pchris@ucbarpa.Berkeley.EDU (Chris Perleberg) (10/15/88)
It seems that the NeXT machine may have a few problems: 1) Outdated Processor Technology: NeXT just missed the wave of fast RISC processors. The 5 MIPS 68030 is completely out performed by the currently available RISC chips (Motorola, MIPS, Sparc) that run at approximately 20 VAX (they claim) MIPS. In a year or two, ECL versions of some of these RISC chips will be running at 40 to 50 MIPS. 2) Non-Standard NuBus Implementation: A small company like NeXT can't hope to create a competitive 3rd party board market for a non-standard bus. 3) Non-Standard Software: What software company would develop software for the special features of just one computer (NeXT Step)? How many copies of this software can they possibly sell? 4) Slow Optical Drive: In the past, optical drives have been significantly slower (seek times) than magnetic drives. What is the advantage of the optical drive? Cost must be less than that of the larger 330Mbyte $2K magnetic drive. But NeXT will be hurt once benchmarks come out for its i/o performance (using the optical drive). 5) Software Not Ready: The 9 month delay (optimistically) until solid software exists could kill NeXT, as Sun & Apple prepare competitive systems. Sun will probably keep to open systems and set some new standards, while Apple will probably stay proprietary. 6) Sun (I have heard) has sold 15,000 workstations to universitys. How many can NeXT expect to sell with its slow processor, non-standard bus/software, slow drive, and late software? Solutions to the above problems: What NeXT should do. 1) Develop a RISC based NeXT implementation as soon as possible. The advantage of Unix (Mach) is its idea of source level portability, rather than binary level compatibility. 2) Make "stub" boards that convert standard NuBus boards to the NeXT version of NuBus. These "stubs" would be placed between the NeXT slots and the standard NuBus boards. Longer Term Goal: Make the NeXT NuBus an international standard, much as Apple made its version of NuBus a standard. Possible Solution: Change the bus NOW to a standard, provide board converters for computers with the current bus. 3) Make NeXT proprietary software into standards, and beat Sun at its own game before Sun beats NeXT. This may mean making NeXT Step an overlay on X windows. The important thing is to develop a standard that can be (and is) used by all computers. 4) Give users the option of scrapping the optical drive and replacing it with the 330 Mbyte winchester. 5) Work like hell. Adopt standard software where ever possible. Distribute work to 3rd parties. Standardize where ever possible to beat Sun at its own game and to make the users happy. Allow Apple to be proprietary and dig its own grave. 6) Standardize and Speed-Up -> Increase sales. Chris Perleberg pchris@ucbarpa.berkeley.edu
richard@gryphon.CTS.COM (Richard Sexton) (10/15/88)
In article <26435@ucbvax.BERKELEY.EDU> pchris@ucbarpa.Berkeley.EDU (Chris Perleberg) writes: > >It seems that the NeXT machine may have a few problems: > >4) Slow Optical Drive: In the past, optical drives have been significantly > slower (seek times) than magnetic drives. What is the advantage of the > optical drive? Cost must be less than that of the larger 330Mbyte $2K > magnetic drive. But NeXT will be hurt once benchmarks come out for its > i/o performance (using the optical drive). What makes these drives so slow ? Surely it's not the actuator technology... A more cynical view would have it that there's nothing in a NeXT box that you could'nt put in an amiga (slot). What seems to set it apart is the software. Just like a Mac is different from an AT&T unix PC. -- What on Earth did Walt Disney have against cats ? richard@gryphon.CTS.COM {backbone...err, well connected site}!gryphon!richard
chari@juniper.uucp (Christopher Michael Whatley) (10/15/88)
I agree with most of what you say. But here are a few things I have picked up. In article <26435@ucbvax.BERKELEY.EDU> pchris@ucbarpa.Berkeley.EDU (Chris Perleberg) writes: > >It seems that the NeXT machine may have a few problems: > >1) Outdated Processor Technology: NeXT just missed the wave of fast RISC > processors. The 5 MIPS 68030 is completely out performed by the currently > available RISC chips (Motorola, MIPS, Sparc) that run at approximately > 20 VAX (they claim) MIPS. In a year or two, ECL versions of some of these > RISC chips will be running at 40 to 50 MIPS. They are supposedly developing their own RISC chip that is compatible with the 030. I don't know any more than that. I read this in a rumour column. (Grain-o-salt) Doesn't 20 RISC MIPS equal about 5 CISC MIPS? >3) Non-Standard Software: What software company would develop software for > the special features of just one computer (NeXT Step)? How many copies of > this software can they possibly sell? IBM has licensed NeXTStep for use with AIX on the PS/2 and RT PC. If that becomes an alternative to OS/2 then the software, which supposedly could be ported very easily, would reach a very wide market. I don't really have any idea how many RTs there are though. > >4) Slow Optical Drive: In the past, optical drives have been significantly > slower (seek times) than magnetic drives. What is the advantage of the > optical drive? Cost must be less than that of the larger 330Mbyte $2K > magnetic drive. But NeXT will be hurt once benchmarks come out for its > i/o performance (using the optical drive). I have wondered about this myself. It would be ridiculous to use this media for software distribution unless it was an OS update or something. I think it would be ridiculous to sell WriteNow on a 300mb disk. One of the infoworld reporters said that NeXT was considering using the modem for distribution. Ha! Chris -- $---------------$--------------------------------$-------------------------$ | Chris Whatley | mail chari@juniper.uucp | "Ever seen the chicken | | 512/453-4238 | chari@killer.dallas.tx.us | walk?" -Jeffrey | $---------------$--------------------------------$-------------------------$
dtynan@sultra.UUCP (Der Tynan) (10/15/88)
In article <26435@ucbvax.BERKELEY.EDU>, pchris@ucbarpa.Berkeley.EDU (Chris Perleberg) writes: > > It seems that the NeXT machine may have a few problems: > > 1) Outdated Processor Technology: NeXT just missed the wave of fast RISC > processors. The 5 MIPS 68030 is completely out performed by the currently > available RISC chips (Motorola, MIPS, Sparc) that run at approximately > 20 VAX (they claim) MIPS. In a year or two, ECL versions of some of these > RISC chips will be running at 40 to 50 MIPS. > This kind of argument drives me nuts! You could spend your life designing the "Latest and Greatest" thing into a system. I think you're *way* off the mark calling the 68030 outdated. Most of the world is using 8088's for pete sake! When IBM brought out their (yeugh!) PC, the processor technology was a tiny step up from a Z80. Today, most systems use 68000's. Jobs jumped the gun, and skipped the 68020 completely (as an aside, does the '30 have a built-in MMU?). In a year or two? ECL? That's a lot of good to somebody who has to deliver something on Oct 12! > 2) Non-Standard NuBus Implementation: A small company like NeXT can't hope to > create a competitive 3rd party board market for a non-standard bus. It has nothing to do with the size of the company. It's the size of the user market that determines this. We'll just have to wait and see. At $6500, it may still be too expensive to produce a big market. However, even if the company was nothing more than Jobs and a soldering iron, if the boxes sell, there'll be a competitive board market. > > 3) Non-Standard Software: What software company would develop software for > the special features of just one computer (NeXT Step)? How many copies of > this software can they possibly sell? Apple got away with it with the Mac, IBM pulled it off with the PC. What's the difference? The NeXT is running Mach. I've never seen one of these things, but I know most of my software will work on one, which is more than I could say for the PC when it first came out. Anyone remember the programs to convert 8080 CP/M code to run on an 8088?? > > 4) Slow Optical Drive: In the past, optical drives have been significantly > slower (seek times) than magnetic drives. What is the advantage of the > optical drive? Cost must be less than that of the larger 330Mbyte $2K > magnetic drive. But NeXT will be hurt once benchmarks come out for its > i/o performance (using the optical drive). The advantage of the optical drive, is it's *removable*. I agree, however, that a hard disk is also necessary. I mean, imagine being able to insert the GNU compiler source, and have it mounted in seconds. When you're done, move the object to the (hard-disk based) /bin directory, and take out the source disk. I think, however, that wishing the floppy-disk market away won't work. What happens to those of us who have mountains of old software we refer to now and again? If the market is education, it is important for 'joe user' to be able to send his stuff to his brother in Iowa, with a PC. > > 5) Software Not Ready: The 9 month delay (optimistically) until solid software > exists could kill NeXT, as Sun & Apple prepare competitive systems. Sun > will probably keep to open systems and set some new standards, while Apple > will probably stay proprietary. > The Mac took a year. So what? Right. Sun will have a competitive system AND "solid software" in less than nine months. Sure. The comparison I saw, showed Sun's answer as costing $13K. AND less performance. Anyway, see above. The basic tools provided will cover a lot of uses (more than MacPaint could :-) > 6) Sun (I have heard) has sold 15,000 workstations to universitys. How many > can NeXT expect to sell with its slow processor, non-standard bus/software, > slow drive, and late software? > *Slow* processor? Anyone would think from what you say, that the 80386 was a Cray in a leadless package? Don't forget; Mach is best known for being a "distributed" operating system. Also, the NeXT CPU is just a plug-in card. How about four more CPU's?? Mach can handle it. Can Sun OS?? > > Solutions to the above problems: What NeXT should do. > > 1) Develop a RISC based NeXT implementation as soon as possible. The advantage > of Unix (Mach) is its idea of source level portability, rather than binary > level compatibility. I take it you like RISC technology. I think you'd be a lot better off if you built your own. While you're at it, why don't you make it liquid nitrogen- cooled, with an ECL processor, running 100MHz. > > 2) Make "stub" boards that convert standard NuBus boards to the NeXT version of > NuBus. These "stubs" would be placed between the NeXT slots and the > standard NuBus boards. Longer Term Goal: Make the NeXT NuBus an > international standard, much as Apple made its version of NuBus a standard. > Possible Solution: Change the bus NOW to a standard, provide board > converters for computers with the current bus. > Ever hear of EBCDIC? Or the 3.75" micro-floppy? IBM has tried to "make a standard" more than once. Anyway, why not cripple it completely, and base it on an 8-bit S100 standard?? > > 3) Make NeXT proprietary software into standards, and beat Sun at its own game > before Sun beats NeXT. This may mean making NeXT Step an overlay on X > windows. The important thing is to develop a standard that can be (and is) > used by all computers. > Oh, I get it. There's a place in Delaware, where you send your source-listing, and a check for $50,000. By return mail, you get a "Certificate of Standard- ization" (No insult to the people of Delaware intended). You can *also* get a PhD in the same package. > 4) Give users the option of scrapping the optical drive and replacing it with > the 330 Mbyte winchester. Finally, a halfway decent comment. > > 5) Work like hell. Adopt standard software where ever possible. Distribute > work to 3rd parties. Standardize where ever possible to beat Sun at its > own game and to make the users happy. Allow Apple to be proprietary and > dig its own grave. > > Chris Perleberg > pchris@ucbarpa.berkeley.edu Sorry, Chris, but your article reeked of "It's good -- now, lets make it *even* better". There comes a phase in every project, when it's time to shoot the engineers and build the thing (stolen sans permission from someone). Where were you when IBM brought the industry *back* five years in '82? At last, someone offers a decent system, that actually *looks* like a system, instead of a toy, and everyone jumps on it, because it's not a VAX (or a Sun :-) The strange thing is, I don't even *like* Steve Jobs (nor do I have a cool $6500!). - Der -- Reply: dtynan@sultra.UUCP (Der Tynan @ Tynan Computers) {mips,pyramid}!sultra!dtynan Cast a cold eye on life, on death. Horseman, pass by... [WBY]
spolsky-joel@CS.YALE.EDU (Joel Spolsky) (10/15/88)
In article <26435@ucbvax.BERKELEY.EDU> pchris@ucbarpa.Berkeley.EDU (Chris Perleberg) writes ( I deleted a lot ): >It seems that the NeXT machine may have a few problems: >3) Non-Standard Software: What software company would develop software for > the special features of just one computer (NeXT Step)? How many copies of > this software can they possibly sell? This is ridiculous. [Enter 1982 mode] Who would bother writing software for the IBM/PC? How many copies of this software can they possibly sell? [End 1982 mode] As it turns out there's lots of software written for not-so-standard machines. Funny isn't it. >4) Slow Optical Drive: In the past, optical drives have been significantly > slower (seek times) than magnetic drives. What is the advantage of the > optical drive? (a) first 256M = $1495 (b) next 256M = $50 Any questions? >6) Sun (I have heard) has sold 15,000 workstations to universitys. How many > can NeXT expect to sell with its slow processor, non-standard > bus/software slow drive, and late software? Loads. It seems as if this slow-non-standard-late machine costs about half what the competition has to offer. See the Wall Street Journal. >Solutions to the above problems: What NeXT should do. > >1)Develop a RISC based NeXT implementation as soon as possible. The advantage > of Unix (Mach) is its idea of source level portability, rather than binary > level compatibility. Easy. The fast NuBus is there to allow several processors to coexist. I'll bet that ultimately you'll be able to plug in a RISC-y board, a lisp-machine board, etc, like the Mac-II/TI explorer hybrid. THAT would be nice for academic enviroments. Also, they plan to enable future versions of the OS to manage multiprocessing by assigning threads to different processors. > >2) Make "stub" boards that convert standard NuBus boards to the NeXT version of > NuBus. These "stubs" would be placed between the NeXT slots and the > standard NuBus boards. Longer Term Goal: Make the NeXT NuBus an > international standard, much as Apple made its version of NuBus a standard. > Possible Solution: Change the bus NOW to a standard, provide board > converters for computers with the current bus. > Weeeel, it looks to me as if the NuBus is really designed for multiprocessors, not for plugging in game ports and clock-calendars. The (very standard) SCSI bus can be used for peripherals. I'm not too worried about the incompatible NuBus; I'd rather have the fast NuBus. >4) Give users the option of scrapping the optical drive and replacing it with > the 330 Mbyte winchester. Hmm. How much would a 330M winchester + tape cartridge system cost? more than $1495? Actually what I though would be neat is to install a cheapo 40 Meg hard disk for swapping, /tmp, and important bits of the kernal. That would speed the thing up greatly for only several hundred dollars, and you'd still have the removable optical cartridges. There's even a 5.25" full-height bay for such a drive. Also when you took out the optical disk you could still run obscure useless commands like "ls" and "cat". ;-) +----------------+---------------------------------------------------+ | Joel Spolsky | bitnet: spolsky@yalecs uucp: ...!yale!spolsky | | | arpa: spolsky@yale.edu voicenet: 203-436-1483 | +----------------+---------------------------------------------------+ #include <disclaimer.h>
dorn@fabscal.UUCP (Alan Dorn Hetzel) (10/15/88)
In article <26435@ucbvax.BERKELEY.EDU> pchris@ucbarpa.Berkeley.EDU (Chris Perleberg) writes: > >It seems that the NeXT machine may have a few problems: > >1) Outdated Processor Technology: NeXT just missed the wave of fast RISC > processors. The 5 MIPS 68030 is completely out performed by the currently > available RISC chips (Motorola, MIPS, Sparc) that run at approximately > 20 VAX (they claim) MIPS. In a year or two, ECL versions of some of these > RISC chips will be running at 40 to 50 MIPS. > On this point, I'm not yet convinced that RISC machines are the only way to go. They will have their uses, but don't expect to see the death of CISC processors anytime soon. RISC places some significant additional strain on memory bandwidth for instruction fetch, and while in some programs this can be offset by instruction caches, in other programs it cannot... Also, Motorola and the other purveyors of 68xxx family products can be expected to keep improving their line as well (68040, higher clock rates, etc.) There is also the fact that the NeXT is a passive backplane system, with the CPU on a removable, (and therefore replaceable) board. Mach also supports multiple processors. While not exactly related, the system also contains a 56001 DSP processor, which will aid GREATLY in some types of processing which micros haven't been much good at previously. >2) Non-Standard NuBus Implementation: A small company like NeXT can't hope to > create a competitive 3rd party board market for a non-standard bus. > "a small company like NeXT" ... backed by H. Ross Perot. (Also, a small company like Apple has started a couple of busses itself) >3) Non-Standard Software: What software company would develop software for > the special features of just one computer (NeXT Step)? How many copies of > this software can they possibly sell? > Well, Mach is unix like enough to make portation of significant quantities of unix software not too onerous a task. As far as NeXT Step, I imagine it will do OK, considering the fact that IBM (yes, Big Blue) has licensed it to use on some of their systems as a user interface. I figure they might sell a FEW copies. >4) Slow Optical Drive: In the past, optical drives have been significantly > slower (seek times) than magnetic drives. What is the advantage of the > optical drive? Cost must be less than that of the larger 330Mbyte $2K > magnetic drive. But NeXT will be hurt once benchmarks come out for its > i/o performance (using the optical drive). > a) It may not be as slow as you think, electro-optical drives are near the leading edge of the optical pack in speed. b) one (1) advantage of the optical drive is that it is a 256Mb REMOVABLE disk subsystem. More blank disks cost about $50 each. c) Yes, costs will be MUCH less than the 330Mb magnetic disk, which still takes multiple platters and heads; The optical disk uses one platter. b) Even assuming their benchmarks will be dismal (and I don't think they will), purchase decisions are not made by benchmarks alone. >5) Software Not Ready: The 9 month delay (optimistically) until solid software > exists could kill NeXT, as Sun & Apple prepare competitive systems. Sun > will probably keep to open systems and set some new standards, while Apple > will probably stay proprietary. > Late software hurts, but they appear to have waited long enough with the introduction of the system itself that their software timetable can be met. >6) Sun (I have heard) has sold 15,000 workstations to universitys. How many > can NeXT expect to sell with its slow processor, non-standard bus/software, > slow drive, and late software? > > We will have to wait and see, but considering the fact that NeXT is going to concentrate exclusively on universities as a market, they may do pretty well. (A footnote, one local university here has already ordered 100 of them) >Solutions to the above problems: What NeXT should do. > > >1) Develop a RISC based NeXT implementation as soon as possible. The advantage > of Unix (Mach) is its idea of source level portability, rather than binary > level compatibility. > Oh, pish, you have RISC on the brain. A second (or third) processor board can always be added if warranted. You were talking about lack of software, well, there is a lot less software for RISC than for the 68xxx family. You complain the software will be late, but here you want to make it later..! >2) Make "stub" boards that convert standard NuBus boards to the NeXT version of > NuBus. These "stubs" would be placed between the NeXT slots and the > standard NuBus boards. Longer Term Goal: Make the NeXT NuBus an > international standard, much as Apple made its version of NuBus a standard. > Possible Solution: Change the bus NOW to a standard, provide board > converters for computers with the current bus. > I'm not sure there is any difference between their bus and the NuBus except clock rate. If that's the case, the 68000 is pretty tolerant of slower than expected devices, so maybe slower boards will work. If not, consider that IBM got away with introducing another bus, and now there are cards for it. Also, given the SCSI port for disks, the wide array of goodies already on board, and the substantial amount of memory in the base system, many users may be able to "limp" along with the basic system for quite some time. >3) Make NeXT proprietary software into standards, and beat Sun at its own game > before Sun beats NeXT. This may mean making NeXT Step an overlay on X > windows. The important thing is to develop a standard that can be (and is) > used by all computers. > Given the nature of the display hardware, X-Windows won't be any big deal to port. Also, display postscript (even though i'm not that fond of it myself), has good potential for becoming a display standard interface (at the lower level). And at the risk of becoming repetitive, NeXT Step will get a big boost towards becoming a standard by IBM's licensing of it. >4) Give users the option of scrapping the optical drive and replacing it with > the 330 Mbyte winchester. > The optical disk drive is, among other things, the software distribution media for the system (no more swapping a zillion floppies to install your favorite LARGE software package). Economies of scale attained by placing one as the "floppy" of each system make it quite inexpensive. I suspect many users will want two of them. Their winchester prices are a little on the high side (hell, quite a bit), but given the SCSI interface, I see strong third party availability of mass storage and tape backup and such. >5) Work like hell. Adopt standard software where ever possible. Distribute > work to 3rd parties. Standardize where ever possible to beat Sun at its > own game and to make the users happy. Allow Apple to be proprietary and > dig its own grave. > Given the nature of workaholics like Jobs and Perot, I expect they have been working like hell and then some since they started this thing. The deal with IBM demonstrates their intent to standardize. As far as universities go, Mach is a pretty reasonable choice for an OpSys standard, given its multiple processor support and ongoing development at CMU. >6) Standardize and Speed-Up -> Increase sales. > > Hey, if they could control the sales numbers, they would have it made for sure. Since they can only sell so many at a time and still support them well (they have to grow), I applaud their decision to go for the university market first. (Stepping down of my soapbox, I remain) Dorn gatech.edu!fabscal!dorn
mash@mips.COM (John Mashey) (10/16/88)
In article <5498@juniper.uucp> chari@juniper.UUCP (Christopher Michael Whatley) writes: ... >>1) Outdated Processor Technology: NeXT just missed the wave of fast RISC >> processors. The 5 MIPS 68030 is completely out performed by the currently >> available RISC chips (Motorola, MIPS, Sparc) that run at approximately >> 20 VAX (they claim) MIPS. In a year or two, ECL versions of some of these >> RISC chips will be running at 40 to 50 MIPS. > >They are supposedly developing their own RISC chip that is compatible with the >030. I don't know any more than that. I read this in a rumour column. This defies all logic. a) If it's compatible with an 030, it's not a RISC. If it's a design that makes the 68K architecture go faster (quite possible), it's still CISC, but maybe it's called a 68040. Do you think that NeXT thinks they can do that better than Moto? b) Building high-performance RISCs is harder than people thought. Even large companies with huge resources often prefer to buy rather than build. Why would a startup do this when their value-add is clearly in other areas, like user interface, software, and system integration? DON'T BELIEVE RUMOR COLUMNS THAT PRINT TRASH LIKE THIS! >(Grain-o-salt) Doesn't 20 RISC MIPS equal about 5 CISC MIPS? Depends on whose RISC MIPS and CISC MIPS they are, although that's at the edge of the distribution. I know our 20-MIPS M/2000s are 20X VAX-11/780 [i.e., spread over 15X-25X on real programs, versus VAX/VMS compilers] because I've benchmarked them. It certainly will be good to start seeing 68030s so one can benchmark them and figure out what they really are. Finally, ECL RISC chips will indeed be 40-50MIPS (or more) in next few years, but that's rather irrelevant to a NeXT discussion. I'd be amazed to see straight ECL chips in a desktop workstation, especially one aimed to be cheap. ECL chips are suitable for bigger machines, but they'r not cheap, they burn power, and they're HOT: a desktop would probably be an above-the-desktop, as it hovers on its fans. More seriously, you're much more likely to find ECL in servers than workstations. -- -john mashey DISCLAIMER: <generic disclaimer, I speak for me only, etc> UUCP: {ames,decwrl,prls,pyramid}!mips!mash OR mash@mips.com DDD: 408-991-0253 or 408-720-1700, x253 USPS: MIPS Computer Systems, 930 E. Arques, Sunnyvale, CA 94086
aglew@urbsdc.Urbana.Gould.COM (10/16/88)
>Standardize and Speed-Up -> Increase sales. > > Chris Perleberg > pchris@ucbarpa.berkeley.edu Well, I don't like writing this, since I believe in standardization, but may I point up some obvious corollaries?: - Standardize and speed-up (but speed up a bit more slowly than your competitors) => lost sales and customer migration - Standardize and speed-up (but choose the wrong area to work on future development, eg. providing bubble memory cards instead of new floppies) => lost sales and customer migration. - Avoid standards and speed-up => increase sales (but maybe not as much as you could have if you standardized, unless your performance is overwhelmingly better.) - Avoid standards and fail to speed up => customer loss, but more slowly than they would have if you standardized. - Standardize and speed-up, plus have a few proprietary features that are really attractive to use => increase sales, gain customers from other systems, and then "trap" them on your system. Can anyone guess why POSIX is such an incomplete standard, and what strategy the big vendors are going to try to use? Andy "Krazy" Glew. at: Motorola Microcomputer Division, Champaign-Urbana Development Center (formerly Gould CSD Urbana Software Development Center). mail: 1101 E. University, Urbana, Illinois 61801, USA. email: (Gould addresses will persist for a while) aglew@gould.com - preferred, if you have MX records aglew@fang.gould.com - if you don't ...!uunet!uiucuxc!ccvaxa!aglew - paths may still be the only way My opinions are my own, and are not the opinions of my employer, or any other organisation. I indicate my company only so that the reader may account for any possible bias I may have towards our products. PS. I promise to shorten this .signature soon.
james@bigtex.cactus.org (James Van Artsdalen) (10/16/88)
In <7774@gryphon.CTS.COM>, richard@gryphon.CTS.COM (Richard Sexton) wrote:
> What makes these drives so slow ? Surely it's not the actuator technology...
It's the mass of the head. Really fast "conventional" hard disks such
as the CDC Wren-III use a rotary voice coil: the arm turns somewhat
akin to a record player. The inertia of the head is therefore
critical. Exiting magentic heads are very lightweight and getting
lighter. Optics are apparently still very much more massive and could
well remain so indefinitely. If NeXT has both magnetic and optical
components on the head, it's not clear how they'll ever get the mass
of the head down as low as conventional disks.
As an aside, one interesting idea for linear voice coil drives is to
place several heads along the arm, dramatically cutting the maximum
seek distance. But it's not clear to me that maximum seek distance is
nearly as important as minimum track-to-track seek time: only stupid
operating systems place data completely at random across a drive (did
someone say AT&T unix? - even MS-DOS does better!).
--
James R. Van Artsdalen james@bigtex.cactus.org "Live Free or Die"
Home: 512-346-2444 Work: 338-8789 9505 Arboretum Blvd Austin TX 78759
henry@utzoo.uucp (Henry Spencer) (10/16/88)
In article <5498@juniper.uucp> chari@juniper.UUCP (Christopher Michael Whatley) writes: >They are supposedly developing their own RISC chip that is compatible with the >030.... Sigh. Isn't it sad: even in comp.arch, people have forgotten what "RISC" means. That's like trying to build a Concorde that's compatible with a donkey cart. >... One of the infoworld >reporters said that NeXT was considering using the modem for distribution. Ha! Why the "Ha!"? It's not a ridiculous idea. The AT&T Software Toolchest works pretty well. And it is cheaper to get X11 from UUNET via modem than from MIT via tape, not to mention faster... It also has the enormous advantage that it largely eliminates the very expensive human operations needed in conventional software distribution. If you ask the Software Toolchest people to send you a tape instead, they will say "forget it, we haven't the manpower". -- The meek can have the Earth; | Henry Spencer at U of Toronto Zoology the rest of us have other plans.|uunet!attcan!utzoo!henry henry@zoo.toronto.edu
bzs@xenna (Barry Shein) (10/16/88)
It does seem obvious that the main advantage of the removeable optical disk in the University environment is with University purchased clusters of machines. A student comes in with his/her disk, sits down at a machine (any machine on campus), plugs it in and boots. Other software etc can be gotten from NFS file servers which would mount on boot. Doesn't really matter which NFS server, any one on the local net will do (a little PARC-like clearinghouse software to broadcast for an appropriate server would take care of that "will the NeXT binaries server please mount itself on my /usr/local partition!") So, you can work near your next (NeXT?) class, then go over to the library and work some more etc. Not sure how mail would work in that scheme although post-office protocols should suffice (contact your mail server machine and fork a background job to suck over all your new mail as you get started, doesn't have to be awfully fast, going thru gateways would be fine, unlike NFS where it would be something to be avoided, the local server could figure out where your mail is for you easily enough.) Thus, the optical disk needs only the basic operating system (even a lot of the utilities could be mounted read-only via NFS, no need to have things like emacs on the local disk etc.) With swap I'll guess you'll use around 100MB or a little less for overhead. That leaves around 150MB for user space, not too bad, especially when compared with your average student's disk quota on a time-sharer (what? A very few MB usually.) The economics are that it shifts the cost of disk space to the student. At a University with 5,000 student users at $50/disk I come up with (scritch scritch) $250,000 in costs now shifted to the students (well, that's a funny number, per-semester? per-year? per-matriculation? also students are getting a lot more disk, but about $50 is probably what their current few MB is worth right now.) Assuming more read/write storage can then be had on the NFS systems, even temporarily, then things can be shuffled about by copying off a removeable disk, putting in a fresh one and rebooting and copying it all back. Pretty gross, but a student's time is pretty cheap (ie. you're also shifting that work to the student.) Given all that backups become stranger, unless there's some way to get one's disk backed up while on-line as a service by the University (doubtful.) Optical disks aren't horribly fragile, but they can get destroyed (eg. sat on in a back pocket, dropped in a bookbag) or lost. I suppose one could grab two machines off-hours and do a disk copy, at worst. Setting up lockers for students to store disks in might be a good idea (living in Boston makes me wonder how these disks respond to large temperature changes, having lived in other places makes me wonder how much they like beach sand.) The laser printer is even stranger, perhaps it was an afterthought. What you really want in that environment is a scheme to send printout to a high-speed printer which is shared (perhaps one located in the university copy-center where a cost accounting system is set up to pay as you go, there might be small satellites set up near each workroom.) Conclusion: In it's current configuration it seems to not be a "student" machine per se but a Computing Center machine, something designed for their model of providing computing (rooms full of machines, let the students buy and manage all the disk space, no need for [much] operations etc.) Not too dumb, quite possibly very clever, not sure where it leaves a certain trend towards owning the machine, working from a dorm room etc. If that were the interest it would have an ISDN connector. -Barry Shein, ||Encore||
hagerman@maxwell.ece.cmu.edu (John Hagerman) (10/16/88)
I don't understand all the fuss about the optical disk. This is a young technology, and it will be a very unusual technology if it doesn't quickly become cheaper. I'm rather amazed at a removable magneto-optical disk for only $50. As for software distribution, even if the disk price doesn't come down there will soon be a number of options. If educational institutions really are the target, then grabbing software off the network seems most likely. Otherwise, get a cheap floppy drive. John.Hagerman@Maxwell.ece.cmu.edu
henry@utzoo.uucp (Henry Spencer) (10/16/88)
In article <9287@bigtex.cactus.org> james@bigtex.cactus.org (James Van Artsdalen) writes: >As an aside, one interesting idea for linear voice coil drives is to >place several heads along the arm, dramatically cutting the maximum >seek distance... Uh, Jim, you may not be aware of this, but the Fujitsu Eagle has two heads per surface on its linear actuator. Not a new idea. -- The meek can have the Earth; | Henry Spencer at U of Toronto Zoology the rest of us have other plans.|uunet!attcan!utzoo!henry henry@zoo.toronto.edu
ns@cat.cmu.edu (Nicholas Spies) (10/16/88)
In article <9287@bigtex.cactus.org> james@bigtex.cactus.org (James Van Artsdalen) writes: ... >lighter. Optics are apparently still very much more massive and could >well remain so indefinitely. ... It's not too difficult to imagine that the optics of a WORM drive read/write head could be reduced to a slender optical fiber leading to a stationary laser. Another approach might be to use holographic optics... :-) -- Nicholas Spies ns@cat.cmu.edu.arpa Center for Design of Educational Computing Carnegie Mellon University
crum@lipari.usc.edu (Gary L. Crum) (10/16/88)
In article <3884@encore.UUCP> bzs@xenna (Barry Shein) writes: > >The laser printer is even stranger, perhaps it was an afterthought. >What you really want in that environment is a scheme to send printout >to a high-speed printer which is shared (perhaps one located in the >university copy-center where a cost accounting system is set up to pay >as you go, there might be small satellites set up near each workroom.) > BYTE mentions "you could use a cube with a NeXT laser printer to act as a print server on a network" and "The cube can print to non-NeXT PostScript printers using its serial ports and Unix printer drivers." I think the Berkeley lpd (with /etc/printcap) system is used, so really, the shared printer could be a NeXT laser printer connected to a staff-maintained machine, a LaserWriter IINTX connected to any other UNIX system with lpd, or something complete esoteric such as the DEC LPS-40 fast PostScript printer connected to a VAX running VMS but still accessible from all Suns here at USC. The possibilities are endless. With a localtalk/ethernet gateway such as the Kinetics KFPS-4 and the free Columbia Appletalk Package (CAP) for UNIX, even AppleTalk printers could be used. Actually, CAP's Printer Access Protocol (PAP) implementation is quite fun. A program called "tlw" lets any UNIX user interactively talk to the PostScript interpreter (executive) of any LaserWriter on accessible AppleTalk networks. That leads to my next paragraph... Barry's environment where students boot cubes off their own platters poses many interesting security problems! In such an environment, cubes cannot "trust" each other because users have their own system disks and hence all users are superusers for their respective machines. In existing Sun installations that I have seen, it is technically possible for users to boot of their own disk or tape, but such practice is not intended and is detected by otherwise unexplained machine downtime. Perhaps the environment where users are allowed to boot NeXT cubes off their own software is analogous to existing environments with PCs on ethernet running TCP/IP protocols including NFS, but in the NeXT case the security issues are more apparent because the user has control over a larger and more mature system of software. I sure hope that this free, boot-it-yourself environment becomes the norm, but mixing UNIX and personal computing ideologies sure produces fireworks. I ran MACH on a Sun-3/60C last summer, albeit without NFS. It is interesting that rsh/rlogin and their daemons were absent from the distribution. Furthermore, compiling the source to those programs from 4.3BSD did not result in working services right off because MACH changes the function of ruserok() and other things. The MACH I used even ran most SunOS (pre-4.0) binaries, with notable exceptions being those that used SunView libraries. The brilliant engineers at NeXT and their group of advisors from academia undoubtedly have plans for cube cluster configurations as well thought out as the NeXT system itself, but until we all learn what those plans are, it sure is fun to speculate and "publish" our ideas. Perhaps we could use this forum for the exchange of ideas for application programs that the NeXT machine makes possible, especially those which would be less practical without such a capable, well-packaged computer system. (Keep in mind that readers working for software developers might use ideas, or maybe just chuckle because they are working on such things or better things. The proposal is for the reader in academia like myself, in a position to share ideas.) For example, how about (don't laugh!): * if not already provided, a program to convert Mathematica output to publication-quality PostScript description. This might be easiest with the source code to Mathematica (or rather, easiest for Stephen Wolfram to do). * given the above, a WSYWIG editor extended to keep Mathematical formulas in some Mathematica-readable format but displayed in pretty format. So, papers with mathematics in them could be constructed using cut-and-paste between Mathematica and the editor. Papers could be send over email as ASCII and MacWrite documents are now, and the receiver could cut equations out of documents and "experiment" with them, using Mathematica on his/her system. I guess I would ultimately like to see a standard format for the communication of mathematical proof (logic) without the free use of natural language, but that is too ambitious. * a multi-user interactive sketchpad, primarily for use to convey graphical information while connected by conventional phone. Or, perhaps packets of voice could be exchanged rapidly enough over ethernet along with the graphical information to replace the use of phone for some communication. Scott Dyer of the Ohio Supercomputer Center mentioned something like this multi-user interactive graphical sketchpad in his talk at the HP Graphics Symposium last July. It ran under suntools -- this clearly requires only networked computers. I won't post for a while after this, I promise. Please, somebody let me know if my choice or amount of material was not appropriate. Yay Trojans!
csimmons@hqpyr1.oracle.UUCP (Charles Simmons) (10/16/88)
In article <26435@ucbvax.BERKELEY.EDU> pchris@ucbarpa.Berkeley.EDU (Chris Perleberg) writes: > >It seems that the NeXT machine may have a few problems: > >3) Non-Standard Software: What software company would develop software for > the special features of just one computer (NeXT Step)? How many copies of > this software can they possibly sell? Oops... Note that IBM has licensed NeXT Step and plans to port it to run on top of AIX. Potentially, OSF will go with AIX as its operating system, and this would suggest that NeXT Step would be likely to appear in any OSF offering. In any event, with IBM backing the NeXT software, NeXT doesn't have to worry about this little item. >4) Slow Optical Drive: In the past, optical drives have been significantly > slower (seek times) than magnetic drives. What is the advantage of the > optical drive? Cost must be less than that of the larger 330Mbyte $2K > magnetic drive. But NeXT will be hurt once benchmarks come out for its > i/o performance (using the optical drive). Something to factor into your price performance picture: I believe another article pointed out that the optical drive was under $1500 dollars. So, for $1550, I get 256MB of disk storage. For $1600 I get 512MB of disk storage. For $1700 I get 1 Gigabyte of disk storage. Admittedly, my seek time gets a little bit long (2 to 3 seconds) when I have to swap disks... For the 330 MB winchester, I'm fairly certain that you can't pop the disk in and out of the machine. To get massive amounts of storage, you'ld have to attach a tape drive or floppy disk drive. 256 MB of floppy disk storage, in volume, should cost you around $250. For the tape drive, your seek time is probably somewhat worse than 2 to 3 seconds. So, the advantage of the optical drive is the ability to make available massive amounts of storage at low cost and reasonable accessibility. >6) Sun (I have heard) has sold 15,000 workstations to universitys. How many > can NeXT expect to sell with its slow processor, non-standard bus/software, > slow drive, and late software? My understanding is that NeXT needs to sell 10,000 machines over the next two years. >2) Make "stub" boards that convert standard NuBus boards to the NeXT version of > NuBus. These "stubs" would be placed between the NeXT slots and the > standard NuBus boards. Longer Term Goal: Make the NeXT NuBus an > international standard, much as Apple made its version of NuBus a standard. > Possible Solution: Change the bus NOW to a standard, provide board > converters for computers with the current bus. To some extent, this may have been done. As Barry Lustig points out, NeXT well be making available a chip that implements a NuBus interface. Cost of the chip: about $25. It would be interesting to hear how easy this chip makes it to build boards for the NeXT NuBus, or to convert a board from 10MhZ to 25Mhz. >3) Make NeXT proprietary software into standards, and beat Sun at its own game > before Sun beats NeXT. This may mean making NeXT Step an overlay on X > windows. The important thing is to develop a standard that can be (and is) > used by all computers. As I mentioned above, IBM has already made NeXT proprietary software into a standard. Also, since NeXT runs a Unix-based operating system, it' going to be real easy for companies to port their product from their current Unix based machines to the NeXT machine. The User Interface provided by the machine is supposed to be real real nice in that it makes it extremely simple for programmers to put together a user interface on top of their existing application. > Chris Perleberg
werner@utastro.UUCP (Werner Uhrig) (10/16/88)
[ please stop cross-posting articles with alt.groups; it causes the following unpleasantness to happen: inews: Unknown newsgroup alt.next News saved in /us/a/werner/dead.article Article not posted - exit status 256 fight for comp.sys.next; vote, forward next-articles and hate- mail to the nearest (former) net-god; send out create-control messages for comp.sys.next. remember the Boston Tea Party, the Alamo and the Pentagon Papers; this country is founded and served by the traditions of Civil Disobedience!!! Even Da' Quayle knew better than to serve a bad cause in Vietnam. ...now back to our regularly scheduled program.... ] > >4) Slow Optical Drive: Cost must be less than of the larger 330Mbyte $2K > It would be ridiculous to use this media for software distribution. > [just as] ridiculous as to sell WriteNow on a 300mb disk. An Infoworld > reporters said that NeXT was considering using the modem for distribution. now, a 330Meg Winchester isn't much help either! but you CAN take your optical pseudo-CD to the store and have them record a copy of the software while you pay and pick up your manual. They could even use some form of hardware protection in the "master" disk to prevent pirating (Zeus help us!) as all the software you are likely to ever acquire (pay for) surely will fit together with the OS on one disk. and distributing via modem isn't impractical at all. Consider the update distribution solved also. Go, NeXT !!! actually, what makes anyone assume that you can't use old-style floppy drives with the machine? just because one doesn't come with the machine means nothing - they simply assume that we already have one lying around somewhere ... :-) a new market for those old external 400k drives of the Mac; useful to install software on the NeXT optical platter. -- --------------------> PREFERED-RETURN-ADDRESS-FOLLOWS <--------------------- (ARPA) werner@rascal.ics.utexas.edu (Internet: 128.83.144.1) (INTERNET) werner%rascal.ics.utexas.edu@cs.utexas.edu (UUCP) ..!utastro!werner or ..!uunet!rascal.ics.utexas.edu!werner
chari@juniper.uucp (Christopher Michael Whatley) (10/16/88)
In article <1988Oct16.022032.29382@utzoo.uucp> henry@utzoo.uucp (Henry Spencer) writes: >In article <5498@juniper.uucp> chari@juniper.UUCP (Christopher Michael Whatley) writes: >>[Cisc comatible risc (ok it did sound dumb)] > >Sigh. Isn't it sad: even in comp.arch, people have forgotten what "RISC" >means. That's like trying to build a Concorde that's compatible with a >donkey cart. Excuse me for not being precise enough. To quote.... "It's a sort of RISC/CISC hybrid [which is] pin-compatible with the 68030. [It] supports the motorola instruction set at between 1 and 2 instructions per cycle..." > >>... One of the infoworld >>reporters said that NeXT was considering using the modem for distribution. Ha! > >Why the "Ha!"? It's not a ridiculous idea. The AT&T Software Toolchest... Ok, it's not such a ridiculous idea. However, your average student (me), does not want to call up uunet and sit ther for an hour watching the retry count go up long distance. If the university could have the software available by modem that would be great. You could go into the computer store at school pay your money and get a special password and login-name. That wouldn't be too hard to manage I suppose. Chris (Gee I miss Toronto, I went to U of T last year) -- $---------------$--------------------------------$-------------------------$ | Chris Whatley | mail chari@juniper.uucp | "Ever seen the chicken | | 512/453-4238 | chari@killer.dallas.tx.us | walk?" -Jeffrey | $---------------$--------------------------------$-------------------------$
crum@lipari.usc.edu (Gary L. Crum) (10/16/88)
In article <3884@encore.UUCP> bzs@xenna (Barry Shein) writes: > >The laser printer is even stranger, perhaps it was an afterthought. >What you really want in that environment is a scheme to send printout >to a high-speed printer which is shared (perhaps one located in the >university copy-center where a cost accounting system is set up to pay >as you go, there might be small satellites set up near each workroom.) > BYTE mentions "you could use a cube with a NeXT laser printer to act as a print server on a network" and "The cube can print to non-NeXT PostScript printers using its serial ports and Unix printer drivers." I think the Berkeley lpd (with /etc/printcap) system is used, so really, the shared printer could be a NeXT laser printer connected to a staff-maintained machine, a LaserWriter IINTX connected to any other UNIX system with lpd, or something complete esoteric such as the DEC LPS-40 fast PostScript printer connected to a VAX running VMS but still accessible from all Suns here at USC. The possibilities are endless. With a localtalk/ethernet gateway such as the Kinetics KFPS-4 and the free Columbia Appletalk Package (CAP) for UNIX, even AppleTalk printers could be used. Actually, CAP's Printer Access Protocol (PAP) implementation is quite fun. A program called "tlw" lets any UNIX user interatively talk to the PostScript interpreter (executive) of any LaserWriter on accessible AppleTalk networks. That leads to my next paragraph... Barry's environment where students boot cubes off their own platters poses many interesting security problems! In such an environment, cubes cannot "trust" each other because users have their own system disks and hence all users are superusers for their respective machines. In existing Sun installations that I have seen, it is technically possible for users to boot of their own disk or tape, but such practice is not intended and is detected by otherwise unexplained machine dowtime. Perhaps the environment where users are allowed to boot NeXT cubes off their own software is analogous to existing environments with PCs on ethernet running TCP/IP protocols including NFS, but in the NeXT case the security issues are more apparent because the user has control over a larger and more mature system of software. I sure hope that this free, boot-it-yourself environment becomes the norm, but mixing UNIX and personal computing ideologies sure produces fireworks. I ran MACH on a Sun-3/60C last summer, albeit without NFS. It is interesting that rsh/rlogin and their daemons were absent from the distribution. Furthermore, compiling the source to those programs from 4.3BSD did not result in working services right off because MACH changes the function of ruserok() and other things. The MACH I used even ran most SunOS (pre-4.0) binaries, with notable exceptions being those that used SunView libraries. The brilliant engineers at NeXT and their group of advisors from academia undoubtedly have plans for cube cluster configurations as well thought out as the NeXT system itself, but until we all learn what those plans are, it sure is fun to speculate and "publish" our ideas. Perhaps we could use this forum for the exchange of ideas for application programs that the NeXT machine makes possible, especially those which would be less practical without such a capable, well-packaged computer system. (Keep in mind that readers working for software developers might use ideas, or maybe just chuckle because they are working on such things or better things. The proposal is for the reader in academia like myself, in a position to share ideas.) For example, how about (don't laugh!): * if not already provided, a program to convert Mathematica output to publication-quality PostScript description. This might be easiest with the source code to Mathematica (or rather, easiest for Stephen Wolfram to do). * given the above, a WSYWIG editor extended to keep Mathematical formulas in some Mathematica-readable format but displayed in pretty format. So, papers with mathematics in them could be constructed using cut-and-paste between Mathematica and the editor. Papers could be send over email as ASCII and MacWrite documents are now, and the receiver could cut equations out of documents and "experiment" with them, using Mathematica on his/her system. I guess I would ultimately like to see a standard format for the communication of mathematical proof (logic) without the free use of natural language, but that is too ambitious. * a multi-user interactive sketchpad, primarily for use to convey graphical information while connected by conventional phone. Or, perhaps packets of voice could be exchanged rapidly enough over ethernet along with the graphical information to replace the use of phone for some communication. Scott Dyer of the Ohio Supercomputer Center mentioned something like this multi-user interactive graphical sketchpad in his talk at the HP Graphics Symposium last July. It ran under suntools -- this clearly requires only networked computers. I won't post for a while after this, I promise. Please, somebody let me know if my choice or amount of material was not appropriate. Yay Trojans!
ward@cfa.harvard.EDU (Steve Ward) (10/17/88)
In article <26435@ucbvax.BERKELEY.EDU>, pchris@ucbarpa.Berkeley.EDU (Chris Perleberg) writes: > > It seems that the NeXT machine may have a few problems: > > 1) Outdated Processor Technology: NeXT just missed the wave of fast RISC > processors. The 5 MIPS 68030 is completely out performed by the currently > available RISC chips (Motorola, MIPS, Sparc) that run at approximately > 20 VAX (they claim) MIPS. In a year or two, ECL versions of some of these > RISC chips will be running at 40 to 50 MIPS. > Well, the processor technology is hardly outdated. Being either CISC or RISC is hardly enough, in and of itself, to obsolete an MPU. Cost, performance, availability, and software product base are a few major considerations, for example. I suspect these latter points heavily influenced the MPU choice. 5 MIPS isn't the fastest, but who says it has to be the fastest to be excellent? There are a lot of pieces to a system beyond the MPU itself. > 2) Non-Standard NuBus Implementation: A small company like NeXT can't hope to > create a competitive 3rd party board market for a non-standard bus. > I am unsure of the situation here, but I have heard that the "MAC" NuBus is nonstandard and the NeXT implementation conforms fully to the NuBus standard. This is an area that needs some clarification. Any NuBus experts care to clarify this topic? > 3) Non-Standard Software: What software company would develop software for > the special features of just one computer (NeXT Step)? How many copies of > this software can they possibly sell? > I suspect that Mach is no more or no less a portable Unix environment than BSD Unix flavors and ATT&T Unix flavors (or Sun or DEC, etc). I suppose the big issue here is yet-another-windows/graphics/user interface. I guess things like Phigs and GKS will help here, as would X. If NeXT doesn't bring these items out, then third parties probably will. On balance, NeXT does not look any less "standard" or portable than other workstation -- NeXT does incorporate Display Postscript, a published specification, I believe. > 4) Slow Optical Drive: In the past, optical drives have been significantly > slower (seek times) than magnetic drives. What is the advantage of the > optical drive? Cost must be less than that of the larger 330Mbyte $2K > magnetic drive. But NeXT will be hurt once benchmarks come out for its > i/o performance (using the optical drive). For diskless nodes using only tthe optical drive will probably be okay. This drive is much faster write-wise (according to propaganda - I personally do not yet know) than WORM drives. Standalone systems will (my guess) require a winchester disk drive for performance requirements of most users. [comments on what NeXT should do to fix its problems deleted] I agree that standardization of hardware and software is an important issue. I hope that their NuBus implementation is a standard-conforming one. Also, a high performance RISC machine would be good, but I am sure this will follow if NeXT does well. Frankly, right now, I think product quality, reliability, deliverability, price, and product support will the major factors affecting NeXT success, as their machine seems to functionally and technically good enough to penetrate their target market. We'll know soon enough. Steven M. Ward ward@cfa.harvard.edu
deraadt@dataspan.UUCP (Theo De Raadt) (10/17/88)
In article <5498@juniper.uucp>, chari@juniper.uucp writes: >In article <26435@ucbvax.BERKELEY.EDU> pchris@ucbarpa.Berkeley.EDU writes: >>It seems that the NeXT machine may have a few problems: >> >>1) Outdated Processor Technology: NeXT just missed the wave of fast RISC >> processors. The 5 MIPS 68030 is completely out performed by the currently >> available RISC chips (Motorola, MIPS, Sparc) that run at approximately... > >They are supposedly developing their own RISC chip that is compatible with the >030. I don't know any more than that. I read this in a rumour column. >(Grain-o-salt) Doesn't 20 RISC MIPS equal about 5 CISC MIPS? I am getting sick and tired of this RISC/CISC battle. Come on guys, measure RISC instructions against CISC microcode level instructions and they work out to about the same thing. As far as I can see, it appears that we are always battling memory speeds. Not trying to start a flame war, but 030's are faster than Sun 4's. I puke trying to write assembly on RISC machines. I love CISC though, because you find your bottleneck, and just jump in and start ripping some things out, and rewrite the routine..-- _____ _ ----------------------------------- / / / / \ _ _ /_/_ Theo de Raadt: (403) 289-4620 / /_ _ ___ __/_ /__/ _\ _\ __/ / DATASPAN / / /</_(_) (_/</_/ \_(_/\(_/\(_/_(_/ ..!alberta!calgary!dataspan!deraadt
steve@polyslo.CalPoly.EDU (Steve DeJarnett) (10/17/88)
In article <26435@ucbvax.BERKELEY.EDU> pchris@ucbarpa.Berkeley.EDU (Chris Perleberg) writes: >It seems that the NeXT machine may have a few problems: >3) Non-Standard Software: What software company would develop software for > the special features of just one computer (NeXT Step)? How many copies of > this software can they possibly sell? Well, given that IBM has bought some of the User Interface technology from NeXT (licensed it, I guess is actually the case), if IBM ever makes use of it, you will probably find quite a few developers willing to write programs to work under this environment (of course, if IBM is buying this to keep NeXT from making it (i.e. they won't do anything with it, and NeXT will flounder and die due to no software available for their machine), then NeXT may be short- lived). ------------------------------------------------------------------------------- | Steve DeJarnett | Smart Mailers -> steve@polyslo.CalPoly.EDU | | Computer Systems Lab | Dumb Mailers -> ..!ucbvax!voder!polyslo!steve | | Cal Poly State Univ. |------------------------------------------------| | San Luis Obispo, CA 93407 | BITNET = Because Idiots Type NETwork | -------------------------------------------------------------------------------
rfarris@serene.CTS.COM (Rick Farris) (10/17/88)
In article <1988Oct16.054306.1884@utzoo.uucp> (Henry Spencer) writes: >In article <9287@bigtex.cactus.org> (James Van Artsdalen) writes: >>As an aside, one interesting idea for linear voice coil drives is to >>place several heads along the arm, dramatically cutting the maximum >>seek distance... > >Uh, Jim, you may not be aware of this, but the Fujitsu Eagle has two >heads per surface on its linear actuator. Not a new idea. (Darn, I wish I had the balls to say "Uh, Henry,"...) Actually, mulitiple heads per surface is an old idea. It used to be quite common to have "head-per-track" drives. Of course these were generally single platter machines. I believe the reason for shifting to a multiple platter "head-per-surface" scheme was cost. Actually, there are still a few applications where head-per-track drives are still used. I believe Univac (err, Sperry, errr, Unisys) makes one for military aircraft. 100 MB as I remember. The problem with placing multiple heads on a *moving* actuator is that each head adds mass. It may not have to move as far, but it's slower getting started (for a given servo system) and slower stopping. Any physicists want to talk about the tradeoffs between moving a large mass a short distance, and moving a small mass a long distance? Rick Farris rfarris@serene.cts.com voice (619) 259-6793 POB M KCBIW public access 259-7757 Del Mar CA 92014 ...!uunet!serene!rfarris serene.uucp 259-3704
mash@mips.COM (John Mashey) (10/17/88)
In article <5549@juniper.uucp> chari@juniper.UUCP (Christopher Michael Whatley) writes: >In article <1988Oct16.022032.29382@utzoo.uucp> henry@utzoo.uucp (Henry Spencer) writes: >>In article <5498@juniper.uucp> chari@juniper.UUCP (Christopher Michael Whatley) writes: >>>[Cisc comatible risc (ok it did sound dumb)] >> >>Sigh. Isn't it sad: even in comp.arch, people have forgotten what "RISC" >>means. That's like trying to build a Concorde that's compatible with a >>donkey cart. > >Excuse me for not being precise enough. To quote.... "It's a sort of RISC/CISC >hybrid [which is] pin-compatible with the 68030. [It] supports the motorola >instruction set at between 1 and 2 instructions per cycle..." 1) The 68030 is "RISCier" than the 68020. Specifically, it went from 3 cycle bus access to 2 cycle access, and cycle counts were pared down elsewhere, I believe. Note that, as we've said before, you can always make CISC go faster by heavier pipelining+parallelism, but usually at a cost in hardware complexity and gate-count. The NeXT box does use a 68030. 2) "between 1 and 2 instructions per cycle": I think this is backwards. I might believe 3-4 cycles/instruction. Lots of people are talking about dual-issue micros, which is what you need to get 1-2 instrs/cycle, but I can't think of any that are yet on the market, and making CISCs do that will be truly exciting. -- -john mashey DISCLAIMER: <generic disclaimer, I speak for me only, etc> UUCP: {ames,decwrl,prls,pyramid}!mips!mash OR mash@mips.com DDD: 408-991-0253 or 408-720-1700, x253 USPS: MIPS Computer Systems, 930 E. Arques, Sunnyvale, CA 94086
ajdenner@athena.mit.edu (Alexander J Denner) (10/17/88)
In article <452@oracle.UUCP> csimmons@oracle.UUCP (Charles Simmons) writes: >Oops... Note that IBM has licensed NeXT Step and plans to port it to >run on top of AIX. Potentially, OSF will go with AIX as its >operating system, and this would suggest that NeXT Step would be >likely to appear in any OSF offering. In any event, with IBM backing >the NeXT software, NeXT doesn't have to worry about this little item. > >As I mentioned above, IBM has already made NeXT proprietary software >into a standard. > >Also, since NeXT runs a Unix-based operating system, it' going to be >real easy for companies to port their product from their current >Unix based machines to the NeXT machine. IBM may use NeXT Step, but I would not count on it. IBM has a history of licensing and/or buying any technology or software they may someday think about using. For OSF, IBM has licensed or considered using: Next, Presentation Manager (what!), X Windows, HP (whatever they call their new windowing system), and others. To my knowledge, none of them have been given any special attention. In fact, chances are IBM may not even accept any of the systems above and write or contract a new one. ------------------------------------------------------------------------------ Alexander J. Denner ajdenner@athena.mit.edu 234 Baker House, 342 Memorial Drive mit-eddie!mit-athena!ajdenner Cambridge, MA 02139
jkrueger@daitc.daitc.mil (Jonathan Krueger) (10/17/88)
In article <12834@oberon.USC.EDU> crum@lipari.usc.edu (Gary L. Crum) writes: >a program to convert Mathematica output to >publication-quality PostScript description. Why not to TeX instead? It's less device dependent, licenses need not be obtained from Adobe, and it can be interchanged to a larger variety of systems. And it makes use of an existing system of mathematical typography; why reinvent the wheel? -- Jon
jkrueger@daitc.daitc.mil (Jonathan Krueger) (10/17/88)
In article <452@oracle.UUCP> csimmons@oracle.UUCP (Charles Simmons) writes: >Something to factor into your price performance picture: I believe another >article pointed out that the optical drive was under $1500 dollars. So, >for $1550, I get 256MB of disk storage. For $1600 I get 512MB of disk >storage. For $1700 I get 1 Gigabyte of disk storage. Admittedly, my >seek time gets a little bit long (2 to 3 seconds) when I have to swap disks... Jukebox. I hear a voice saying, "the NeXT jukebox...for when you need the NeXT disk". I also hear a voice saying "backup" and also "our 400 disk box stores 100 gigabytes." Opens up whole new horizons in optimizing allocation of disk resources. Such things exist today, but tend to be expensive and highly vendor-specific, and don't offer rewrite/erase capabilities as far as I know. -- Jon
dorn@fabscal.UUCP (Alan Dorn Hetzel) (10/17/88)
A couple of following comments: 1) Re: ISDN connector if they meant it for dorm use. Nope, ISDN isn't stabilized yet (I know that firsthand) They included an embedded Ethernet connection, which is probably better anyway for an on-campus environment. 2) Re: intended audience on campus... 1) backing up disks --- some of the machines can have two drives. that is an available option. Also, nothing says that the students *have* to keep their files on the optical disk instead of file server, or can't keep copies in multiple places. 2) disk cost .. hey, a couple boxes of decent floppies can set you back almost that much. (let's not even discuss textbook costs here) 3) Don't forget the professor/researched as an intended audience. The 56001 DSP chip should be very attractive to some engineering types... Well, you get the idea.... there are many faces to this thing. It will be fun to watch how it performs in the market. Dorn gatech.edu!fabscal!dorn
mende@athos.rutgers.edu (Bob Mende Pie) (10/17/88)
In article <5549@juniper.uucp> chari@juniper.uucp (Christopher Michael Whatley) writes about the subject of getting updates via modem. > Ok, it's not such a ridiculous idea. However, your average student (me), > does not want to call up uunet and sit ther for an hour watching the > retry count go up long distance. If the university could have the > software available by modem that would be great. You could go into the > computer store at school pay your money and get a special password and > login-name. That wouldn't be too hard to manage I suppose. There are a couple of things that may make this useful... We must remember that NeXT plans to make a university responsible for distribution and I will assume to some extent maintence/support of NeXT machines. The $50 cost of a optical disk is *not* unreasonable if you only have to send out one disk to a university. The university could allow people to get the updated software via modem or by use of a two drive machine. While I think that the optical disk is good, NeXT *SHOULD* adopt a standard format for both 3.5" floppys and 5.25" floppys. I would think the MS-DOS format for floppys, no flames, is a standard (a number of them) that would allow file transfer to those who only have a IBMish machine. I can imagine a box that hangs off your SCSI port, that has a 5.25" 1.44Meg floppy on it as well as plugs to hang more floppys off of. This way vendors could offer their software on 1.44 meg floppys as well as optical disks. /Bob... -- {...}!rutgers!mende mende@aramis.rutgers.edu mende@zodiac.bitnet Hi there! This is just a note from me, to you, to tell you, the person reading this note, that since my operation I have not had a strange thought, well except the one about the MAMBA is a BLUE SUIT.
wetter@cit-vax.Caltech.Edu (Pierce T. Wetter) (10/17/88)
> * if not already provided, a program to convert Mathematica output to >publication-quality PostScript description. This might be easiest with the >source code to Mathematica (or rather, easiest for Stephen Wolfram to do). > > * given the above, a WSYWIG editor extended to keep Mathematical formulas >in some Mathematica-readable format but displayed in pretty format. So, >papers with mathematics in them could be constructed using cut-and-paste >between Mathematica and the editor. Papers could be send over email >as ASCII and MacWrite documents are now, and the receiver could cut >equations out of documents and "experiment" with them, using Mathematica >on his/her system. I guess I would ultimately like to see a standard format >for the communication of mathematical proof (logic) without the free use >of natural language, but that is too ambitious. > First off, Mathematica uses postscript for all its graphs anyway all that is left is the text interface, which given the mac version, probably does that anyway. (for everything but the equations) As to your second point, the author of Milo (an equation mangling program that runs on the mac) plans a Milo-Mathematica interface, thought the current version doesn't have it. (Milo is strong on interface and weak on math, Mathematica is the reverse.) Perhaps if Ron Avitzur (the author of Milo) or I get a NeXt box, we can port Milo to the Box. Or perhaps you could just write a Mathematica to TeX parser. In fact, I have a parser already written that would do the job nicely. Now all I need is $6500. (Actually Caltech price is probably $6500 + 3% + 6.5%=$7100.) Pierce
per@kps.UUCP (FES) (10/18/88)
In article <198@daitc.daitc.mil> jkrueger@daitc.daitc.mil.UUCP (Jonathan Krueger) writes: >In article <12834@oberon.USC.EDU> crum@lipari.usc.edu (Gary L. Crum) writes: >>a program to convert Mathematica output to >>publication-quality PostScript description. > >Why not to TeX instead? It's less device dependent, licenses need not >be obtained from Adobe, and it can be interchanged to a larger variety >of systems. And it makes use of an existing system of mathematical >typography; why reinvent the wheel? > >-- Jon Oh, come on! You must realise that even simple stuff like wheels have gone through some improvments over the years. And surely, text processing systems too. Face it, TeX *is* outdated since a few years. Per --- per`kps.UUCP ---
daveh@cbmvax.UUCP (Dave Haynie) (10/19/88)
in article <521@fabscal.UUCP>, dorn@fabscal.UUCP (Alan Dorn Hetzel) says: >>It seems that the NeXT machine may have a few problems: >> >>1) Outdated Processor Technology: NeXT just missed the wave of fast RISC >> processors. The 5 MIPS 68030 is completely out performed by the currently >> available RISC chips (Motorola, MIPS, Sparc) that run at approximately >> 20 VAX (they claim) MIPS. In a year or two, ECL versions of some of these >> RISC chips will be running at 40 to 50 MIPS. Priced the ~8 MIPS Sun 4 lately? Or the ~14 MIPS 88K chipset. How about an Apollo 10K? RISC machines are starting to get fast, and they're even starting to get down in price, but these two directions haven't met yet. >>6) Sun (I have heard) has sold 15,000 workstations to universitys. How many >> can NeXT expect to sell with its slow processor, non-standard bus/software, >> slow drive, and late software? Since the VAST majority of Suns sold to universities are Sun 3s (68020 based) and below (believe it or not, folks STILL use Sun 2s here and there), I don't think a 68030 based system, even NeXT's, which isn't an especially fast 68030 system (they're running it's memory at about 1/2 the possible speed), will have no trouble competing with the installed 68020 systems. Or a $25,000-$50,000 RISC based workstation. There certainly may be other reasons why a university would by a Sun or Apollo or even Apple rather than a NeXT, but I really don't think it'll be based on CPU speed. > I'm not sure there is any difference between their bus and the NuBus except > clock rate. It's also based on CMOS levels. > If that's the case, the 68000 is pretty tolerant of slower than expected devices, > so maybe slower boards will work. 680x0 CPUs can easily run asynchronously. The NuBus, however, is defined as being synchronous, with a 10MHz clock (with something like a 75/25 duty cycle). Even considering the fast clock alone, it would amazing if any existing NuBus board worked in this bus. It might not be much trouble to redesign existing NuBus boards to work in this bus, probably just change around some state machines to take into account the faster clock rate (probably does add more states, though). > If not, consider that IBM got away with introducing another bus, and now there > are cards for it. It may not even matter. It'll be quite some time before it makes any sense for 3rd party groups to build NeXTBus cards. But it may make lots of sense for NeXT to do so, and they've at least not locked themselves into something as slow as true NuBus. > Dorn > gatech.edu!fabscal!dorn -- Dave Haynie "The 32 Bit Guy" Commodore-Amiga "The Crew That Never Rests" {uunet|pyramid|rutgers}!cbmvax!daveh PLINK: D-DAVE H BIX: hazy "I can't relax, 'cause I'm a Boinger!"
daveh@cbmvax.UUCP (Dave Haynie) (10/19/88)
in article <1249@cfa.cfa.harvard.EDU>, ward@cfa.harvard.EDU (Steve Ward) says: > Summary: NeXT comments >> 2) Non-Standard NuBus Implementation: A small company like NeXT can't hope to >> create a competitive 3rd party board market for a non-standard bus. > I am unsure of the situation here, but I have heard that the "MAC" > NuBus is nonstandard and the NeXT implementation conforms fully to the > NuBus standard. This is an area that needs some clarification. Any > NuBus experts care to clarify this topic? I'm no NuBus expert, but I have read quite a bit on the topic. The most non-standard thing Apple did was change the form factor of the card from it's original Eurocard form (what's in the TI explorer, and probably in the NeXT) to a PC-ish form. I believe this form factor has now been adopted in some official capacity. The other thing they did was add a Non Mastered Interrupt line to each slot. Standard NuBus interrupts are weird; the board issuing the interrupt becomes the bus master and bangs on a chunk of memory that belongs to the board it wishes to interrupt. This does make sense in a system of multiple masters, but not so much on something like a Mac II, where the motherboard CPU is basically the default master (though there is no real default master on NuBus). The NeXT machine changes the NuBus clock from 10MHz to 25MHz, and the bus drivers from Bipolar to CMOS. That should be sufficient to render it incompatible with any existing NuBus cards. > Steven M. Ward ward@cfa.harvard.edu -- Dave Haynie "The 32 Bit Guy" Commodore-Amiga "The Crew That Never Rests" {uunet|pyramid|rutgers}!cbmvax!daveh PLINK: D-DAVE H BIX: hazy "I can't relax, 'cause I'm a Boinger!"
krish@jetsun.WEITEK.COM (Krishnan Sridhar) (10/19/88)
In article <5941@winchester.mips.COM>, mash@mips.COM (John Mashey) writes: > In article <5498@juniper.uucp> chari@juniper.UUCP (Christopher Michael Whatley) writes: > ... > >They are supposedly developing their own RISC chip that is compatible with the > >030. I don't know any more than that. I read this in a rumour column. > > This defies all logic. > a) If it's compatible with an 030, it's not a RISC. If it's a design > that makes the 68K architecture go faster (quite possible), it's still > CISC, but maybe it's called a 68040. Do you think that NeXT thinks they > can do that better than Moto? Not quite true. A RISC version of 68K can be made and with appropriate compiler, can be made fully compatibile with their ancestors. Frankly, RISC is not such a "narrow" concept as you make it appear. The concepts of RISC are fairly general and can be adapted to newer designs, still maintaining compatibility. Note, the compiler has to be a lot smarter, but given the fact that RISC itself requires a smart compiler to fully realize it's potentials, I don't think it's that big a deal. - krish
tim@crackle.amd.com (Tim Olson) (10/19/88)
In article <5024@cbmvax.UUCP> daveh@cbmvax.UUCP (Dave Haynie) writes: | >>1) Outdated Processor Technology: NeXT just missed the wave of fast RISC | >> processors. The 5 MIPS 68030 is completely out performed by the currently | >> available RISC chips (Motorola, MIPS, Sparc) that run at approximately | >> 20 VAX (they claim) MIPS. In a year or two, ECL versions of some of these | >> RISC chips will be running at 40 to 50 MIPS. | | Priced the ~8 MIPS Sun 4 lately? Or the ~14 MIPS 88K chipset. How about | an Apollo 10K? RISC machines are starting to get fast, and they're even | starting to get down in price, but these two directions haven't met yet. It seems to me that most RISC processors are less expensive then their CISC counterparts. The Am29000 pricing for 100 piece quantities is 16MHz $174 20MHz $230 25MHz $349 I'm sure that LSI Logic could also show you very low prices on their RISC chips. Last I heard, the 68030 was in the $300+ price range. -- Tim Olson Advanced Micro Devices (tim@crackle.amd.com)
die@cpoint.UUCP (David I. Emery) (10/19/88)
Most previously announced erasable optical media (such as Tandy's) allow only a limited number of write erase cycles before the media deteriorates enough to signficantly effect error rates. I beleive the Tandy product was qouted as only supporting around 10 write erase cycles. Does the NeXT/(Cannon?) media have any limit on the number of write erase cycles allowed ? It appears to be erased by thermal techniques (heating above the Curie point), how many heating/cooling cycles can the material undergo without changes in crystal structure that effect data SNR (annealing effects perhaps) ? Does this mean that the NeXT optical floppies have the possibility of wearing out in spots ? Doe the file system do anything to compensate. Can the disks be erased by modest high temperatures (such as leaving one in a hot car (160-180 F) or bright sun ? David I. Emery Clearpoint Research Corp. 99 South Street, Hopkinton Ma. 01748 1-508-435-2000 {decvax, cybvax0, mirror}!frog!cpoint!die -- David I. Emery Clearpoint Research Corp. 99 South Street, Hopkinton Ma. 01748 1-508-435-2000 {decvax, cybvax0, mirror}!frog!cpoint!die
firth@sei.cmu.edu (Robert Firth) (10/19/88)
In article <250@dataspan.UUCP> deraadt@dataspan.UUCP (Theo De Raadt) writes: >I puke trying to write assembly on RISC machines... I'd be interested to know whether that's a generally held view. In the course of porting a compiler and runtime to the MIPSCo M/500 RISC machine, I wrote a fair amount of assembler, and found it both easy and pleasant. Far less bothersome that, say, Vax or MC68020 assembler. Even the second pass reorganising the code was straightforward (this is adviseable since the assembler's own reorganiser is not always very astute). The most pleasant part was that I didn't have to keep computing instruction and address mode timings ("now, is it faster to do a clear or a move of zero? should I scale the index myself or use the scaled-index mode? do I buy anything making this address a CSE?" And so on)
gregg@ihlpb.ATT.COM (Wonderly) (10/19/88)
From article <3884@encore.UUCP>, by bzs@xenna (Barry Shein): > It does seem obvious that the main advantage of the removeable optical > disk in the University environment is with University purchased > clusters of machines. > > A student comes in with his/her disk, sits down at a machine (any > machine on campus), plugs it in and boots. Other software etc can be > gotten from NFS file servers which would mount on boot. Doesn't really > matter which NFS server, any one on the local net will do (a little > PARC-like clearinghouse software to broadcast for an appropriate > server would take care of that "will the NeXT binaries server please > mount itself on my /usr/local partition!") This brings up some interesting prospects. There are certain security issues, but these could probably be resolved by accessing a remote password file so that a student with a NeXT computer couldn't create a bogus password entry to read others' mail. A remote password file would also centralize the creation and maintainence of unique ID's. The comp center could then mass produce disks that contained the appropriate environment to boot and access the remote file(s) and file systems, and just sell them through the campus bookstore or whatever; very interesting. If there was no identification on a disk as to who the use was, then several students could share a disk, if they felt that the cost was prohibitive. The only real problem is that people who like to work at home would still need to have a NeXT machine, because a terminal (or PC with emulator) would no longer do the trick. Personally, I think it is better for the students to work together. One can learn a lot from setting around and listening to the guru's talk. -- Gregg Wonderly AT&T Bell Laboratories DOMAIN: gregg@ihlpb.att.com IH2D217 - (312) 979-2794 UUCP: att!ihlpb!gregg
rsexton@uceng.UC.EDU (robert sexton) (10/19/88)
While RISC may be cheaper(smaller design, less silicon) what you are really doing is shifting the cost burden onto the rest of the system. The high memory bandwidth of the RISC design means more high speed memory, bigger high-speed caches. With a CISC design, you put all of the high speed silicon on one chip, lowering the cost of all the support circuitry and memory. -- Robert Sexton, University of Cincinnati rsexton@uceng.uc.edu tut.cis.ohio-state.edu!uccba!uceng!rsexton Box Full O' Transputers... The Breakfast with MIPS I do not speak for UC, They don't speak for me.
guy@auspex.UUCP (Guy Harris) (10/19/88)
>I am getting sick and tired of this RISC/CISC battle. Come on guys, measure >RISC instructions against CISC microcode level instructions and they work out >to about the same thing. No, measure how fast a given application runs; that's what counts. Note also that I, or a compiler, can get at the RISC instructions, while neither I nor a compiler can generally get at a CISC machine's microinstructions (assuming it has microinstructions). Also note that the microinstruction sets for two different implementations of a given CISC architecture are often (usually?) not the same, which means if you *could* get at the microinstructions you would have to recompile (or rewrite) microprograms to get them to run on the new implementation. >As far as I can see, it appears that we are always battling memory >speeds. Not clear. >Not trying to start a flame war, but 030's are faster than Sun 4's. To which '030 machine, and to which Sun-4, are you referring? At the time the Sun-4/260 came out, no available '030 machine was faster because there weren't any '030 machines. What happens when the next Sun-4 comes out? Note Allen Baum's argument that RISC lets you get to market faster with a machine of a given speed, which means that by the time a CISC machine catches up you're on your way to a new faster implementation. Also, you might compare '030s against MIPS-based machines; are they faster than them, as well? >I puke trying to write assembly on RISC machines. Fortunately, I rarely had to do so, and these days fewer and fewer people writing applications have to do so. These days, "ease of writing assembly code" is less and less of a figure of merit for an instruction set. >I love CISC though, because you find your bottleneck, and just jump in >and start ripping some things out, and rewrite the routine.. You can do that with RISC as well.
paul@unisoft.UUCP (n) (10/19/88)
In article <5025@cbmvax.UUCP> daveh@cbmvax.UUCP (Dave Haynie) writes: >Interrupt line to each slot. Standard NuBus interrupts are weird; the board >issuing the interrupt becomes the bus master and bangs on a chunk of memory >that belongs to the board it wishes to interrupt. This does make sense in >a system of multiple masters, but not so much on something like a Mac II, >where the motherboard CPU is basically the default master (though there is >no real default master on NuBus). Actually the problem is more the amount of board real-estate required to make a board that is a master, you have to add a couple of extra state machines, arbitration logic, drivers for the address lines etc etc it was really a good idea to just provide a pin you yank low when you want to signal an interrupt ..... my minimal slave logic is normally around 6 chips ... >The NeXT machine changes the NuBus clock from 10MHz to 25MHz, and the bus >drivers from Bipolar to CMOS. That should be sufficient to render it >incompatible with any existing NuBus cards. yes I agree, if it weren't for this I think all my cards would run simply by replacing the PALs with faster parts (and with more internal wait states - isn't programmable logic great). Of course all the panel connectors would be unreachable ..... Since it sounds like the minimal NeXTBus add-in card is going to have to have 2 96 pin connectors it means that their sizes are going to be larger at the bottom end (my cards are half cards - gets the prices way down) plus they may be unmanageable if you put a tiny card in a deep slot. Add to this the need for bus-master logic (NeXT's $25 chip -> $100 to the consumer, I assume this chip contains the state machines but not the drivers so add $5-10 more), and the faster interface (faster parts). This will probably mean that the base prices of cards will be higher - also since the volumes will be low for quite a while, I too don't expect a mad rush of 3rd parties (from outside Higher Ed at least) designing hardware for it. Paul -- Paul Campbell, UniSoft Corp. 6121 Hollis, Emeryville, Ca ..ucbvax!unisoft!paul Nothing here represents the opinions of UniSoft or its employees (except me) "Gorbachev is returning to the heritage of the great Lenin" - Ronald Reagan 1988 (then the Wasington Post attacked RR [from the right] for being a Leninist)
daveh@cbmvax.UUCP (Dave Haynie) (10/20/88)
in article <23298@amdcad.AMD.COM>, tim@crackle.amd.com (Tim Olson) says: > | Priced the ~8 MIPS Sun 4 lately? Or the ~14 MIPS 88K chipset. How about > | an Apollo 10K? RISC machines are starting to get fast, and they're even > | starting to get down in price, but these two directions haven't met yet. > It seems to me that most RISC processors are less expensive then their > CISC counterparts. The Am29000 pricing for 100 piece quantities is > 16MHz $174 > 20MHz $230 > 25MHz $349 > I'm sure that LSI Logic could also show you very low prices on their > RISC chips. Last I heard, the 68030 was in the $300+ price range. Alot of it depends on quantity. I'm sure NeXT and Apple are buying their 68030s more that 100 at a time. Many of the ASIC houses making RISCs are output limited. And with most of the RISC designs, once you pay the additional cost of caches and MMUs, you're way out of the 68030 league, cost wise. Complete systems I've seen with both MIPS and 88k put you at around $1000 for the CPU subsystem. > -- Tim Olson > Advanced Micro Devices > (tim@crackle.amd.com) -- Dave Haynie "The 32 Bit Guy" Commodore-Amiga "The Crew That Never Rests" {uunet|pyramid|rutgers}!cbmvax!daveh PLINK: D-DAVE H BIX: hazy "I can't relax, 'cause I'm a Boinger!"
ronc@fai.UUCP (Ronald O. Christian) (10/20/88)
In article <5498@juniper.uucp> chari@juniper.UUCP (Christopher Michael Whatley) writes: >IBM has licensed NeXTStep for use with AIX on the PS/2 and RT PC. If that >becomes an alternative to OS/2 then the software, which supposedly could be >ported very easily, would reach a very wide market. A thought just occurred to me. Could this be why Microsoft is so p***ed off at Jobs? Ron -- Ronald O. Christian (Fujitsu America Inc., San Jose, Calif.) {amdahl, pyramid, sun, unisoft, uunet}!fai!ronc -or- ronc@fai.com Calling all Fujitsu Usenet sites! Contact fai!ronc or ronc@fai.com to establish uucp connection.
rogerk@mips.COM (Roger B.A. Klorese) (10/20/88)
In article <250@dataspan.UUCP> deraadt@dataspan.UUCP (Theo De Raadt) writes: >I am getting sick and tired of this RISC/CISC battle. Come on guys, measure >RISC instructions against CISC microcode level instructions and they work out >to about the same thing. As far as I can see, it appears that we are always >battling memory speeds. Not trying to start a flame war, but 030's are >faster than Sun 4's. Yes, but Sun 4's are not representative of leading-edge RISC design, or of efficient RISC architecture. >I puke trying to write assembly on RISC machines. Funny, while I don't recommend doing it (as our compilers can usually do at least as good a job as you can), I can show you testimonials to the ease of writing MIPS assembler. Remember, SPARC ~= RISC. -- Roger B.A. Klorese MIPS Computer Systems, Inc. {ames,decwrl,prls,pyramid}!mips!rogerk 928 E. Arques Ave. rogerk@mips.COM (rogerk%mips.COM@ames.arc.nasa.gov) Sunnyvale, CA 94086 I don't think we're in toto anymore, Kansas. +1 408 991-7802
whh@pbhya.PacBell.COM (Wilson Heydt) (10/20/88)
In article <Oct.17.00.51.55.1988.10300@athos.rutgers.edu>, mende@athos.rutgers.edu (Bob Mende Pie) writes: > > I would think the MS-DOS format for > floppys, no flames, is a standard (a number of them) that would allow file > transfer to those who only have a IBMish machine. I can imagine a box that > hangs off your SCSI port, that has a 5.25" 1.44Meg floppy on it as well as ^^^^^^^^^^^^^ Minor quibble--5.25" high density is 1.2 MB. 3.5" is 1.44 MB. > plugs to hang more floppys off of. Those boxes are already made. They're called "clones." The cost starts at $500. And since the Next machine has a couple of serial ports-- connection is already solved. --Hal ========================================================================= Hal Heydt | "Hafnium plus Holmium is Analyst, Pacific*Bell | one-point-five, I think." 415-645-7708 | --Dr. Jane Robinson {att,bellcore,sun,ames,pyramid}!pacbell!pbhya!whh
jmd@granite.dec.com (John Danskin) (10/21/88)
::In article <7399@aw.sei.cmu.edu> firth@bd.sei.cmu.edu (Robert Firth) writes: ::>In article <250@dataspan.UUCP> deraadt@dataspan.UUCP (Theo De Raadt) writes: ::> ::>>I puke trying to write assembly on RISC machines... ::> ::>I'd be interested to know whether that's a generally held view. ::>... I wrote a fair amount of assembler, and found it both ::>easy and pleasant. Far less bothersome that, say, Vax or MC68020 ::>assembler. :: I thought exactly the same thing. I have to write assembly code or microcode for lots of different machines, and I have had the easiest time with risc machines. Largely because of the regular architecture, easily determinable timing, and lack of squillions of features that you MUST use to achieve performance. -- John Danskin | jmd@decwrl.dec.com or decwrl!jmd DEC Advanced Technology Development | (415) 853-6724 100 Hamilton Avenue | My comments are my own. Palo Alto, CA 94306 | I do not speak for DEC.
brazil@pawl18.pawl.rpi.edu (Timothy E. Onders) (10/21/88)
The Cannon drive is not a true Optical drive. It is an optically enhanced magnetic storage device. Through the use of lasers, the tracks can be made as small as the width of the beam. The actual reading a writing is done with a magnetic field, much the same as it is done in a normal hard drive. So far, Tandy has been the only manufacturer to come close to a true erasable optical media. Theirs seems to be based on phase change media, which changes from clear to opaque and back depending on the power and wavelength of the laser beam it is exposed to. As to fears of high temperatures erasing the disk, rest assured that Cannon picked a material with a high enough curie point that you won't have to worry too much about that. Besides, even if it is heated up to the point, you would still need a magnetic field to damage the data. Although this is the first true market penetration for a read/write optical-type media, the actuall technology was in use as early as 3 or 4 years ago. Then, as now, the phase change medium seems more promising, since it allows for a lighter R/W head since it is not necessary to have a magnetic coil, as well as the optics, not to mention the fact that, as Tandy suggested, Phase change media could be read by present read-only devices, since the data is stored in much the same way as on a conventional CD. Timothy E. Onders brazil@pawl.rpi.edu
kds@blabla.intel.com (Ken Shoemaker) (10/22/88)
Of course, another obvious reason for not using any of the currently available RISCs (except for maybe the 29000) is system implementation costs. If you don't put a cache on a MIPS, the system won't run like a jackrabbit. With the SPARC you need to implement your own MMU. And the component costs of 88000s and Clippers with their custom MMU/cache chips are almost half the cost of the whole machine! And then we can start to talk about availability of production volumes of reasonably debugged parts... -------- If you break a law to prove a law, you're on pretty shakey moral grounds -- Ian Shoales Ken Shoemaker, Microprocessor Design, Intel Corp., Santa Clara, California uucp: ...{hplabs|decwrl|amdcad|qantel|pur-ee|hacgate|oliveb}!intelca!mipos3!kds
ns@cat.cmu.edu (Nicholas Spies) (10/22/88)
In article <1490@imagine.PAWL.RPI.EDU> brazil@pawl18.pawl.rpi.edu (Timothy E. Onders) writes: >The Cannon drive is not a true Optical drive. It is an optically enhanced >magnetic storage device. Through the use of lasers, the tracks can be made as >small as the width of the beam. The actual reading a writing is done with >a magnetic field, much the same as it is done in a normal hard drive. ... In fact, opti-magetic discs are written by raising the disc coating to its Curie point in a magetic field, which alters the polarization of light reflected off (and/or transmitted through?) the coating according to the direction of the magnetic field is was written under. Reading is done optically, not magnetically, by noting the changes of polarization. >a magnetic field, much the same as it is done in a normal hard drive. So >far, Tandy has been the only manufacturer to come close to a true erasable >optical media. Theirs seems to be based on phase change media, which changes >from clear to opaque and back depending on the power and wavelength of the >laser beam it is exposed to. ... >years ago. Then, as now, the phase change medium seems more promising, since >it allows for a lighter R/W head since it is not necessary to have a magnetic >coil, as well as the optics, not to mention the fact that, as Tandy suggested, >Phase change media could be read by present read-only devices, since the >data is stored in much the same way as on a conventional CD. Conventional videodiscs and CD's store data as a series of pits of varying length, of a constant depth of 1/4 wavelength deep, so that when they are illuminated by the reading laser, light reflected from the bottom of the pits is out of phase by 1/2 wave, yeilding destructive interference so the pits appear black (or light against a black background?). In any case, current laserdisc technology is reflective and does not depend on the relative opacity of the medium. -- Nicholas Spies ns@cat.cmu.edu.arpa Center for Design of Educational Computing Carnegie Mellon University
friedl@vsi.COM (Stephen J. Friedl) (10/23/88)
In article <1490@imagine.PAWL.RPI.EDU>, brazil@pawl18.pawl.rpi.edu (Timothy E. Onders) writes: > The Canon drive is not a true Optical drive. It is an optically enhanced > magnetic storage device. Through the use of lasers, the tracks can be made as > small as the width of the beam. The actual reading a writing is done with > a magnetic field, much the same as it is done in a normal hard drive. I'm not sure this is entirely correct. Writing is indeed done by a magnetic field, but reading is done by the laser. The following is taken from the latest _Byte_, entirely without permission. The _italics_ represented here are in the original article: "The optical platter is composed of the same clear rigid polycarbonate material that's used in CD-ROMs. Embedded within the platter is a layer of reflective aluminum backing that's overlaid with a magneto-optical substrate. The platter rotates inside the cartridge at 3000 revolutions per minute, 10 times the rotation speed of a CD-ROM, and almost as fast as a hard disk drive. "How does the magneto-optical drive work? A single laser performs both read and write operations. To write data to the disk, the drive first applies a magnetic field to the platter. The orientation of the magnetic field determines the data to be written to the platter -- either a 0 or a 1. The magnetic field is first oriented to write 0s at the start of what's called the _erase pass_. "The laser uses a high-power beam to heat a sector on the platter's substrate to its _Curie point_ -- the temperature at which the crystals in the substrate "forget" their previous orientation and reorient themselves to the surrounding magnetic field. All the data in the target sector is thus erased to 0s. "Next, a magnetic field is oriented to write 1s in the _write pass_, and at every spot in the sector where a bit must be set to a 1, the laser again heats the substrate to the Curie point. Finally, the sector is read in a _verify pass_ to check the accuracy of the data. "To read data off the platter, the drive removes the magnetic field, and the laser directs a low-intensity beam at the platter. The beam travels through the substrate and is reflected off the aluminum backing. However, in a phenomenon known as the Kerr effect, the crystal alignment in the magneto-optical substrate alters the polarization of the reflected beam. The amount of beam polarization determines its intensity as it passes through a polarizing filter to a photodetector. The beam intensity indicates whether a 1 or 0 was read at the spot on the platter." -------- The _Byte_ article (Nov 1988) is pretty detailed. Steve -- Steve Friedl V-Systems, Inc. +1 714 545 6442 3B2-kind-of-guy friedl@vsi.com {backbones}!vsi.com!friedl attmail!vsi!friedl ---------Nancy Reagan on the Three Stooges: "Just say Moe"---------
henry@utzoo.uucp (Henry Spencer) (10/23/88)
In article <5024@cbmvax.UUCP> daveh@cbmvax.UUCP (Dave Haynie) writes: >Since the VAST majority of Suns sold to universities are Sun 3s (68020 based) >and below (believe it or not, folks STILL use Sun 2s here and there), I don't >think a 68030 based system, even NeXT's, which isn't an especially fast 68030 >system (they're running it's memory at about 1/2 the possible speed), will have >no trouble competing with the installed 68020 systems... Your statement is correct as it reads (they will not have no trouble, i.e. they will have trouble), but that's probably not what you meant. The 68030 is *not* orders of magnitude faster than the 68020, especially if you are comparing fast 68020 systems with mediocre 68030 systems. (Do remember that you should probably normalize for memory access times, not raw clock rate, when comparing implementations.) It's not at all obvious to me that NeXT is going to blow Sun away. -- The meek can have the Earth; | Henry Spencer at U of Toronto Zoology the rest of us have other plans.|uunet!attcan!utzoo!henry henry@zoo.toronto.edu
henry@utzoo.uucp (Henry Spencer) (10/23/88)
In article <198@daitc.daitc.mil> jkrueger@daitc.daitc.mil.UUCP (Jonathan Krueger) writes: >>a program to convert Mathematica output to >>publication-quality PostScript description. > >Why not to TeX instead? It's less device dependent, licenses need not >be obtained from Adobe, and it can be interchanged to a larger variety >of systems. And it makes use of an existing system of mathematical >typography; why reinvent the wheel? Name three printers that accept TeX input. Like it or lump it, PostScript is clearly the de facto standard for a printer control language for sophisticated typesetting for the foreseeable future. TeX is Yet Another Old Batch-oriented Text Formatter. -- The meek can have the Earth; | Henry Spencer at U of Toronto Zoology the rest of us have other plans.|uunet!attcan!utzoo!henry henry@zoo.toronto.edu
jkrueger@daitc.daitc.mil (Jonathan Krueger) (10/24/88)
In article <1988Oct23.004659.20720@utzoo.uucp>, henry@utzoo (Henry Spencer) writes: Name three printers that accept TeX input. Name three processors that execute C source. Jonathan Krueger -- Point of light #344 If you see this point of light being operated in an unsafe manner: ...uunet!daitc!jkrueger jkrueger@daitc.daitc.mil (703) 998-4777
daveh@cbmvax.UUCP (Dave Haynie) (10/25/88)
in article <1988Oct23.002818.20634@utzoo.uucp>, henry@utzoo.uucp (Henry Spencer) says: > Xref: cbmvax comp.arch:7110 alt.next:228 > In article <5024@cbmvax.UUCP> daveh@cbmvax.UUCP (Dave Haynie) writes: >>no trouble competing with the installed 68020 systems... ^^ Whoops. > Your statement is correct as it reads (they will not have no trouble, i.e. > they will have trouble), but that's probably not what you meant. The 68030 > is *not* orders of magnitude faster than the 68020, A 68030 can run BETTER than TWICE the speed of a 68020+68851 MMU, if you let it. Sun 3s have a somewhat faster MMU, but no matter what you do, a 25MHz 68020 isn't going to be able to read a longword every two clock cycles. A 68030 using burst fills and both caches will usually preform about as fast with a few wait states as a 68020 with no wait states, running at the same speed. > especially if you are > comparing fast 68020 systems with mediocre 68030 systems. (Do remember that > you should probably normalize for memory access times, not raw clock rate, > when comparing implementations.) It's not at all obvious to me that NeXT > is going to blow Sun away. Of course, and I never implied that they would. I did imply that a NeXT machine is more than a match for a good portion of the Sun systems out there, all Sun 2s, Sun 3/50s, Sun 3/60s, and in fact, any Sun that's not based on a 25MHz 68020 with a no wait state MMU and no wait state RAM or a good chunk of external cache. Then you may have an even match. Of course, given an even match and even price, I'd pick the Sun. But hardware wise, they will compare favorably. > The meek can have the Earth; | Henry Spencer at U of Toronto Zoology > the rest of us have other plans.|uunet!attcan!utzoo!henry henry@zoo.toronto.edu -- Dave Haynie "The 32 Bit Guy" Commodore-Amiga "The Crew That Never Rests" {uunet|pyramid|rutgers}!cbmvax!daveh PLINK: D-DAVE H BIX: hazy Amiga -- It's not just a job, it's an obsession
larryh@tekgvs.GVS.TEK.COM (Larry Hutchinson) (10/26/88)
In article <896@vsi.COM> friedl@vsi.COM (Stephen J. Friedl) writes: > is reflected off the aluminum backing. However, in a > phenomenon known as the Kerr effect, the crystal alignment > in the magneto-optical substrate alters the polarization of > the reflected beam. The amount of beam polarization Does anyone know if this bit about the Kerr effect is true? Since the Kerr effect (pronounced car, I am told) is an electro-optical effect I don't see where the magnetic field comes into play (except to change the crystal domain direction). I keep wondering if someone is confusing Faraday rotation (which is magnetic field dependent) with Kerr. Anybody have any references? Larry Hutchinson, Tektronix, Inc. PO Box 500, MS 50-383, Beaverton, OR 97077 UUCP: [uunet|ucbvax|decvax|hplabs]!tektronix!tekgvs!larryh ARPA: larryh%tekgvs.TEK.COM@RELAY.CS.NET CSNet: larryh@tekgvs.TEK.COM
dorourke@polyslo.CalPoly.EDU (David M. O'Rourke) (10/27/88)
In article <4144@tekgvs.GVS.TEK.COM> larryh@tekgvs.GVS.TEK.COM (Larry Hutchinson) writes: >Does anyone know if this bit about the Kerr effect is true? Since the >Kerr effect (pronounced car, I am told) is an electro-optical effect >I don't see where the magnetic field comes into play (except to change >the crystal domain direction). I keep wondering if someone is confusing >Faraday rotation (which is magnetic field dependent) with Kerr. >Anybody have any references? IEEE Spectrum, Optical Disk become Writeable, 1988, Feburary I think. And the posters are correct, it is the Kerr effect. I don't understand all of the physics involved, but I can visulize the base effect. But I'll go out on a limb about {disclaimer coming: please no flames!!} "SOMETHING I KNOW VERY LITTLE" about. Could it be that Kerr's effect has it's roots somewhere in Faraday's law? Anyways the IEEE article is very enlightning. It explains exactly how the drive in the Next machine works, or at least a close relative of the drive in the NeXT machine. If more people would read the article then there wouldn't be so much confusion. -- David M. O'Rourke dorourke@polyslo.calpoly.edu "If it doesn't do Windows, then it's not a computer!!!" Disclaimer: I don't represent the school. All opinions are mine!
mark@cygnet.CYGNETSYSTEMS (Mark Quattrocchi) (11/02/88)
In article <199@daitc.daitc.mil> jkrueger@daitc.daitc.mil.UUCP (Jonathan Krueger) writes: >In article <452@oracle.UUCP> csimmons@oracle.UUCP (Charles Simmons) writes: >>Something to factor into your price performance picture: I believe another >>article pointed out that the optical drive was under $1500 dollars. So, >>for $1550, I get 256MB of disk storage. For $1600 I get 512MB of disk >>storage. For $1700 I get 1 Gigabyte of disk storage. Admittedly, my >>seek time gets a little bit long (2 to 3 seconds) when I have to swap disks... > >Jukebox. I hear a voice saying, "the NeXT jukebox...for when you need >the NeXT disk". I also hear a voice saying "backup" and also "our 400 >disk box stores 100 gigabytes." Opens up whole new horizons in >optimizing allocation of disk resources. Such things exist today, but >tend to be expensive and highly vendor-specific, and don't offer >rewrite/erase capabilities as far as I know. > Jukebox... Did I hear the magic word? Cygnet, the leader in optical disk jukeboxes does make them for both 5.25" and 12" drives. We support just about any brand of drives that exists (if not now we will shortly). Our 5.25" jukebox will fit under a desk and holds 25-28 cartridges depending on the drive type. The time to exchange cartridges is less than 4 seconds (the worlds fastest). We even have software that allows multiple jukebox management that runs under UNIX. So you too can have a NeXT jukebox, just send all your money and 10 corn flake box tops to ... me :-) Mark Quattrocchi