We had a bad week. Someone (and I use the term loosely) broke into our old car and stole a small piece of the ignition, rendering poor Watson the Datsun undrivable. The perpetrator didn’t take anything else, but did break some plastic and mess up the passenger door outside lock. Needless to say, we weren’t pleased, and I spent most of two days dealing with it, having Watson towed to a shop, finding and installing a new ignition, and putting him back together. Then on Saturday I made the mistake of playing basketball with a group of people who dwarfed my 6’0" height and 150 lb. weight. It was, and still is, a painful experience. We also forgot to include the phone number for Amaze in last week’s issue, so it’s at the end of this article. Not all is gloomy though, and we did hear some good information at the latest dBUG meeting, at which Aldus honcho Paul Brainerd gave yet another "future of computing" talk but with a semi-working version of something he loosely called Multimedia PageMaker. This issue’s date, Monday, November 18th, is officially my 24th birthday, which will be fun, although not as much fun as if we were back in Ithaca, an option we seriously considered as Watson sat glumly sans ignition. As some unemployed friends say, the worst that can happen is that we’ll die and crows will peck out our eyeballs. 🙂 Anyway, on to the issue.
One quick piece of administrivia: the caretakers of the FTP site at sumex-aim.stanford.edu have decided to reorganize the site slightly. TidBITS issues will be located in the /info-mac/digest/tb/. directory from now on. So if you get your fix from sumex, make a note of the new spot.
There’s been some confusion over results from Speedometer 3.0 (an excellent shareware utility for rating the speeds of Macs) especially in the case of the Quadras. Apparently, if you turn off the caches on a Quadra, it runs at about the speed of an SE/30 or a IIci running a color monitor in 256 colors from internal video. This has bothered a lot of people, but Mike O’Dell offers this technical explanation. "A bit of Computer Architecture 101: Without the caches running, the 68040 is limited to the speed of memory, so expecting it to go noticeably faster than a cache-less 68020 on the same memory system is pretty unreasonable. Yes, the 68040 shortened the cycle counts for some instructions, but all that means is that it gets to wait on memory more often. The limit is the memory bandwidth. Keep in mind that the access time of memory chips is only a fraction of the full cycle time of the memory system, which is usually two to three times the access time (i.e., 70 ns access yields about 200 ns write cycles, about 150 ns read cycles), and the memory system is busy for the full cycle time, unless it is bank-interleaved. Bank interleaving allows each bank to be doing a cycle simultaneously, assuming the reference stride hits the banks right. So, when you disable the primary mechanism used to reduce memory latency, you shouldn’t be surprised the chip runs as slow as a chip without those features."
Jay Lieske writes with a useful tip for Nisus users: "Nisus 3.05 and later uses a file called "Nisus Text Stationery" for ‘TEXT’ files that lack Nisus formatting – i.e. files from the net, C source code, etc. Set up a file with that name in your Nisus folder to have the default layout and font you want." [Thanks, Jay. Speaking of Nisus and Paragon Concepts, Paragon is working on a new program to radically simplify organizing and managing files in the disk hierarchy. We haven’t heard any details yet, but they posted to the Nisus list asking for name suggestions.]
Murph Sewall writes about AppleScript, the scripting module that Apple originally promised for System 7, "Apple may still envision AppleScript in a "future system" upgrade, but it hasn’t received any mention among the features expected in System 7.1, including QuickTime, ATM, MODE32, and <thank goodness> a fix for that non-relocatable code that makes System 7’s System memory use appear to grow without bound, plus of course, the improved SANE <math> that’s already in 7.0.1. Latest word is that 7.1 will not be announced at January’s Macworld (would you believe March?)."
Amaze — 206/820-7007
Mike O’Dell, whose email address we lost, sorry.
Jay Lieske — [email protected]
Murph Sewall — [email protected]