dan@hpnmdla.HP.COM (Dan Pleasant) (07/06/90)
Last week I requested information on the MPW C++ compiler. I got several responses, and there was some interest expressed in a summary, so here it is. The bottom line is that MPW C++ is a stable, useful product, but very slow. My thanks to all those who responded. (My e-mail thanks to a couple of you bounced.) Dan Pleasant ------------------ I do not own C++ (yet), as I have had too many experiences with buying beta releases from APDA, and then having to fork out extra bucks to get the final release. I have heard that it is pretty slow at compilation, and there are likely a few bugs (but I do not recall specifically seeing anything posted to this effect -- bugs, that is; it is slow). One thing I read back in March or so was an announcement from Apple that the final version (MPW C++ 1.0 Final) was to be available on June 15th. It wasn't, but should you go ahead and get C++, bear in mind that it will be upgraded fairly soon, and you may as well wait... -- --Frank Malczewski (fdm@wlv.imsd.contel.com) (malczews@nunki.usc.edu) ---------------------- I have used MPW C++ for some smaller projects and am currently using it for a commercial project. I have never encountered big problems, it seems quite useable. Only problem it's sloooooow ! Matthias ---------------------- In comp.sys.mac.programmer you write: >Why has MPW C++ remained in beta for so long? I dunno. Bugs I guess. I have not found too many though. Also, Apple may have decided to do 2.1 (doubt it though). >How buggy is it? Not too bad. I have found one that is REAL annoying, but work aroundable. very solid for a beta product. >Does anybody have a list of known bugs? I don't. If you get one, toss me a copy? >Most important: Is MPW C++ ready for "prime time"? Would you use it >to write a commercial product? I am. At least two, probably four. It's only really major lacking is a source level deebugger that understandes (and thinks in) C++. SADE ain't it. +C -- Cory Kempf I do speak for the company (sometimes). Three Letter Company 603 883 2474 email: cory@three.mv.com, harvard!zinn!three!cory -------------------------- >Why has MPW C++ remained in beta for so long? >How buggy is it? >Does anybody have a list of known bugs? >Most important: Is MPW C++ ready for "prime time"? Would you use it >to write a commercial product? I can't answer for Apple why C++ has been in Beta for so long. Most of the bugs are pretty esoteric. I use MPW C++ and MacApp for every application (that is not a driver or init, etc.) One reason for the delay is the perfection of the much awaited load/dump feature. It will really speed up the compile time. Another reason is that they want to have a very tight product without bugs. It'll be worth the wait and the upgrades are usually reasonable. I really love MPW C++ and I highly recommend it. ==================================================================== |Matt Mashyna | As you see, science once again robs us of | |mm5l@andrew.cmu.edu | our jobs. They put a micro-chip in my place.| |Carnegie Mellon | | ==================================================================== -------------------- Howdy, I'm going to pass on some second-hand information or "opinions" that may be useful. My impressions are that C++ is "beta" mostly because it is a complex system based on the latest AT&T C++ 2.0 specifications - apparently the specifications have outpaced their implementations. At the same time Apple wanted to be among the first companies to be AT&T-compatible (wow a definite rarity for Apple) so they released it in "beta" form within weeks after AT&T published the new specifications. Apparently, AT&T kept changing its unpublished specifications at the last minute (Apple had non-disclosure access) which lead to a lot of uncertainty about the implementation. If your problem is complex enough or if you want to work with new AT&T C++ 2.0 specifications, you may experience known or unknown bugs. For the most part, however, MPW C++ is stable and Apple engineers have been known to brag that the MPW implementation is more stable than AT&T's Unix implementation. Since I may be wrong, I will lend perspective on my opinion: A graduate student here at EPP at CMU wanted to know in March what was the fastest way to write a graphically-oriented Mac circuit layout program by June. It was almost HyperCard, then it was almost cT, an educational language developed here at CMU. Somehow I convinced him to use MacApp and he committed himself to his knowledge of C++. He didn't know a thing about MacApp (an object class library written in MPW Object Pascal with a set of MPW C++ headers to access the class library). He stunned himself by finishing the project a mere 2 weeks behind schedule, or in about 3-1/2 months. Most of his time was spent learning MacApp - his biggest "C++" problem was with bugs in the C++ headers for MacApp, not in C++ itself. His application is about 600K in size and is pretty sophisticated, but it isn't too complex and it doesn't fool around with any new C++ 2.0 features (e.g. multiple inheritance) - he stuck with the features that enabled him to use MacApp (HandleObjects), and with those features he had learned in classes teaching AT&T's previous specification of C++ under Unix. I am planning to use C++ with MacApp on my next programming project. - Brian