 In article <6590301@hplsla.HP.COM> jima@hplsla.HP.COM (Jim Adcock) writes: >Can anyone give any examples of someone who has been programming in C++ for >more than a couple of months who would willing program in any other language, >let alone any other dialect of C? Yes, me. Willing isn't quite the right word though. Different applications sometimes demand different solutions. I know you are all C++ fanatics here, but most of the features you like so well in C++ are done better in Eiffel, Smalltalk or Common Lisp with [Flavors, CommonLoops, CLOS, CommonObjects]. So if you really mean willing, well I'm not willing to program in anything but CommonLisp with Flavors, but that won't keep a paycheck coming in. :-) I am an object-oriented fanatic, but not necessarily a C++ fanatic. C++ is great, but it is a very immature product. I started programming with C++ right after it was first released to the academic community so I guess you could say I have used it a bit. When I started as leader of the project I'm working on now (~200,000 lines of code, 5-8 programmers, very high reliability and confidence level, multiple platforms) I looked closely at ALL available C++ products (Oregon, Oasys/Glockenspiel 1.2, AT&T 1.2, g++). None was suitable for commercial development. g++ had copyright/support problems. Oregon was way beta and full of bugs. Oasys/Glockenspiel failed my C++ validation suite. AT&T also had support problems. There were no symbolic debuggers (big problem) that worked on the actual C++ source. There was no decent documentation (everyone knows Bjarne's book is a little weak as a reference manual). All the compilers/translators were very young and I did not have the confidence in them that I do in a much more mature C product. Also the knowledge base for C++ and object-oriented stuff in general was and still is very weak. I feel I understand object-oriented programming very well, but I've been using it and doing research in it for five years. It can be very hard to teach and many people have a hard time understanding the concepts. So yes there were very valid reasons not to choose C++ for some projects. Many of those reasons are quickly going away which is great, but any software manager/group leader should carefully consider ALL the issues, not just the technical ones. By the way we used C with some home-built object oriented extensions. We will go to C++ when the computer vendors release supported binary versions with complete development environments (debuggers, profilers, etc.), which they are starting to do now. Jim Nusbaum Radiation Research Laboratory Loma Linda University Medical Center internet: nusbaum%proton.uucp@ucrmath.ucr.edu uucp: ...!ucrmath!proton!nusbaum