Path: utzoo!attcan!uunet!jarthur!nntp-server.caltech.edu!palmer From: palmer@nntp-server.caltech.edu (David Palmer) Newsgroups: comp.sys.mac.programmer Subject: Re: multi-platform revision control Keywords: rcs, unix, macintosh, ibm, projector, a/ux Message-ID: <1990Sep30.231919.7872@nntp-server.caltech.edu> Date: 30 Sep 90 23:19:19 GMT References: <1540@camex.COM> <1990Sep30.183252.9896@midway.uchicago.edu> Organization: California Institute of Technology, Pasadena Lines: 49 francis@arthur.uchicago.edu (Francis Stracke) writes: >In article <1540@camex.COM> kent@camex.com (Kent Borg) writes: >>Have you ever wondered why Microsoft applications look like they were >>not written by Macintosh users? They weren't. Apparently Word (and >>the other products out of Washington state at least) are written under >>MS-DOS. Microsoft has their own cross-development environment that >Not only are they cross-developing; they're not even compiling into >native code (as of Word 3, anyway). Word 3 has lots of PCOD resources, >with stuff in them that looks like strings that belongs in code, from >which I infer (am I right?) that they're using UCSD P-System. Back in 1984, when I was working on Word 1.0 for the Mac, we did our cross development on 68000-based Xenix boxes. (The cycle was debug-compile-download-run). The compiler (C of course) was developed in-house, as far as I know it was not based on the UCSD P-System, except in spirit. It ordinarily produced M-Code (same idea as P-Code, M stands for Microsoft) which was interpreted, but by enclosing a block, or section of code, in double-curly braces ('{{' and '}}') you could toggle between M-code and native-code compilation. The interpreter was written in hand-optimized assembler. This system had a lot of advantages, you could optimize with fine granularity either for speed or for space. Native-code segments were several times faster than M-Code, which was several times more compact. For Macs with 128k of RAM (the fat Mac was still in the rumor stage) and a single 400 kbyte drive (which also had to carry the System and Finder), we usually optimized for space, since if code is too large, any time savings will be eaten up by swapping it in and out of RAM. For things like displaying text, we compiled to Native code (I was responsible for optimizations to that section, and Word displayed text almost twice as fast as the MacWrite of the time.) For user interface stuff, M-Code was used. (If it takes 5 ms to set a check-box, instead of 1 ms, it really doesn't matter because the user won't see the box get set until the next screen retrace, 15 ms later.) So reports that Microsoft developed Mac software on IBM PCs are pure calumny. The user interface (at least for Word 1.0) was developed by expatriates from Xerox PARC, who had ealier designed the first WYSIWYG editor, BRAVO. -- David Palmer palmer@gap.cco.caltech.edu ...rutgers!cit-vax!gap.cco.caltech.edu!palmer I have the power to cloud men's minds -- or at least my own.