Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!hellgate.utah.edu!csn!alumni.colorado.edu!fozzard From: fozzard@alumni.colorado.edu (Richard Fozzard) Newsgroups: comp.sys.mac.misc Subject: Re: Mac LC or NEXT Message-ID: <1991Feb4.201342.28566@csn.org> Date: 4 Feb 91 20:13:42 GMT References: <13255@medusa.cs.purdue.edu> Sender: news@csn.org (news) Organization: University of Colorado, Boulder Lines: 47 Nntp-Posting-Host: alumni.colorado.edu In article melling@cs.psu.edu (Michael D Mellinger) writes: >bag low-level programming and device-dependant imaging models. >Computer performance is such that we can afford to waste a few mips on >Display Postscript, OO programming, and other high level languages. >You might say "C forever", but will your competition? Maybe so, but how many people will actually use your device-independent approaches? There is a limit to the performance hit that people are willing to take for device independence. I have talked with several NeXT developers, and they all have had to put special code in their programs that bypass DP and do things with straight bitmaps (or other technique) to get acceptable screen performance. Such code would presumably not port to NeWS, of course. And from the very beginning PC and Mac programmers (especially for games where performance is key) have bypassed the "proper" way of displaying to the screen. The hard fact is that programmers for a given platform will always "cheat" to get better performance, no matter how fast the platform. BTW, I find it ironic that a NeXTNuT would be defending programming for device independence, when a NeXTStep/Objective C program is about as device-dependent as you can get. The NeXT company and most of its user community has rejected X out-of-hand - NeXT is probably the only workstation manufacturer to not provide a company-supported X. I like the NeXT very much, and have used it to develop some small programs. It is a wonderful development environment, the best on the market, I'd say. Yet the code runs ONLY on a NeXT (and the very few IBM 6000s that NeXTStep has shipped for). That is exactly why I have failed to convince my superiors to buy any. The programs I build using Guide/OpenWindows on a Sun run on anything (even PCs and Macs) that can run an X-server. Programs I build in HyperCard run on Macs and PCs, and when Spinnaker finishes their Plus for X later this year, everything else. By refusing to provide a hardware-independent NeXTStep X toolkit, NeXT has negated the most significant advantage of its machine! No doubt, NeXTStep on top of X would run rather slowly, but at least it would RUN! And it seems to me that this would only help sales of NeXT hardware, as the fastest way to both write and run such programs. -- ======================================================================== Richard Fozzard "Serendipity empowers" Univ of Colorado/CIRES/NOAA R/E/FS 325 Broadway, Boulder, CO 80303 fozzard@boulder.colorado.edu (303)497-6011 or 444-3168