jdb@mordor.UUCP (John Bruner) (10/30/85)
[I'm cross-posting this to "net.micro.mac" and directing followups to that newsgroup only.] Before I start, I'd like to say that if you are used to developing C programs on UNIX, using the Mac will be a big step down. It is *slow*. It is *painful* (beware of leaving your source discs in the Mac when running your program -- in several cases a program crash due to an uninitialized pointer has caused my Mac to freak out and trash my discs). The Toolbox doesn't handle errors very well (e.g. DrawPicture on a still-open picture will produce a disc- destroying crash). [Much of this is to be expected, given the difference in hardware between (say) a VAX with Fujitsu Eagles and a non-MMU micro with floppy discs. I hope affordable micros with memory management start showing up in the marketplace soon.] The November 1985 issue of BYTE does have a lengthy comparison of Macintosh C compilers. The author favors the Manx Aztec C development system. The Manx product is a good one; however, I believe that the review was somewhat unfair to the Megamax compiler. (I use the Megamax compiler, but I have no other connection with Megamax, Inc.) One point which argues against the Megamax compiler (and which was understated in the review) is that it defines "short" integers to be 8 bits wide. This is contrary to the conventional practice in other compilers (and will be in violation of the ANSI standard when that standard becomes available). This is a serious "misfeature" and I wish Megamax could be persuaded to fix it. A big factor in favor of the Megamax compiler is the lack of copy protection. This has made their development system very flexible. A second major factor is price. The review compares the most expensive development system available for each of the five compilers it covered. The difference between the (undiscounted) prices of the Megamax compiler and Aztec developer's system is $200. It was unfair to cite the RAMdisc compile time of the Aztec C compiler without providing similar figures for the other compilers. Perhaps the other compilers do not come with RAM discs, but for $200 (the difference in price between the Aztec developer's sytem and the Megamax system) one can purchase all of the available RAM discs for the Macintosh (some are even public domain). The lack of copy-protection on the Megamax compiler probably enables it to use the RAMdisc more effectively than the Aztec compiler can. One negative assertion about the Megamax system -- the lack of sources for the library routines -- is untrue. The sources can be purchased for (I believe) $75. ($375 is still cheaper than $500.) I have found Megamax to be an good company to deal with. With the exception of short integers, I'm [obviously, I guess] a satisfied user of their product. -- John Bruner (S-1 Project, Lawrence Livermore National Laboratory) MILNET: jdb@mordor [jdb@s1-c.ARPA] (415) 422-0758 UUCP: ...!ucbvax!dual!mordor!jdb ...!seismo!mordor!jdb
dad@mit-vax.UUCP (David Duff) (11/05/85)
In article <4141@mordor.UUCP> jdb@mordor.UUCP (John Bruner) writes: ... >Before I start, I'd like to say that if you are used to developing >C programs on UNIX, using the Mac will be a big step down. It is >*slow*. It is *painful* (beware of leaving your source discs ... This is a meaningless statement. Perhaps what you mean is "if you are used to working alone on a large VAX, the Macintosh may be slower." I think Aztec C is faster than several UNIX systems I have worked on.