joel@gould9.UUCP (Joel West) (10/28/86)
As John Sculley himself confirmed, Apple is working hard to undo the
Steve Jobs' Macintosh-as-a-toaster no-slot mistake. (People don't spend
$2,500 for toasters; a CD-ROM/audio CD player at $400 would be a more
likely candidate for a mass-market home computer.) Some sort of
announcement in 1987 is expected, presumably before the non-competition
agreement with Jobs expires in summer '87.
The rumor column in 10/27 InfoWorld contains yet another claim that
the slotted Mac will use TI's NuBus. An observer I trust say that's
pure disinformation (presumably to catch leaks), that the product
is VME all the way.
I know the Sun-3 is VME; aren't most of the other 680xx boxes
also VME? To my knowledge, only LMI (Symbolics?) and TI use Nu.
Would anyone care to comment on the technical advantages to using
either bus? From a marketing standpoint, I would think VME would
offer a strong advantage.
--
Joel West MCI Mail: 282-8879
Western Software Technology, POB 2733, Vista, CA 92083
{cbosgd, ihnp4, pyramid, sdcsvax, ucla-cs} !gould9!joel
joel%gould9.uucp@NOSC.ARPAcch@sei.cmu.edu (Clifford Huff) (10/28/86)
Keywords: NuBus, VME, slotted Mac In regards to the question of the next generation Macintosh using VME or NuBus, you should find it interesting that between the IEEE standards committees for VME and NuBus, Apple is very active and only active on the NuBus (proposed IEEE-696 Std). This is very interesting! Apple has made suggestions to the IEEE-696 committee to make interfacing to NuBus an even easier task and has propose a second smaller microcomputer interface card. I believe that Apple is very concerned about making a bus decision which they will have to live with for many years. And when Apple looked at VME they saw too many long term problems that outweighed any short term marketing benefits of using VME. Apple sees in NuBus a much more straight forward bus implementation and with their recommendations a much easier bus to interface with by 3rd parties. I also believe that a recent issue of Computer Design reported that Apple has signed an agreement with TI concerning the NuBus. This would be a very strange move for Apple if they were heading in the VME path. -- Cliff Huff Software Engineering Institute Carnegie-Mellon University Pittsburgh, PA 15213 ARPRANET: cch@sei.cmu.edu
larryh@tekcbi.UUCP (10/28/86)
In article <842@gould9.UUCP> joel@gould9.UUCP (Joel West) writes: > >Would anyone care to comment on the technical advantages to using >either bus? From a marketing standpoint, I would think VME would >offer a strong advantage. And while we are at it, could someone comment on how to get a copy of the bus spec? I, for one, would be interrested in desiging a Nu bus to VMEbus converter/buffer.
cch@sei.cmu.edu (Clifford Huff) (10/29/86)
I made an earlier posting about the proposed IEEE standard for the NuBus
being IEEE-696. I was wrong (IEEE-696 is for the old S-100 bus).
I believe the correct IEEE number for the NuBus is IEEE-896.
Does anyone know for certain?
Sorry about the incorrect information...
--
Cliff Huff
Software Engineering Institute
Carnegie-Mellon University
Pittsburgh, PA 15213
ARPRANET: cch@sei.cmu.edugnu@hoptoad.uucp (John Gilmore) (10/29/86)
In article <842@gould9.UUCP>, joel@gould9.UUCP (Joel West) writes: > Would anyone care to comment on the technical advantages to using > either bus? From a marketing standpoint, I would think VME would > offer a strong advantage. The advantages to a company of using a standard bus are mainly these: * Your customers can plug in lots of cards to customize their system. * Your customers can benefit from new technologies faster because you (or they) can get new cards from third parties. Disks, disk controllers, networking cards, tapes, serial ports, etc are all easier to buy than to build, and the companies that specialize in building them often do a better job than a system manufacturer like Apple would. (Remember the horrible Apple hard disks? Remember the slow Mac floppies? The Mac network that only talks to itself? Third parties would go bankrupt if they tried to sell such stuff.) The Nu bus is not a standard, and has been adopted by few companies. The VMEbus is, and seems to be the "Multibus of the '80s" as far as peripheral card availability. By the way, it is a "VMEbus", and the cards that go in it are "VMEbus cards", not "VME cards". You don't talk about plugging Uni cards into your Unibus, do you? The other important characteristic of a system bus is that it not get in the way. But this is less important than the above. Things that you really need to be fast (e.g. main memory) can be put on a private bus. It's a real drag waiting a few years to be able to plug in a simple thing like a cheap or fast hard disk because nobody has built the controller for your system. Mac and Lisa owners (pre-mac plus) know about this. -- John Gilmore {sun,ptsfa,lll-crg,ihnp4}!hoptoad!gnu jgilmore@lll-crg.arpa Overheard at a funeral: "I know this may be an awkward time, but do you recall him ever mentioning source code?" -- Charles Addams
hunter@oakhill.UUCP (Hunter Scales) (10/30/86)
In article <367@aw.sei.cmu.edu.sei.cmu.edu> cch@cg.sei.cmu.edu.UUCP (PUT YOUR NAME HERE) writes: > >I believe the correct IEEE number for the NuBus is IEEE-896. > >Does anyone know for certain? IEEE P896 is the Futurebus standard. -- Motorola Semiconductor Inc. Hunter Scales Austin, Texas {ihnp4,seismo,ctvax,gatech}!ut-sally!oakhill!hunter (I am responsible for myself and my dog and no-one else)
johnt@hammer.TEK.COM (John Theus) (10/30/86)
In article <1240@hoptoad.uucp> gnu@hoptoad.uucp (John Gilmore) writes: > >The Nu bus is not a standard, and has been adopted by few companies. >The VMEbus is, and seems to be the "Multibus of the '80s" as far as >peripheral card availability. By the way, it is a "VMEbus", and the >cards that go in it are "VMEbus cards", not "VME cards". You don't >talk about plugging Uni cards into your Unibus, do you? And NuBus (IEEE 1196; IEEE 896 is Futurebus) is not Nu bus. I don't know your definition of standard, but as far as the IEEE is concerned neither is a standard, and both are awaiting approval by the Computer Society's Microprocessor Standards Committee (MSC). Multibus II is also at this stage, while Futurebus has been approved by the MSC. If your definition of standard is based on the number of companies using the bus, then yes, VMEbus is much more widely used. If your definition of standard is based on the number of options and incompatabilities allowed by a specification, then NuBus wins hands down. With NuBus you don't have the problems of differing address and data widths, and a free for all in the use of the address space. NuBus was developed at MIT. Western Digital went into the business of building boards for MIT and the NuMachine, and for LMI. TI bought this business from Western Digital, and uses NuBus in their AI machines. The principal players at WD, and then TI, now work at Corollary. George White of Corollary is the chairman of the IEEE 1196 committee. NuBus, along with Multibus II and Futurebus, can all be classified as modern, full feature buses, while VMEbus is a wide Multibus with an improved electrical environment. John Theus Futurebus Parallel Protocol Coordinator Tektronix, Inc.
lat@druil.UUCP (TepperL) (10/30/86)
In article <1240@hoptoad.uucp>, gnu@hoptoad.UUCP writes: > The advantages to a company of using a standard bus are mainly these: > > * Your customers can plug in lots of cards to customize their system. > > * Your customers can benefit from new technologies faster because you > (or they) can get new cards from third parties. Disks, disk controllers, > networking cards, tapes, serial ports, etc are all easier to buy than to > build, and the companies that specialize in building them often do a > better job than a system manufacturer like Apple would. (Remember the > horrible Apple hard disks? Remember the slow Mac floppies? The Mac > network that only talks to itself? Third parties would go bankrupt if > they tried to sell such stuff.) There's an additional advantage to a company making a system using a standard bus, and its an advantage that companies can relate to (although many haven't yet figured it out): People will *buy* a system using a standard bus. When the time comes to look at a peripheral, use of a standard bus means there will be different vendors to choose from. Different vendors means competition and that means lower prices. Look at the available prices for SCSI disks for the Mac and you'll get the drift. John's point about getting new technology faster is well taken. Just consider what might have happened had the original Mac been designed using the Multibus: instant Ethernet, instant hard disks, instant streaming tape. You could've put a Fujitsu Eagle on a Mac if you had the money. There have been 1024 x 1024 x 8-bit, full color image processing boards available for the Multibus for several years now. Actually, John didn't take it far enough. This isn't new technology, it's current technology that wasn't available because somebody thought it would be more FUN FOR THEM to design a new bus. It's my belief that a Mac with a Multibus would have buried the IBM PC, rather than giving it some stiff competition like the real Mac did. In addition, a whole lot of the criticism leveled at the Mac by the trade press would never have even occurred. Face up to it: a lot of corporate DP managers that buy PCs read the trade press. A Multibus with just 2 extra slots would have meant a lot and would have been worth paying for. But NOOOOOOOO! Instead, Apple locked people into a proprietary "bus" (if that's what one chooses to call it) when they couldn't deliver the goods: hard disks) for well over a year, a SECOND floppy for months. That hurt us and it hurt Apple. The NIH syndrome (Not Invented Here) has burned many a company. There's only one company in the world that can dare to try locking customers into its products, and that's IBM. I don't like it, but it's a FACT. An aside: From the little I know, it sounds like the VMEbus is a better bus (now) than the (older) Multibus is now. I used it in my discussion above because it was probably the best available choice when the Mac was designed. It would still serve well today for a small computer like the Mac or IBM PC. -- Larry Tepper {ihnp4 | allegra}!drutx!druil!lat +1-303-538-1759
munson@ernie.Berkeley.EDU (Ethan Munson) (10/31/86)
In article <796@oakhill.UUCP> hunter@oakhill.UUCP (Hunter Scales) writes: >In article <367@aw.sei.cmu.edu.sei.cmu.edu> cch@cg.sei.cmu.edu.UUCP (PUT YOUR NAME HERE) writes: >> >>I believe the correct IEEE number for the NuBus is IEEE-896. >> >>Does anyone know for certain? > > IEEE P896 is the Futurebus standard. >-- >Motorola Semiconductor Inc. Hunter Scales >Austin, Texas {ihnp4,seismo,ctvax,gatech}!ut-sally!oakhill!hunter > >(I am responsible for myself and my dog and no-one else) Here at Berkeley, I was able to read this message four separate times (this was the fourth). Is it the poster's fault, the net's, or our news managing software's? Ethan Munson munson@ernie.berkeley.edu
munson@ernie.Berkeley.EDU (Ethan Munson) (10/31/86)
In article <203@druil.UUCP> lat@druil.UUCP (TepperL) writes: > >But NOOOOOOOO! > >Instead, Apple locked people into a proprietary "bus" (if that's what >one chooses to call it) when they couldn't deliver the goods: hard >disks) for well over a year, a SECOND floppy for months. That hurt >us and it hurt Apple. The NIH syndrome (Not Invented Here) has >burned many a company. > >There's only one company in the world that can dare to try locking >customers into its products, and that's IBM. I don't like it, but >it's a FACT. > >-- >Larry Tepper {ihnp4 | allegra}!drutx!druil!lat +1-303-538-1759 While I think that Mr.Tepper's points about the value of having an accessible, "standard" bus are quite correct, I think it may be hasty to blame Apple overmuch for the decision. Apple made a machine with slots long before they made the Mac, the II series. It gave them incredible headaches because people would buy discount interface cards, or even premium quality cards and then discover that their favorite software wouldn't run (I for instance could not use both an ALS Dispatcher serial card and Softech UCSD p-System simultaneously, without significant patches to the p-System). Jobs apparently hated this problem and his decisions on the Mac reflect it. The Mac is supposed to be easy to use, complex interactions between cards, OS, and software are not a feature of an easy to use machine. So . . . the Mac has a complex set of requirements for how software is supposed to work that should make software independent of machine version and the machine is closed. The result has been difficulty penetrating the corporate market and years of poor hard disk performance (serial ports are slow). However, because almost all commercial software adheres to the standards set by Apple and because some of it is very good, it is very difficult to successfully market software that does not adhere to the standards. This might not have been the case if the Mac had been open from the start. Also, since it has been impossible to add an 8086 card, the Mac has had to viewed as standing apart from the Clone Crowd. Is this good? Is this bad? Ask me or John Sculley in five-ten years. Ethan Munson UCB CS Grad Student munson@ernie.berkeley.edu
joel@gould9.UUCP (Joel West) (11/01/86)
In article <203@druil.UUCP>, lat@druil.UUCP (TepperL) writes: > John's point about getting new technology faster is well taken. > Just consider what might have happened had the original Mac been > designed using the Multibus: instant Ethernet, instant hard disks, > instant streaming tape. You could've put a Fujitsu Eagle on a Mac > if you had the money. There have been 1024 x 1024 x 8-bit, full color > image processing boards available for the Multibus for several years > now. > > It's my belief that a Mac with a Multibus would have buried the IBM PC, > rather than giving it some stiff competition like the real Mac did. I won't argue the dis/advantages of the antiquated Multibus. But there is a good reason why Multibus Mac would not have done well against the IBM PC. Can you say 'too expensive'? Slots are not cheap; they require additional gold and power supply and cabinet. And I'm sure anyone in business will tell you that the cost of parts and assembly gets rolled up 4-10 times before it reaches dealer retail. More significantly, the cost and size of boards for Multibus are probably not suitable for the Mac. (Note the previous posting which pointed out that Apple has proposed microcomputer-sized NuBus cards.) It wouldn't be the same size, and it wouldn't be the same cost. I'd say offhand, the two things Apple could have included with the original Mac to improve sales were: 1. Provide socketed memory or empty sockets, with built-in jumpers for higher-density chips. 2. Provide an external port for high-speed peripherals, e.g. SCSI. thus allow hard disks for business use. They call this the Macintosh Plus, and it's outselling the 512e by about 3:1. A lot of us like to second guess major companies; it's easy to do, since most people on the net have never run one. My inclination would be to examine how a strategy could have been fine-tuned, since major companies will tend to make their big decisions based on ego (Jobs) or sheer weight of bureaucracy. -- Joel West MCI Mail: 282-8879 Western Software Technology, POB 2733, Vista, CA 92083 {cbosgd, ihnp4, pyramid, sdcsvax, ucla-cs} !gould9!joel joel%gould9.uucp@NOSC.ARPA
shebanow@ernie.Berkeley.EDU (Mike Shebanow) (11/02/86)
(sacrificial lamb line)
While I can't say one way or another what kind of bus the new Mac
will use, I will put in my own two cents about VMEbus:
I have been writing software for VMEbus based 68000/68020
systems for about three years now. All of this software has been
related in some form or another to image processing. After using
the VMEbus for such an extended period of time, I have two points
to make in favor of NuBus.
A: VMEbus is still too slow for many image processing
and other IO and/or memory intensive applications. Every
application I have worked on has been VMEbus limited.
On a 68020, adding memory on a VMEbus board exacts
a severe performance penalty (4 or more wait states).
Even using a fast memory bus like VMX32 doesn't give
you access to 0 wait state memory. The NuBus, which
is about 3x faster than VME, will make a large difference.
B: Existing VMEBus cards will not be cheaper than the
new cards that Apple and new third party developers
will bring out, regardless of which bus Apple uses.
Even simple IO cards for VMEbus cost at least $1000
currently (by simple I mean a 4x8 parallel board with
no onboard intelligence). Smart VMEbus boards like disk
controllers typically cost $2000 or more. For a personal
computer like Apple's, these prices are way out of line.
Although I would expect the introduction of a mass
market VMEbus computer to do wonders for VMEbus pricing,
I don't expect that the small companies who make VMEbus
boards today could compete with companies like Apple
and AST when it comes to cutting manufacturing costs.
I think the second point is the more important one: no matter
which "high performance" bus Apple chooses, they will be forced
to do everything from scratch. So why not pick the bus which
has the better bandwidth, so that the machine will be as good
as it can possibly be?
I vote for NuBus (or FutureBus).
Andrew Shebanow
bus wierdo at large.robinson@ecsvax.UUCP (Gerard Robinson) (11/04/86)
Its a hard call to make, but if the Mac had been open in the beginning, would it have retained its essence? Multibus boards were made by OEMs for system integrators, not end users. Multibus boards are large and potentially would intimidate the users, as well as make the systems larger (definitely not Mac). Granted, I was no friend or supporter of Apples until they chose to survive by removing Steve Jobs, but the original vision of the Mac as the "computer for the rest of the world" was an important one to establish. Now it can expand.
henry@utzoo.UUCP (Henry Spencer) (11/06/86)
> NuBus, along with Multibus II and Futurebus, can all be classified as > modern, full feature buses, while VMEbus is a wide Multibus with an improved > electrical environment. Personally, I don't *want* a "full feature" bus! I want something simple and fast which doesn't make me use half the board (or expensive single-sourced slow-delivery buggy semi-vaporware VVVLSI) for bus interfaces. It looks to me like the bus-design community is going through a delayed version of the processor-architecture community's evolution. The "full feature" buses are the Intel 432 of bus design. It's high time for a RISC bus. (I may be being a bit hard on the NuBus; as I recall, it's not as bad as the rest. Multibus II and Futurebus are definitely 432buses, and VMEbus avoids that distinction only if you ignore the more recent tailfins.) -- Henry Spencer @ U of Toronto Zoology {allegra,ihnp4,decvax,pyramid}!utzoo!henry
chuq@sun.uucp (Chuq Von Rospach; Lord of the OtherRealms) (11/06/86)
> Its a hard call to make, but if the Mac had been open in the beginning, would > it have retained its essence? Even more important: If the original Mac had been open, would anyone have bought it? remember that an open bus adds a LOT of complexity to a system, since you can't put everything on a single board anymore -- you have added costs for a larger power supply, a backplane, for FCC RF interference reductions, for all sorts of things. If the Mac had come out at $5,000 each, with 128K in it and no software, would YOU have bought it? Would all of the other people who did buy the Mac buy it at double the cost? If the Lisa is any indication, the answer it obvious...
news@cit-vax.Caltech.Edu (Usenet netnews) (11/07/86)
Organization : Calfornia Institute of Technology Keywords: From: woody@tybalt.caltech.edu (William Edward Woody) Path: tybalt.caltech.edu!woody In article <8980@sun.uucp> chuq@sun.uucp (Chuq Von Rospach; Lord of the OtherRealms) writes: >> Its a hard call to make, but if the Mac had been open in the beginning, would >> it have retained its essence? > >Even more important: If the original Mac had been open, would anyone have >bought it? remember that an open bus adds a LOT of complexity to a system, >since you can't put everything on a single board anymore -- you have added >costs for a larger power supply, a backplane, for FCC RF interference >reductions, for all sorts of things. > >If the Mac had come out at $5,000 each, with 128K in it and no software, >would YOU have bought it? Would all of the other people who did buy the Mac >buy it at double the cost? > >If the Lisa is any indication, the answer it obvious... If the Apple II, IIe, etc., is any indication, the answer is not so obvious. If you are willing to change the size and shape of the Mac, and do the hardware addressing _right_, you would have had a B/W Amega with a good operating system. But I like my cute little Macintosh. I hope the new Macintosh 2 (comming Real Soon Now...) doesn't become as large and klunky as an IBM AT... - William Woody Mac! > ][n && /|\ woody@romeo.caltech.edu
henkp@nikhefk.uucp (Henk Peek) (11/08/86)
In article <7288@utzoo.UUCP> henry@utzoo.UUCP (Henry Spencer) writes: >> NuBus, along with Multibus II and Futurebus, can all be classified as >> modern, full feature buses, while VMEbus is a wide Multibus with an improved > Where can I get a the current futurebus and NuBus "work" standards. I have only an old futurebus concept. Are there differences between the TI-NuBus and MIT-Nubus? henk peek henkp@nikhefk.UUCP seismo!mcvax!nikhefk!henkp.UUCP
henry@utzoo.UUCP (Henry Spencer) (11/11/86)
The genesis of the Mac's closed-box philosophy can be traced back to an
Apple internal paper, published a few years ago in one of the ACM SIG
publications, titled something like "Making a million computers a year".
A fascinating paper; the problems of producing computers in that kind of
volume are mind-boggling. (For example, if you keep a one-month inventory
of $2000 computers, that's $166 MILLION dollars tied up in inventory alone!)
Harbingers of the Mac are all through that paper. The closed-box approach
comes from two considerations: (1) the people who will buy computers at
that kind of production volume ("the rest of us") want one-piece turnkey
hardware that does not need assembly or configuring; (2) maintenance and
support for that many computers is a horrible nightmare unless the
configuration is utterly standardized.
Seen in this light, the shift to a slotted Mac is *not* the correction of
a technical mistake; the original technical reasoning was correct. The
mistake was in marketing, to wit the assumption of selling truly vast
numbers of computers to unsophisticated customers. Given more modest
sales to generally more knowledgeable customers, the closed-box philosophy
is less appropriate. (Actually, I think the closed-box Mac could still
have been a conspicuous success if it had had (a) more memory, (b) either
a built-in hard disk or a straightforward way of attaching fast mass-storage
peripherals [e.g. SCSI], and perhaps (c) memory-management hardware to blur
the limits of physical memory. Apple's biggest mistake was not the closed
box, but a closed box that didn't have quite enough inside it.)
--
Henry Spencer @ U of Toronto Zoology
{allegra,ihnp4,decvax,pyramid}!utzoo!henry