I stumbled acros these the other day and saved them, I'm sure you all know the content and theme but here are some anyway - yes they are years old.
I shouldn't think there would be a problem showing them here?
Excerpts from Dave Haynie's commodore postings on google newsroups
A300/600
The A300 was an on-going project, to build a C64-like Amiga. There was an expansion edge/port of some kind. GRR was messing around with a built-in,
dirt-cheap Genlock, but I don't know if it would have made the cut or not. I think the floppy was originally external, and other things.
When Sydnes took over, that very design was changed.
Now, they didn't instantly rename it, so chances are, the A300 name stuck on the PCB. Really doesn't mean anything -- I could put any name I like on a PCB (well, they did get after us to remote the thing on the backside of the A2000 motherboard, under the HAYNIE/FISHER bit, that said "The Few, The Proud, The Remaining... then everyone's initials, after the big '85-'86 layoffs).
Bottom line: the A300 was not an A500 replacement, it would have sold below the A500 in price, enough to keep the A500 around. The A600 was supposed to be $50-$60 cheaper than the A500, but it came in at about that much more expensive than the A500. Whoops.
Ok, really, there was nothing particularly wrong with the A600. The PCMCIA slot wasnt' a horrible idea, it made technical sense, given that things designed to go there were low-power. But in '91, it didn't make marketings sense -- they were all very expensive. The design wasn't flawed or anything, far as I know. It was just bad, considering the A500 was better regarded in the marketplace. They shouldn't have tried to make it the A500 replacement. GRR designed them both, they're both good designs, obviously based on specs written by two very different
administrations.
The PCMCIA slot was flawed - in their infinite wisdom, Commodore incorporated the slot before the standard was finalised. *Both the 600 and the 1200 have non-standard slots that don't work with the majority of expansion cards. *Now that the cards are really cheap, though, I've found the slots to be useful. *I've got a PCMCIA modem and a network card, and I can hot-plug any of my PCMCIA-capable Amigas into a LAN. The IDE port is also fantastic. *Yes, it may be a nasty implementation of IDE, but it does give the 600 access to standard 2.5" drives. *I can wander down to my local PC store and still buy hard drives for the 600, but there's no way I'll ever be able to get a hard drive for my 500.
I think the main problem with the 600 is that it was designed to be the replacement for a computer that didn't need replacing. *The 600 offered nothing for the average user (ie. games players) that the 500 couldn't do. *It may be a very different design, but from an end-user's perspective the 600 is a 500 in a smaller box.
As I (and most on this side of the pond) see it, the blame is clearly on Ali. *I don't know how hard he works. *But given the results, it would have been better if he didn't. *When he was brought in, he did his best to install his people throughout corporate and international management. *So if you have a problem with these people, the blame is on Ali, the guy who hired most of them.
I can tell you that his efforts in Engineering led directly to the demise of Commodore. *Before Ali brought in Sydnes, things were running very smoothly toward a release of AA systems in the spring of 1992, including the A1000+ ($1000 32-bit AA system) and A3000+ (an A3000 update with AA, DSP3210, and MC68040). *
There was an ultra low end A300 (low cost sub-A500), planned to complement the line, not replace the A500. *After Sydnes, who has hand picked by Ali despite his total lack of any qualifications for the position, was had AA purposely delayed by 6 month, the A4000 with its IDE bus rather than the A3000+, etc.
*This delay caused the big financial woes of 1993, since there weren't enough AA systems ready (you can't ramp up that fast, especially when chips like Lisa, Bridgette, Gayle, etc are made outside of Commodore and need to be ordered well in advance, which they weren't). *No one wanted the A600 (which we likened to the Amiga version of the PLUS/4 -- so wrong the engineers knew it), but it was forced out in place of the A500, and the A500 cancelled. *Ali did this. *And the A600 cost MORE than the A500, let us not forget. Sydnes wasted time building what we called the A1000jr (important symbolism here -- Sydnes was the PCjr guy at IBM), an ECS based slighly scaled down A3000. *This one was so wrong no country would order any.
Augi (former C-A HW engr who feels better now that he's told the world about a great machine that he put a lot of sweat into only for it to be scrapped for no rational reason at all. *It was scrapped because the new management had an * * *emotional investment in the casework for a PC that was cancelled! *They wanted to use their precious PC case,damn it, and the only way they could justify the slip in the Amiga schedule [the case needed major re-tooling for * * * *an amiga to fit in] was to scrap the two machines that were *well along in development, and declare that a new machine had to be started from scratch. *So you got the A4000, which inherited its price from the A3000+ and its IDE * * * from the A1000+.)
A4000CR
The A4000 CR motherboard was actually out before C= went under. However, the cost reductions were fairly trivial -- they put
the 68030 on the motherboard (strangely like the A3000), and rearranged the add-on PAL logic to use a cheaper PAL in one place, I
believe.
However, those PALs shouldn't have been there in the first place. They could easily have been pulled into a Gary chip built with the gate array technology of 1992, rather than the CSG gate arrays of 1990. Meanwhile, the very expensive A3640 card was not cost reduced, even though a design was nearly complete, and the development could have been paid for in two months of production (this would have also boosted performance, and supported '040s to 40MHz and '060s to 66MHz). The basic Amiga I/O glue, like floppy and the 8520s, should have been put into an ASIC years ago. There was all kind of stuff that could have been cost reduced, but wasn't.
The C64 underwent at least five redesigns between its introduction and 1986 or so. Each time it was updated to take advantage of cheap but modern logic not possible or practical earlier on. There was a production engineering group in Japan responsible for this on-going cost reduction program. That's what the Amiga should have had. It was never possible for the new product Engineering teams to solve all the cost problems, they had time-to-market constraints, as well as a generally uncooperative management when it came to even simple integration in the initial design.
>>Amiga cd?
Well, it works. *It's not a full replacement for SCSI. *You don't get DMA driven IDE, it's programmed I/O and eats all kinds of bandwidth. *Disk speed
won't suffer much, system speed will if you had something else for the CPU to be doing at the time. *Bottom line though is cheapness; IDE adds little
cost to a system if implemented right, with SCSI you have to pay for a SCSI chip, and DMA if you want SCSI to outperform IDE.
IDE has lots of problems with standards, mainly in the area of how fast a disk will go. *It is, after all, a hack, one of these after-the-fact standards (of course, most things are, that's where SCSI came from too).
>Perhaps when SCSI finally reaches the standards' stage...
It's standard enough to be useful, far more so than lots of other really useful things. *If you wait for a 100% bulletproof standard, it'll be obselete
by the time its out.
>At the moment it most certainly makes no sense for a home computer (I have >SCSI on all mine and I see very liitle advantage over the IDE I had on a PC.
If you have only one drive, than the only real advantage to SCSI is the higher throughput you'll get with a DMA-driven SCSI implementation. *At home,
I have three drives, so I'm already beyond an IDE system (but I'm not typical).That's on an A3000; I don't think the bulk of A600 owners will know the
difference, much less care. *When I was 16, a hard disk of any kind on my home system would have been plenty. *Hell, affordable floppies would have been an
amazing thing...
>In fact SCSI on a PC leads to a lot of nightmares unless you >know what you are doing and/or have a MC/EISA PC).
Well, don't blame SCSI for PC problems. *The PCs work fine with hardware they they know about, and horribly with most everything else.
>I'm really getting annoyed at all those in this group who seem to think home >users needs 486's, 040's, 030's,etc...with scanners, SCSI,etc...Come down to reality.
Most of the folks in this group want that kind of stuff at their home. *But they're not typical home users, they're for the most part already computer
people. *A computer person knows what to do with a computer and can use up a good portions of a really expensive system. *You wouldn't expect sports car
designers to drive Ford Escorts, or movie makers to settle for 19" TV screens and cheap-ass VHS at home either. *The Masses, on the other hand, spend most
of their money on the cheap stuff, which they don't understand all that well in the first place.
>My NeXT and Apple/Mac computers use SCSI. It's no big deal.
Well, certainly for the NeXT it is. *You're not going to be happy with paging to an IDE drive on a UNIX system, everything else has to shut down and wait
for the hard drive. *The Mac SCSI implementations are, indeed, no big deal;they'd generally do as well or better, performance-wise, with IDE.
>CD-ROM's do not need SCSI?.
But they need some kind of interface. *Unless you have PC bus and can plug in a card that implements a custom device, or glomp on a box like the CDTV-thing
for A500, SCSI is your only option for CD-ROM. *Tape backup can work over a floppy port if you're willing to wait. *Scanners need a high speed interface,
IEEE-488 is a popular alternative to SCSI, but again, if you had SCSI, you wouldn't need to buy another interface card (then again, you could run the
hard disk over IEEE-488 if you wanted to, like PETs and our old HP development machines did). *A600 owners don't need any of this. *Most higher end systems
do. *We have assorted scanners around here, along with laser disks, ultra high density floppies, fast tape drives, Syquests, etc. *Most of which come
only in SCSI or custom interface varities.
Dave Haynie * * * | Chief Toady, Frog Pond Media Consulting 2002
>>Amiga like performance in 1985 from the PC?
It didn't exist.
And the claim doesn't have any place in today's market, anyway. Look out there.. see any simplicity? I see Pentium 4s with 125 million transistors, nVidia and ATi chips with 70-something pipleline stages and a similar number of chips. We're not just talkin' "more complex than the Amiga", we're talking "more complex by many orders of magnitude" than the Amiga.
The problems were elsewhere, and with 20/20 hindsight and an actual involvement in this stuff, I can tell you precisely where the problems were. On the chip front, it wasn't the complexity of the chips, not at all. It was the simple fact that Commodore management didn't spend the necessary R&D money to maintain development of the custom chips in any useful way. Look at modern tech companies, like nVidia, ATi, Intel, or AMD. See how much money they plow back into R&D. Had Commodore done this at similar levels, there would have no problem delivering new, interesting chipsets.
Second problem was software: the details of the custom chips should never have been revealed to programmers. This would have removed any need to maintain binary compatibility. You don't worry about this today, of course, because one ATi chip may or may not be anything like the next, but no one cares, your drivers account for any differences. That is the right way to do it, and it was well established as the right way to do it before the Amiga ever shipped. That's the one area of the whole OS that should have been done differently, before 1.0 shipped, IMHO.
>- First, the basic A1000 hardware should have been given a boost with extra RAM and a hard drive as soon as the cost of this non-proprietary hardware fell.
The big problem here: with about 150,000 or so Amiga 1000s in the world, only some of those owners buying memory upgrades, and five or
six companies offering them, there was no way the prices could have fallen. Not only that, but memory was wicked expensive in the
mid-to-late 80s. DRAM was in a worldwide supply crunch, and the prices reflected it. While I don't disagree with the "it would have been
nice" factor, it also wasn't even remotely possible.
>The custom graphic chips are often described as an evolutionary dead end. *In fact, they had considerable room for improvement while retaining a very high level of backward compatibility, as AA showed.
Well, AA's compatibility was good, sure, but it was very much evolutionary, and that showed. AA would have been great in 1988, ok in
1990, but in 1992, it was an also-ran, capabilities-wise. That goes back to budget.. C= only had so much money to spend on improvements
(especially considering the huge salaries the top management got -- keep in mind, at this time, none of it was performance-keyed, and
folks like Gould and eventually, Ali were making substantially more than the top dogs at Apple, IBM, Compaq, etc.
AAA, on the other hand, spent huge bits of technology on being compatible. Much of that was totally impossible anyway, at least in some configurations. For example, the 64-bit setup (one Andrea, one Mary, two Lindas, and two Monicas.. I found this: http://amiga.emugaming.com/amigaaaa.html with some of my drawings and a shot from "The Deathbed Vigil" on it... seems correct, though they missed the fact that AAA was actually in 1988, long before AA began).
>Jay Miner's team had already finished the ECS upgrade in early 1987.
Much of ECS was actually done in West Chester. Some of the delay was the simple fact that Amiga Los Gatos wasn't using modern CAD tools.
They had to capture, simulate, and verify all the old stuff before anything new could take place. The other problem was more serious: the
original Amiga chips, and ECS, were done in Commodore's 1.5 micron high-speed NMOS process. They were very much pushing the limits on
what that process could achieve. Moving to CMOS was critical, but keep in mind -- these were ALL transistor-level designs, not the gate-level
stuff people do today. Basically, you don't take an NMOS chip and redesign it into a CMOS chip, you start over from scratch.
>Commodore's clever leaders then sacked the original Amiga team and managed to lose the ECS blue prints.
None of that happened. Again, ECS was done in West Chester, with some consulting with the original team. For that matter, it was the second
major Amiga project done there, since Fat Agnus was designed in West Chester (Bob Raible did the main architecture, Victor Andrade did the
chip, as I recall).
>> ECS saw the light of day in 1990 only after prototype chips were reverse engineered.
Also not true. There was no reverse engineering done. ECS chips were running well before 1990, and they were phased into the A2000
production as a running change.
>Had Miner et al been retained with a modest R&D budget (which was instead ****ed away on second rate PC clones) I don't think it requires a huge imaginative leap to envision AA in 1988/9.
Well, you're right about the budget in general. However, very little was spent on the PC clones development. The point of that was to allow C= to be a "one-stop shopping" supplier. One big problem Apple had in Europe was their lack of a PC -- many large companies, in the day, wanted to deal with a single vendor for all of their PC needs. If they wanted PCs and Macs, and you didn't make both, they'd have to
choose... and it never went against the PC. This is also what drove the Amiga UNIX project. PC development was done in-house, but it was primarily systems work. The chips stuff they did was very simple gate-array development (which means, systems engineers like me, Greg Berlin, George Robbins, etc.... only those in the PC group. Not chip designers) for glue logic. By the 1990s, they were outsourcing the
PCs, much like everyone else (Dell, Compaq, Gateway, etc).
>Denise is the only thing that changes?.
Incorrect. Denise did pretty much everything Denise could ever do. That's exactly why the Lisa chip in the AA set was the only totally-new chip in AA. To go much beyond Denise (say, to add just the simple upgrade from 16 to 256 LUT registers), CMOS was the only possible choice. It couldn't have been a simple add-on to Denise. Of course, even if it was, you would have had to tweak Agnus to deliver an increased bandwidth (memory access in AA is 4x faster than in ECS or OCS, if you didn't know).
>With something like Dave Haynie's A3000+ in 1989?
Certainly, if C= management had treated Engineering with the same kind of investment that, well, pretty much everyone else in high-tech done,things could have happened differently. We never would have had AA in 1989, but with proper investment, maybe AAA (which began in 1988, but probably could have started a year or two earlier, with the right backing). That certainly wouldn't have sucked.
>I think Commodore would have found enough breathing room to transition to PowerPC/retargetable graphic cards, just as Apple did.
There's no telling. For one, it's not an instant assumption that we would have gone to PowerPC. That was probably the right decision in 1990-1991, sure. Much before then, MIPS would have been the obvious choice; too much beyond that, and x86 woudl have been the only reasonable choice. The system architecture I designed in 1991, which was intended to be the transitional machine (moving from dedicated
AA/AAA chips to Hombre and commodity RTG + PCI chips, moving from 68K to "something else") was CPU independent.
>The fact that they lasted until 1994 making the choices they did shows how brilliant the 1985 design was, and I think you do it a great disservice.
Guess I missed that part.
One thing I did, because, well, I could do it (leading high-end Amiga development had its advantages, and pretty much anything that Jeff Porter, me, and sometimes Greg Berlin agreed on would fly, at least in the days before Ali & Sydnes decided to kill the company with their incompetance and micromanagement), was to ensure each Amiga system was long lived. So we had the CPU slots (which Apple eventually copied, but the x86 world never did, even though IBM actually published a very good paper on this in the early 90s), video slots, expansion, etc. I absolutely knew that the C= way (and fortunately, I learned that way in the TED and C128 days, so as not to have to experiment on the Amiga) was never going to get us a new model every year. Of course, that only goes so far. My goal was to make your machines useful for at least five years. I didn't count on 10-15 being necessary
Dave Haynie * * * | Chief Toady, Frog Pond Media Consulting dhay...@jersey.net| Take Back Freedom! Bush no more in 2004!
>>Should Commodore of bought Sinclair as a cheap better computer than the C64?
"Better" is a relative term. Some people really just do like the C64 or the Apple ][ better than any other computer. That really is their business. I have personally found that a ZX81 makes the best frickin' doorstop of any computer known to man, though I haven't necessarily fired up a ZX81 for its computational use, or tried every possible computer as a means of holding my door ajar.
>> But I just like to say that the 68000 is 16 bit, I know it 32 *bit internally but its classification is 16 bit.
Well, there are three reasonable "bit-ness" viewpoints.
>I wish people would either stick to the same rule for all CPUs which is to informally designate them by their external data bus width only, or they use the form: max general purpose data register width / external data bus width.
No, not really. For one, no one really cares all that much about the external bus width, unless they're a hardware guy like me actually hooking the sucker up. That is one valid classification (external data bus width in hardware), but it's really just of interest to HW people. To everyone else, it's just an implementation detail and perhaps a performance optimization (eg, a 32-bit bus probably goes faster than a 16-bit bus, if the chip inside can actually use data that fast).
Anyway, the second legit via is the programmer's model. The 68000, for all intents and purposes, is a 32-bit chip to the programmer. The 68020 is certainly a slightly better and, sometimes, considerably better 32-bit chip, but that's the main difference.
Then there's the CPU designer's view. If you call the 68000 a 16-bit chip, it really should be because the chip only has 16-bit data paths and 16-bit ALUs (three of them), not because it has a 16-bit bus interface. Same reason the Z-80A is an 8-bit chip, even though it can do a few 16-bit things, too -- internal functions happen 8-bits at a time.
As for address bus width, not salient to the discussion of chip "size". There are 32-bit CPUs with 12-bit addresses, or 36+ bit addresses. I'm using an 8-bit and a 16-bit CPU in a design, currently, neither of which expose either an address or data bus to the outside world. The address bus says more about the application for the particular instance of a particular chip than it does anything about the CPU core, programming, etc.
>>Bill Gates is one of the reasons for why Amiga died 8 years ago?.
Not really. Microsoft was powerful in the 80s, but not invincible. Commodore was killed off by software pirates on the outside, and gross
mismanagement on the inside, before Gates & Microsoft even became a factor.
>Why not have a day of sorrow/remembering, on the day when its 10
> years since the bloodbath of CBM. (just an idea)
Which bloodbath? The "Last Big Layoff?" You'll have to wait until April 27th, 2004 for the day of the big layoff (it wasn't THAT huge in
terms of people, but it was more than half of the remaining staff, so percentage wise, it was the largest). Or April 30th, 2004, for the day
of the actual bankruptcy filing.
>>has anyone tried to implement NTFS for the Amiga? *
I rather doubt it. There are religious issues
>> : Could the Amiga *handle such a file system?
Sure.
>> There is no reason why it couldn't, but why not run a native FS like PFS3 (commercial) or SFS (free)?
What makes those any more "native" than NTFS? They weren't written by any of the various Amiga companies. PFS began life as a sort of ill-conceived UNIX file system for the Amiga. It had a few goods ideas, but some really bad ones as well, which made it far worse than a properly tuned FFS. It wasn't even remotely recoverable, and it wasn't remotely error-free enough for that not to be an issue.
Supposedly new versions are better, but you'd need substantial changes to make it a good file system, structurally. And it's limited to small files (<4GB). And how about attributes? For the Amiga to stop disappearing, it needs to get modern.
>> I can readily recommend SFS, it makes my life much easier.
SFS claims to offer true journalling, which is a big improvement. Though I still would demand the recoverability -- sometimes you want deleted files back. And do you want to bet on anything being bug free, or have a fail-safe. Also, no attributes, no large files (<4GB). And on the web site, they claim there are problems with large blocks on FFS, but don't actually point them out. Having used large blocks for
over 10 years without a single "other problem", I would like to know what they're talking about.
> * Well, I thought it might kind of neat to install removeable hard drive cages in my Amiga and PC, and be able to swap hard drives between them in this say. *NTFS is more robust than FAT32 so I thought that might be a better choice, if it was available to me!
Surprisingly, NTFS is an excellent file system. It supports 64-bit file sizes, and it has the curious aspect of everything in the file system being a named attribute. So it's absolutely trivial to add Amiga-like features, such as Amiga dates, protections, etc. and still maintain full compatibility with the simpler schemes used in NT. FAT16 made sense, since it is so commonly used on floppies and memory cards. At present, FAT32 would be pointless, other than for similar data exchange.
>> August 1985 vs. September 1985. No one at Amiga or C= imagined for a minute that the Amiga was supposed to be a C64/C128 followup. It was to the C64 very much what the Mac was to the Apple ][... the
>> future. As one of the C64 revivalists/nostalgics I have to agree. In the 80s I didn't expect a new computer to be compatible with an old one from the same company.
And, in fact, they never were. Well, there was a level of compatibility between some of the CBM machines, but not 100%. The C64
didn't try to run VIC-20 programs. Atari upgrades were hit and miss. Apple did a little better with the Apple ][ series, but they were also
less aggressive on improvements.
The C128, of course, was compatible -- actually, with both the C64 and the old Kaypro CP/M machines to a large extent, if you want to get technical. But this was also a different era, or at least the beginning of one. While we were developing the C128, we saw where other systems were going, with a real OS and all. Hell, we even sort of had one, with GEOS on the C64 (and eventually, 128).
>I started out with the VIC20 in 1982. When me and my brother could afford a C64 in 1984 we sold all our VIC stuff without a second thought. Why keep an obsolete computer when the future was here?
If you bought new, it was expensive, but eventually, it wasn't a big price hit to upgrade, compared to Amigas, much less Macs or PCs in the
day. C= was also fairly good at keeping the peripherals compatible. Until the Amiga, there had only been the IEEE peripherals and the serial peripherals (the C= serial bus, of course, was basically the IEEE bus implemented in bit-banging serial).
>Same thing when the Amiga became affordable with the 500 in 1987. We never even considered the possiblity that it could have been compatible with the 64. We wanted an Amiga so we sold our C64, 1541
>drive, MPS 801 printer and all the disks. Why would we want to hang on to all that old stuff when we had an Amiga?
Right. And there really was no precedence, in those days, of compatibility with upgrades, at least not in the consumer market. So there was no expectation. We're all used to compatible upgrades now (even occasionally in video games!), but that was a product of a more complex design.
>To answer the original question: the reasons for me starting to play with C64s and 128s and not as much with Amigas (though I have a few of them too) are that they have a greater sentimental value to me,
>they're easier to find at flea markets, there is (it seems) a livelier community around them, and (this is subjective) I find them more accessible.
I think these systems, the C64 and C128, were the last of the sort of computer you could totally master yourself, and that is part of their allure. I started on computers in 1973... first on a programmable HP calculator (with core memory and a CRT). That was easy -- I learned everything there was to learn in a week or so. So I got my Dad to hook me up with a computer at Bell Labs, which turned out to be a fairly
serious scientific computer, but what did I know. This one was way too complex, and the information was hard to come by.
Some years later, my best friend bought a PET 2001, and man did I go to town on that one. That was a computer that I could program out the wazoo, no worries about 300 baud, thermal paper, etc. And you could learn each and every secret about the machine, only, not in a week. So it was challenging. That was the case with most of the systems of those days. I used others; I still like the C= best (and not just because my name's hidden inside one of 'em...).
>Something to do with not having to wait for the computer to boot from disk and not having to use a mouse and a GUI (my Amiga 500s actually feel slower than my C128s and C64s), I think.
And NO ONE knows everything about an Amiga. Otherwise, I never would have had a chance becoming something of an expert on a piece or two of
it. It was cool, for users, because all thse "stuff" extended what programmers could deliver, but it was far harder for anyone to get The
Big Picture.
Dave Haynie * * * | Chief Toady, Frog Pond Media Consulting dhay...@jersey.net| Take Back Freedom! Bush no more in 2004! "Deathbed Vigil" now on DVD! See http://www.frogpondmedia.com
>Hi Dave, *I have been away for a few years now and haven't been able to >follow the Amiga scene. *Still holding annual digs at your place?
I have been... not sure about the timing this year, but I'm thinking
of doing it again. Wouldn't be until late August or September.
> Where is the old crowd hanging out online now?
There's nothing really central for the C= crowd. I'm on the Team Amiga
list, which has a number of people from the old days, mostly talking
about things unrelated to the Amiga these days.
I shouldn't think there would be a problem showing them here?
Excerpts from Dave Haynie's commodore postings on google newsroups
A300/600
The A300 was an on-going project, to build a C64-like Amiga. There was an expansion edge/port of some kind. GRR was messing around with a built-in,
dirt-cheap Genlock, but I don't know if it would have made the cut or not. I think the floppy was originally external, and other things.
When Sydnes took over, that very design was changed.
Now, they didn't instantly rename it, so chances are, the A300 name stuck on the PCB. Really doesn't mean anything -- I could put any name I like on a PCB (well, they did get after us to remote the thing on the backside of the A2000 motherboard, under the HAYNIE/FISHER bit, that said "The Few, The Proud, The Remaining... then everyone's initials, after the big '85-'86 layoffs).
Bottom line: the A300 was not an A500 replacement, it would have sold below the A500 in price, enough to keep the A500 around. The A600 was supposed to be $50-$60 cheaper than the A500, but it came in at about that much more expensive than the A500. Whoops.
Ok, really, there was nothing particularly wrong with the A600. The PCMCIA slot wasnt' a horrible idea, it made technical sense, given that things designed to go there were low-power. But in '91, it didn't make marketings sense -- they were all very expensive. The design wasn't flawed or anything, far as I know. It was just bad, considering the A500 was better regarded in the marketplace. They shouldn't have tried to make it the A500 replacement. GRR designed them both, they're both good designs, obviously based on specs written by two very different
administrations.
The PCMCIA slot was flawed - in their infinite wisdom, Commodore incorporated the slot before the standard was finalised. *Both the 600 and the 1200 have non-standard slots that don't work with the majority of expansion cards. *Now that the cards are really cheap, though, I've found the slots to be useful. *I've got a PCMCIA modem and a network card, and I can hot-plug any of my PCMCIA-capable Amigas into a LAN. The IDE port is also fantastic. *Yes, it may be a nasty implementation of IDE, but it does give the 600 access to standard 2.5" drives. *I can wander down to my local PC store and still buy hard drives for the 600, but there's no way I'll ever be able to get a hard drive for my 500.
I think the main problem with the 600 is that it was designed to be the replacement for a computer that didn't need replacing. *The 600 offered nothing for the average user (ie. games players) that the 500 couldn't do. *It may be a very different design, but from an end-user's perspective the 600 is a 500 in a smaller box.
As I (and most on this side of the pond) see it, the blame is clearly on Ali. *I don't know how hard he works. *But given the results, it would have been better if he didn't. *When he was brought in, he did his best to install his people throughout corporate and international management. *So if you have a problem with these people, the blame is on Ali, the guy who hired most of them.
I can tell you that his efforts in Engineering led directly to the demise of Commodore. *Before Ali brought in Sydnes, things were running very smoothly toward a release of AA systems in the spring of 1992, including the A1000+ ($1000 32-bit AA system) and A3000+ (an A3000 update with AA, DSP3210, and MC68040). *
There was an ultra low end A300 (low cost sub-A500), planned to complement the line, not replace the A500. *After Sydnes, who has hand picked by Ali despite his total lack of any qualifications for the position, was had AA purposely delayed by 6 month, the A4000 with its IDE bus rather than the A3000+, etc.
*This delay caused the big financial woes of 1993, since there weren't enough AA systems ready (you can't ramp up that fast, especially when chips like Lisa, Bridgette, Gayle, etc are made outside of Commodore and need to be ordered well in advance, which they weren't). *No one wanted the A600 (which we likened to the Amiga version of the PLUS/4 -- so wrong the engineers knew it), but it was forced out in place of the A500, and the A500 cancelled. *Ali did this. *And the A600 cost MORE than the A500, let us not forget. Sydnes wasted time building what we called the A1000jr (important symbolism here -- Sydnes was the PCjr guy at IBM), an ECS based slighly scaled down A3000. *This one was so wrong no country would order any.
Augi (former C-A HW engr who feels better now that he's told the world about a great machine that he put a lot of sweat into only for it to be scrapped for no rational reason at all. *It was scrapped because the new management had an * * *emotional investment in the casework for a PC that was cancelled! *They wanted to use their precious PC case,damn it, and the only way they could justify the slip in the Amiga schedule [the case needed major re-tooling for * * * *an amiga to fit in] was to scrap the two machines that were *well along in development, and declare that a new machine had to be started from scratch. *So you got the A4000, which inherited its price from the A3000+ and its IDE * * * from the A1000+.)
A4000CR
The A4000 CR motherboard was actually out before C= went under. However, the cost reductions were fairly trivial -- they put
the 68030 on the motherboard (strangely like the A3000), and rearranged the add-on PAL logic to use a cheaper PAL in one place, I
believe.
However, those PALs shouldn't have been there in the first place. They could easily have been pulled into a Gary chip built with the gate array technology of 1992, rather than the CSG gate arrays of 1990. Meanwhile, the very expensive A3640 card was not cost reduced, even though a design was nearly complete, and the development could have been paid for in two months of production (this would have also boosted performance, and supported '040s to 40MHz and '060s to 66MHz). The basic Amiga I/O glue, like floppy and the 8520s, should have been put into an ASIC years ago. There was all kind of stuff that could have been cost reduced, but wasn't.
The C64 underwent at least five redesigns between its introduction and 1986 or so. Each time it was updated to take advantage of cheap but modern logic not possible or practical earlier on. There was a production engineering group in Japan responsible for this on-going cost reduction program. That's what the Amiga should have had. It was never possible for the new product Engineering teams to solve all the cost problems, they had time-to-market constraints, as well as a generally uncooperative management when it came to even simple integration in the initial design.
>>Amiga cd?
Well, it works. *It's not a full replacement for SCSI. *You don't get DMA driven IDE, it's programmed I/O and eats all kinds of bandwidth. *Disk speed
won't suffer much, system speed will if you had something else for the CPU to be doing at the time. *Bottom line though is cheapness; IDE adds little
cost to a system if implemented right, with SCSI you have to pay for a SCSI chip, and DMA if you want SCSI to outperform IDE.
IDE has lots of problems with standards, mainly in the area of how fast a disk will go. *It is, after all, a hack, one of these after-the-fact standards (of course, most things are, that's where SCSI came from too).
>Perhaps when SCSI finally reaches the standards' stage...
It's standard enough to be useful, far more so than lots of other really useful things. *If you wait for a 100% bulletproof standard, it'll be obselete
by the time its out.
>At the moment it most certainly makes no sense for a home computer (I have >SCSI on all mine and I see very liitle advantage over the IDE I had on a PC.
If you have only one drive, than the only real advantage to SCSI is the higher throughput you'll get with a DMA-driven SCSI implementation. *At home,
I have three drives, so I'm already beyond an IDE system (but I'm not typical).That's on an A3000; I don't think the bulk of A600 owners will know the
difference, much less care. *When I was 16, a hard disk of any kind on my home system would have been plenty. *Hell, affordable floppies would have been an
amazing thing...
>In fact SCSI on a PC leads to a lot of nightmares unless you >know what you are doing and/or have a MC/EISA PC).
Well, don't blame SCSI for PC problems. *The PCs work fine with hardware they they know about, and horribly with most everything else.
>I'm really getting annoyed at all those in this group who seem to think home >users needs 486's, 040's, 030's,etc...with scanners, SCSI,etc...Come down to reality.
Most of the folks in this group want that kind of stuff at their home. *But they're not typical home users, they're for the most part already computer
people. *A computer person knows what to do with a computer and can use up a good portions of a really expensive system. *You wouldn't expect sports car
designers to drive Ford Escorts, or movie makers to settle for 19" TV screens and cheap-ass VHS at home either. *The Masses, on the other hand, spend most
of their money on the cheap stuff, which they don't understand all that well in the first place.
>My NeXT and Apple/Mac computers use SCSI. It's no big deal.
Well, certainly for the NeXT it is. *You're not going to be happy with paging to an IDE drive on a UNIX system, everything else has to shut down and wait
for the hard drive. *The Mac SCSI implementations are, indeed, no big deal;they'd generally do as well or better, performance-wise, with IDE.
>CD-ROM's do not need SCSI?.
But they need some kind of interface. *Unless you have PC bus and can plug in a card that implements a custom device, or glomp on a box like the CDTV-thing
for A500, SCSI is your only option for CD-ROM. *Tape backup can work over a floppy port if you're willing to wait. *Scanners need a high speed interface,
IEEE-488 is a popular alternative to SCSI, but again, if you had SCSI, you wouldn't need to buy another interface card (then again, you could run the
hard disk over IEEE-488 if you wanted to, like PETs and our old HP development machines did). *A600 owners don't need any of this. *Most higher end systems
do. *We have assorted scanners around here, along with laser disks, ultra high density floppies, fast tape drives, Syquests, etc. *Most of which come
only in SCSI or custom interface varities.
Dave Haynie * * * | Chief Toady, Frog Pond Media Consulting 2002
>>Amiga like performance in 1985 from the PC?
It didn't exist.
And the claim doesn't have any place in today's market, anyway. Look out there.. see any simplicity? I see Pentium 4s with 125 million transistors, nVidia and ATi chips with 70-something pipleline stages and a similar number of chips. We're not just talkin' "more complex than the Amiga", we're talking "more complex by many orders of magnitude" than the Amiga.
The problems were elsewhere, and with 20/20 hindsight and an actual involvement in this stuff, I can tell you precisely where the problems were. On the chip front, it wasn't the complexity of the chips, not at all. It was the simple fact that Commodore management didn't spend the necessary R&D money to maintain development of the custom chips in any useful way. Look at modern tech companies, like nVidia, ATi, Intel, or AMD. See how much money they plow back into R&D. Had Commodore done this at similar levels, there would have no problem delivering new, interesting chipsets.
Second problem was software: the details of the custom chips should never have been revealed to programmers. This would have removed any need to maintain binary compatibility. You don't worry about this today, of course, because one ATi chip may or may not be anything like the next, but no one cares, your drivers account for any differences. That is the right way to do it, and it was well established as the right way to do it before the Amiga ever shipped. That's the one area of the whole OS that should have been done differently, before 1.0 shipped, IMHO.
>- First, the basic A1000 hardware should have been given a boost with extra RAM and a hard drive as soon as the cost of this non-proprietary hardware fell.
The big problem here: with about 150,000 or so Amiga 1000s in the world, only some of those owners buying memory upgrades, and five or
six companies offering them, there was no way the prices could have fallen. Not only that, but memory was wicked expensive in the
mid-to-late 80s. DRAM was in a worldwide supply crunch, and the prices reflected it. While I don't disagree with the "it would have been
nice" factor, it also wasn't even remotely possible.
>The custom graphic chips are often described as an evolutionary dead end. *In fact, they had considerable room for improvement while retaining a very high level of backward compatibility, as AA showed.
Well, AA's compatibility was good, sure, but it was very much evolutionary, and that showed. AA would have been great in 1988, ok in
1990, but in 1992, it was an also-ran, capabilities-wise. That goes back to budget.. C= only had so much money to spend on improvements
(especially considering the huge salaries the top management got -- keep in mind, at this time, none of it was performance-keyed, and
folks like Gould and eventually, Ali were making substantially more than the top dogs at Apple, IBM, Compaq, etc.
AAA, on the other hand, spent huge bits of technology on being compatible. Much of that was totally impossible anyway, at least in some configurations. For example, the 64-bit setup (one Andrea, one Mary, two Lindas, and two Monicas.. I found this: http://amiga.emugaming.com/amigaaaa.html with some of my drawings and a shot from "The Deathbed Vigil" on it... seems correct, though they missed the fact that AAA was actually in 1988, long before AA began).
>Jay Miner's team had already finished the ECS upgrade in early 1987.
Much of ECS was actually done in West Chester. Some of the delay was the simple fact that Amiga Los Gatos wasn't using modern CAD tools.
They had to capture, simulate, and verify all the old stuff before anything new could take place. The other problem was more serious: the
original Amiga chips, and ECS, were done in Commodore's 1.5 micron high-speed NMOS process. They were very much pushing the limits on
what that process could achieve. Moving to CMOS was critical, but keep in mind -- these were ALL transistor-level designs, not the gate-level
stuff people do today. Basically, you don't take an NMOS chip and redesign it into a CMOS chip, you start over from scratch.
>Commodore's clever leaders then sacked the original Amiga team and managed to lose the ECS blue prints.
None of that happened. Again, ECS was done in West Chester, with some consulting with the original team. For that matter, it was the second
major Amiga project done there, since Fat Agnus was designed in West Chester (Bob Raible did the main architecture, Victor Andrade did the
chip, as I recall).
>> ECS saw the light of day in 1990 only after prototype chips were reverse engineered.
Also not true. There was no reverse engineering done. ECS chips were running well before 1990, and they were phased into the A2000
production as a running change.
>Had Miner et al been retained with a modest R&D budget (which was instead ****ed away on second rate PC clones) I don't think it requires a huge imaginative leap to envision AA in 1988/9.
Well, you're right about the budget in general. However, very little was spent on the PC clones development. The point of that was to allow C= to be a "one-stop shopping" supplier. One big problem Apple had in Europe was their lack of a PC -- many large companies, in the day, wanted to deal with a single vendor for all of their PC needs. If they wanted PCs and Macs, and you didn't make both, they'd have to
choose... and it never went against the PC. This is also what drove the Amiga UNIX project. PC development was done in-house, but it was primarily systems work. The chips stuff they did was very simple gate-array development (which means, systems engineers like me, Greg Berlin, George Robbins, etc.... only those in the PC group. Not chip designers) for glue logic. By the 1990s, they were outsourcing the
PCs, much like everyone else (Dell, Compaq, Gateway, etc).
>Denise is the only thing that changes?.
Incorrect. Denise did pretty much everything Denise could ever do. That's exactly why the Lisa chip in the AA set was the only totally-new chip in AA. To go much beyond Denise (say, to add just the simple upgrade from 16 to 256 LUT registers), CMOS was the only possible choice. It couldn't have been a simple add-on to Denise. Of course, even if it was, you would have had to tweak Agnus to deliver an increased bandwidth (memory access in AA is 4x faster than in ECS or OCS, if you didn't know).
>With something like Dave Haynie's A3000+ in 1989?
Certainly, if C= management had treated Engineering with the same kind of investment that, well, pretty much everyone else in high-tech done,things could have happened differently. We never would have had AA in 1989, but with proper investment, maybe AAA (which began in 1988, but probably could have started a year or two earlier, with the right backing). That certainly wouldn't have sucked.
>I think Commodore would have found enough breathing room to transition to PowerPC/retargetable graphic cards, just as Apple did.
There's no telling. For one, it's not an instant assumption that we would have gone to PowerPC. That was probably the right decision in 1990-1991, sure. Much before then, MIPS would have been the obvious choice; too much beyond that, and x86 woudl have been the only reasonable choice. The system architecture I designed in 1991, which was intended to be the transitional machine (moving from dedicated
AA/AAA chips to Hombre and commodity RTG + PCI chips, moving from 68K to "something else") was CPU independent.
>The fact that they lasted until 1994 making the choices they did shows how brilliant the 1985 design was, and I think you do it a great disservice.
Guess I missed that part.
One thing I did, because, well, I could do it (leading high-end Amiga development had its advantages, and pretty much anything that Jeff Porter, me, and sometimes Greg Berlin agreed on would fly, at least in the days before Ali & Sydnes decided to kill the company with their incompetance and micromanagement), was to ensure each Amiga system was long lived. So we had the CPU slots (which Apple eventually copied, but the x86 world never did, even though IBM actually published a very good paper on this in the early 90s), video slots, expansion, etc. I absolutely knew that the C= way (and fortunately, I learned that way in the TED and C128 days, so as not to have to experiment on the Amiga) was never going to get us a new model every year. Of course, that only goes so far. My goal was to make your machines useful for at least five years. I didn't count on 10-15 being necessary
Dave Haynie * * * | Chief Toady, Frog Pond Media Consulting dhay...@jersey.net| Take Back Freedom! Bush no more in 2004!
>>Should Commodore of bought Sinclair as a cheap better computer than the C64?
"Better" is a relative term. Some people really just do like the C64 or the Apple ][ better than any other computer. That really is their business. I have personally found that a ZX81 makes the best frickin' doorstop of any computer known to man, though I haven't necessarily fired up a ZX81 for its computational use, or tried every possible computer as a means of holding my door ajar.
>> But I just like to say that the 68000 is 16 bit, I know it 32 *bit internally but its classification is 16 bit.
Well, there are three reasonable "bit-ness" viewpoints.
>I wish people would either stick to the same rule for all CPUs which is to informally designate them by their external data bus width only, or they use the form: max general purpose data register width / external data bus width.
No, not really. For one, no one really cares all that much about the external bus width, unless they're a hardware guy like me actually hooking the sucker up. That is one valid classification (external data bus width in hardware), but it's really just of interest to HW people. To everyone else, it's just an implementation detail and perhaps a performance optimization (eg, a 32-bit bus probably goes faster than a 16-bit bus, if the chip inside can actually use data that fast).
Anyway, the second legit via is the programmer's model. The 68000, for all intents and purposes, is a 32-bit chip to the programmer. The 68020 is certainly a slightly better and, sometimes, considerably better 32-bit chip, but that's the main difference.
Then there's the CPU designer's view. If you call the 68000 a 16-bit chip, it really should be because the chip only has 16-bit data paths and 16-bit ALUs (three of them), not because it has a 16-bit bus interface. Same reason the Z-80A is an 8-bit chip, even though it can do a few 16-bit things, too -- internal functions happen 8-bits at a time.
As for address bus width, not salient to the discussion of chip "size". There are 32-bit CPUs with 12-bit addresses, or 36+ bit addresses. I'm using an 8-bit and a 16-bit CPU in a design, currently, neither of which expose either an address or data bus to the outside world. The address bus says more about the application for the particular instance of a particular chip than it does anything about the CPU core, programming, etc.
>>Bill Gates is one of the reasons for why Amiga died 8 years ago?.
Not really. Microsoft was powerful in the 80s, but not invincible. Commodore was killed off by software pirates on the outside, and gross
mismanagement on the inside, before Gates & Microsoft even became a factor.
>Why not have a day of sorrow/remembering, on the day when its 10
> years since the bloodbath of CBM. (just an idea)
Which bloodbath? The "Last Big Layoff?" You'll have to wait until April 27th, 2004 for the day of the big layoff (it wasn't THAT huge in
terms of people, but it was more than half of the remaining staff, so percentage wise, it was the largest). Or April 30th, 2004, for the day
of the actual bankruptcy filing.
>>has anyone tried to implement NTFS for the Amiga? *
I rather doubt it. There are religious issues
>> : Could the Amiga *handle such a file system?
Sure.
>> There is no reason why it couldn't, but why not run a native FS like PFS3 (commercial) or SFS (free)?
What makes those any more "native" than NTFS? They weren't written by any of the various Amiga companies. PFS began life as a sort of ill-conceived UNIX file system for the Amiga. It had a few goods ideas, but some really bad ones as well, which made it far worse than a properly tuned FFS. It wasn't even remotely recoverable, and it wasn't remotely error-free enough for that not to be an issue.
Supposedly new versions are better, but you'd need substantial changes to make it a good file system, structurally. And it's limited to small files (<4GB). And how about attributes? For the Amiga to stop disappearing, it needs to get modern.
>> I can readily recommend SFS, it makes my life much easier.
SFS claims to offer true journalling, which is a big improvement. Though I still would demand the recoverability -- sometimes you want deleted files back. And do you want to bet on anything being bug free, or have a fail-safe. Also, no attributes, no large files (<4GB). And on the web site, they claim there are problems with large blocks on FFS, but don't actually point them out. Having used large blocks for
over 10 years without a single "other problem", I would like to know what they're talking about.
> * Well, I thought it might kind of neat to install removeable hard drive cages in my Amiga and PC, and be able to swap hard drives between them in this say. *NTFS is more robust than FAT32 so I thought that might be a better choice, if it was available to me!
Surprisingly, NTFS is an excellent file system. It supports 64-bit file sizes, and it has the curious aspect of everything in the file system being a named attribute. So it's absolutely trivial to add Amiga-like features, such as Amiga dates, protections, etc. and still maintain full compatibility with the simpler schemes used in NT. FAT16 made sense, since it is so commonly used on floppies and memory cards. At present, FAT32 would be pointless, other than for similar data exchange.
>> August 1985 vs. September 1985. No one at Amiga or C= imagined for a minute that the Amiga was supposed to be a C64/C128 followup. It was to the C64 very much what the Mac was to the Apple ][... the
>> future. As one of the C64 revivalists/nostalgics I have to agree. In the 80s I didn't expect a new computer to be compatible with an old one from the same company.
And, in fact, they never were. Well, there was a level of compatibility between some of the CBM machines, but not 100%. The C64
didn't try to run VIC-20 programs. Atari upgrades were hit and miss. Apple did a little better with the Apple ][ series, but they were also
less aggressive on improvements.
The C128, of course, was compatible -- actually, with both the C64 and the old Kaypro CP/M machines to a large extent, if you want to get technical. But this was also a different era, or at least the beginning of one. While we were developing the C128, we saw where other systems were going, with a real OS and all. Hell, we even sort of had one, with GEOS on the C64 (and eventually, 128).
>I started out with the VIC20 in 1982. When me and my brother could afford a C64 in 1984 we sold all our VIC stuff without a second thought. Why keep an obsolete computer when the future was here?
If you bought new, it was expensive, but eventually, it wasn't a big price hit to upgrade, compared to Amigas, much less Macs or PCs in the
day. C= was also fairly good at keeping the peripherals compatible. Until the Amiga, there had only been the IEEE peripherals and the serial peripherals (the C= serial bus, of course, was basically the IEEE bus implemented in bit-banging serial).
>Same thing when the Amiga became affordable with the 500 in 1987. We never even considered the possiblity that it could have been compatible with the 64. We wanted an Amiga so we sold our C64, 1541
>drive, MPS 801 printer and all the disks. Why would we want to hang on to all that old stuff when we had an Amiga?
Right. And there really was no precedence, in those days, of compatibility with upgrades, at least not in the consumer market. So there was no expectation. We're all used to compatible upgrades now (even occasionally in video games!), but that was a product of a more complex design.
>To answer the original question: the reasons for me starting to play with C64s and 128s and not as much with Amigas (though I have a few of them too) are that they have a greater sentimental value to me,
>they're easier to find at flea markets, there is (it seems) a livelier community around them, and (this is subjective) I find them more accessible.
I think these systems, the C64 and C128, were the last of the sort of computer you could totally master yourself, and that is part of their allure. I started on computers in 1973... first on a programmable HP calculator (with core memory and a CRT). That was easy -- I learned everything there was to learn in a week or so. So I got my Dad to hook me up with a computer at Bell Labs, which turned out to be a fairly
serious scientific computer, but what did I know. This one was way too complex, and the information was hard to come by.
Some years later, my best friend bought a PET 2001, and man did I go to town on that one. That was a computer that I could program out the wazoo, no worries about 300 baud, thermal paper, etc. And you could learn each and every secret about the machine, only, not in a week. So it was challenging. That was the case with most of the systems of those days. I used others; I still like the C= best (and not just because my name's hidden inside one of 'em...).
>Something to do with not having to wait for the computer to boot from disk and not having to use a mouse and a GUI (my Amiga 500s actually feel slower than my C128s and C64s), I think.
And NO ONE knows everything about an Amiga. Otherwise, I never would have had a chance becoming something of an expert on a piece or two of
it. It was cool, for users, because all thse "stuff" extended what programmers could deliver, but it was far harder for anyone to get The
Big Picture.
Dave Haynie * * * | Chief Toady, Frog Pond Media Consulting dhay...@jersey.net| Take Back Freedom! Bush no more in 2004! "Deathbed Vigil" now on DVD! See http://www.frogpondmedia.com
>Hi Dave, *I have been away for a few years now and haven't been able to >follow the Amiga scene. *Still holding annual digs at your place?
I have been... not sure about the timing this year, but I'm thinking
of doing it again. Wouldn't be until late August or September.
> Where is the old crowd hanging out online now?
There's nothing really central for the C= crowd. I'm on the Team Amiga
list, which has a number of people from the old days, mostly talking
about things unrelated to the Amiga these days.
Last edited:
Its astonishing what went on
