Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!usc!apple!agate!ucbvax!hcrlgw.crl.hitachi.co.JP!joe From: joe@hcrlgw.crl.hitachi.co.JP (Dwight Joe) Newsgroups: comp.sys.apple2 Subject: re: New Computer? Message-ID: <9012130536.AA19123@hcrlgw.crl.hitachi.co.jp> Date: 13 Dec 90 05:36:48 GMT Sender: daemon@ucbvax.BERKELEY.EDU Lines: 103 Motivation: What I find really vexing about the NeXTstation and the Amiga 3000 is that they are great machines _and_ that they are both at least $3000.00. Most college students can't afford that prices, never minding the $6000.00 price tag of the NeXTcube. If there is a major market among college students for machines in that price range, then we might as well admit that Edward Bennett, the former secretary of education, was right; he claimed that most students have got more than enough money to go to college and are even spending excess money on stereos and slick new automobiles. Let's face it. Most students are going to a state-supported university, the primary reasoning being that it is cheap. At one such school in the south, the students were complaining--loudly-- that tuition and miscellaneous fees were rising from about ~$400 to ~$600. Most people--most people not reading the USENET--cannot afford $3000 for a slick hi-tech NeXTstation or an Amiga 3000. Despite all the complaints about the MAC classic, which costs $750 at educational pricing, it's consumer demand far exceeds the supply. Do you why? Most students cannot afford to shell out more than $1000 for that slick NeXTstation, Amiga 3000, Mac II, etc. The other thing that I find really vexing is that some people seem to think that more processing power can be over-kill. They even think that the 6502 is just plain great. The trouble is that, if you ever decide that you want to do anything more than BASIC programming and very simple wordprocessing, then you can't. Why? Because running an FFT program, playing an elaborate video/arcade game with hundreds of moving features on the screen, and wordstar 4.0 will require significantly more processing power. In other words, regardless of how clever a programmer you might be, the hardware defines the upper limit of what you can do. I can assure you that wordstar 4.0 written for, say, a 4004 will be painfully slow regardless of how clever you might be as a programmer. Proposal: What I'm venturing is the following proposal for a new computer. First, it should use, within a given price range, the most powerful microprocessor available. Second, the total price for a system with CRT (preferably flat-panel), disk drive (preferably a micro-floppy drive), and main unit should be less than $1000. Let's keep, in mind, the fact that our target market is students. We've got to keep the price down on everything: hardward _and_ software. So, we'd like our system to be as compatible with current systems as possible to ease the transition of software packages from those systems to our new system. That means that the operating system should be UNIX, MS-DOS, OS/2, or compatible operating systems. Don't forget that writing completely new software packages for our system will force those packages to be highly priced because (1) the installed base of machines will be smaller than for a main-stream machine like PC clone and (2) a high ransom must be paid to Microsoft to bear the risk for writing to that smaller base, in addition to the expense of training people to operate on a new platform. Keeping that in mind, we should select a standard microprocessor. The 80486 and 68040 are out of the question. They are too expensive. I recommend the Sparc chip. Its specs can be licensed from Sun which has generously supported an open architecture. From those specs, we can design a cheap-o (Don't forget that we are building this system for students, most of whom are _poor_) RISC processor--no superscalar features and no branch-prediction. Heck, we can even use larger feature sizes, like 1.5 micron, to increase the yield and further reduce the price. Now, piece that together with 1 plain-vanilla 3.5" microfloppy drive, maybe 1 megabyte of memory, a keyboard, a UNIX disk, a C (not C++) programming language, a MAC Classic-like window system, a 640 X 400 pixel video black-and-white (color = too expensive) circuits, and plain CRT (preferably flat-panel). Yeah. It is very plain, but it is very cheap and _VERY_ powerful. If this sounds like a MAC Classic, upgraded to a RISC chip and UNIX, you're right in thinking that. I fail to see why the MAC Classic-type personal computer can't be upgraded to a RISC system. The el-cheapo Sparc chip, which we build with big feature sizes, should be even cheaper than a 68000, which requires even more transistors than a Sparc chip. The main memory should be fairly cheap by the time that the system is actually assembled. The price of 1 megabit chips is declining as 4 M and 16 M chips increase their market share. How about it Apple? Well, actually, how about it Sun (since Apple has already committed to the MAC Classic)? For $1500, a color system is available. (If anyone responds, please set their distribution to "global". Otherwise, I definitely won't be able to read the responses. Then again, I may not be able to read the responses anyhow due to a policy of censorship at my site. Don't ask me. I don't make the policy.)