If you think a RealAudio server on a C64 is weird, this will totally cook you're noodle. There's a whole new C64 in development, designed in programmable logic by Jeri Ellsworth (http://www.geocities.com/cm_easy). This is absolutely for real -- I saw a prototype last year, and the first PCB last month at the Amiga Expo show near Baltimore. Jeri's brain-picked many if not most of the original C128 team on the details -- we were a little more accessible than the C64 people.
Crazier yet, it hooks to a PS/2 mouse and keyboard (the keyboard interface can emulate the C64/C128 port scan, something I actually worked out and proposed for the C128D, since the hardware to serialize this cost less than the cable we used to run the whole scanning matrix to the dumb keyboard, but it wasn't done). It hooks to an ATA hard drive, and has one PCI slot (I/O mode only, intented it seems for Ethernet support).
Look at all the damn fools using computers today! Of course they're too easy.
Man, when I started out (1973), you had to be really, really dedicated to use a computer. Everyone who used one knew how to program (ok, mostly because they didn't do much else). When I hacked into my first UNIX system (1975, one of the hundreds at Bell Labs in Holmdel), half the users didn't even have passwords. Didn't need 'em, UNIX (System III, maybe something even earlier) was just to hard to use.
Now anyone can learn a computer. My Mom knows three operatings systems (AmigaOS, MacOS, Windows). What's this world coming to? They were hard to create, they should be hard to use. In fact, most of you shouldn't be on a computer, either. I think the test should be "can I design one from scratch". In case there's no silicon FAB in your backyard, you can use FPGAs...
It's all relative. The Duron, for one, is a fairly fast chip -- nothing more or less than an Athlon "Thunderbird" with less L2 cache. The floating point still smokes a PIII.
As for "professional", unless you're involved in video/MPEG-2 rendering, realtime digital audio processing, or heavy-duty CAD rendering, there's a fine chance that any CPU still made will be just dandy for professional use. Most professionals spend their time in compilers, word processors, spreadsheets, or other non-CPU-intensive apps.
Most people find Internet just dandy on anything made in the last 5, maybe 10 years. Despite all of Intel's consumer ads, most of the time you're bandwidth limited... and I'm saying this from the other end of a serious optical pipe, not a 56K dialup.
As for games, sure, some of those kiddie demand the fastest performance they can get their hands on. So they're running 1GHz Athlons if they're dedicated. Compared to most computers in most regular consumer homes, the Duron is very fast. After all, it's mainstream, and it's relatively new. You might expect Intel's next major Celeron upgrade to run a little faster than the Duron, if they can swing it.
Re:That's cuz computers work on magic, not science
on
Hackers And Mysticism?
·
· Score: 1
So, if programmers are sorcerers, what does that make us computer hardware engineers? And, in understanding how the computer operates, does that may me mad, relative to normal life?
In a typical RTOS, you have atomic operations on semaphores and other very small chunks, and that's it (well, ideally). So essentially, your atomicity is on a data level, it's not generally a processor/interrupt/hardware issue. It's also possible, but more complex, on multiprocessor systems.
Much of this is based on the OS design. The original RTLinux people looked at implementing realtime in plain old Linux, but decided it couldn't be done without massive work and risk to the API integrity. Of course, they were after hard realtime. Mainstream Linux would benefit greatly from "as good or better than Windows" realtime; hard realtime is only truely necessary for certain classes of embedded apps.
Another thing they don't mention, but should, is the need for a reentrant kernel and real, actual, useful, and working threads in Linux. This goes hand in hand with well written realtime/near realtime applications. Being preemptable is a whole lot more useful when the multitasking is far more fine grained than you have in today's Linux. Also, kernel-mode threads (as you have in BeOS and NT) are particularly useful for realtime-sensitive data.
Ummm... not to be a Microsoft supporter or anything, but NT/Win2K, while not a hard realtime system, is dramatically better at functional realtime than any mainstream version of Linux. Not in the BeOS or QNX league, but good enough for things like hard core Digital Audio Workstation applications (when you have 64 audio tracks and 128 MIDI channels running low latency and glitch free under Linux, Linux will have a real chance in this market, but it's challenged enough today playing back a single MP3 without glitching).
Technically, you're correct. There are two acid tests you can run on a patent before doing anything else: is there prior art, or is it "obvious to one skilled in the art".
However, the patent office is notoriously bad at investigating prior art. They basically just look at existing patents, not "what's out there in the real world". Second problem is that there's no requirement that the patent examiner of record actually be "skilled in the art". The PTO was issuing software patents for 10-15 years before they had any actual examiners with software engineering credentials, and seem to be even further over the edge on business practice patents.
Here's one heinous example I know about. I was a technical consultant to an employer's legal team going up against an IBM suit -- basically, they wanted access to all our technology, so they came after us with a stack of patents. One of these, filed in _1984_!, claimed to patent nothing more or less complex than cut and paste of text between multiple buffers in a text editor. Since our computer shipped with MicroEmacs, we violated this "invention". Never mind that the exact same sequence of keystrokes would have achieved the exact same effect in TECO Emacs, which I was using at CMU back in '79 (derived from earlier versions done at MIT...).
Sure, you can fight a patent with prior art, but the presumption is usually that the patent was researched fully. Clearly an increasingly bad assumption, but that's what you'll find in the system.
My company (Met@box) is doing this: internet appliances with some options. Rather than a fixed modem, we have a "communications module", which could be a modem, or ethernet, or eventually xDSL, cable modem, etc.
But there are issues. In some cases, we are bundling services with the system -- these are not yet something you would likely buy for home use if you already had a PC. In some installations, you're getting our box via another service provider, such as a cable TV company, and may very well be able to share the connection with a PC.
TiVo and more specific purpose units are, and sure, there's no really reason no to allow alternate hookups even there, except perhaps the support issue. They want to sell an appliance -- every unit is basically the same unit, and not offering an Ethernet hookup may save on the support side without really hurting their acceptance among advanced users. After all, I bought my TiVo anyway; I would have preferred an ethernet hookup.
Of course, TiVo may also be taking a shortcut and not running over a full TCP/IP link, which is another issue.
In time, this will change. The phone thing works for the first appliance, gets annoying for the second (I also have a satellite box in the same room), unmanageable much beyond that. As connected appliances grow, they will at first support and later demand some kind of home network. Hopefully, that'll be IP6 based with some choice in options, though there's certainly a chance the consumer electronics giants will embrace one thing (maybe bluetooth, maybe not) so completely that computer users and smaller manufacturers like me will have not choice but to follow suit.
We're entering the transition phase now -- give it a few years.
FYI - the original Amiga chips, done in Commodore's 1.5 micron HNMOS process, topped out at something around 60,000 transistors. The Lisa chip, part of the "AGA" (formerly AA, formerly formerly Pandora) chipset, ran up to around 120,000, as the only CMOS chip in the main Amiga chips (there were CMOS gate arrays in most Amiga systems, glue chips that basically implemented the specific system differences).
Back in 1985, Amiga hit so strongly in the sound and graphics areas because Jay was a genius, but also because no one else really tried to do it at such a level. Commodore could have made their own CPU, like in the 8-bit days, but it made no sense: you couldn't compete, in a market of maybe a million units a year, with a whole CPU family sold in the 10's-100's of millions.
The graphics companies today are like the CPU companies back then -- they're locked in a constantant battle of one-upmanship with each other, working on very complex designs which sell in the 10's of millions for pretty little cash. The only possible way Amiga or anyone else can compete is to do something very different. If your target is still the desktop, that would be a hard one, especially given all the graphics chip talent these guys have sucked up from the workstation world.
You don't think all this was motivated by a few kids downloading MP3s, do you? Sure, it's wrong, even if it isn't necessarily hurting anyone (I've used Napster twice; each time it cost me over $100 in CDs I just HAD to get after listening to the MP3.. but that's just me).
I have an article on ZDnet (so kill me) about this very thing: TalkBack Central: RIAA vs. Artist: Will the real victim please stand up. Basically, the RIAA represents the huge record companies, not the artists. And they're being forced to contend with an unpleasant fact: they're dinosaurs, and the mammals are looking about ready for an uprising.
Lemme 'splain. There was a time, long ago, when a record company was instrumental in the recording process. They found musicians, hooked them up with other musicians, hired songwriters, producers, etc. That still happens to an extent: groups like "N'Sync" were manufactured this way. But they're the exception these days.
So now enter computer technology, which lets the artists work in their own private home studios. This was inevitable from the days in the 60s when folks like the Beatles first used the studio as part of the creative process, not just a place to lay down a few tracks. I can get better quality in my home studio than "the lads" every managed (technical, folks -- the talent just ain't there). So, what DO the record companies do? Primarily, they manufacture and distribute CDs and tapes on a huge scale. Their clout gets their artists in the stores and on the radio. They advance musicians money against eventual royalties for making record and videos, but rarely actually pay for that. Aside from the whole meatspace distribution, they're functional as an engine of transfer in just one more way: up to 95% of the profits made on a record goes into their pockets.
Now enter Napster. Napster has two important things, even if it's morally wrong. First off, it demonstrated a great peer to peer technology, bettered still by GNUtella and others. What this did is fundamental: it moved the problem of electronic distribution from necessarily being a huge server-based project (look at the size of eBay's rig, for example, and they're just dishing up pictures and text) to Just An Application. The second thing they did was on the culture. After Napster, there are millions of college kids, and others, who now think electronic distribution is a fine way to listen to music, not just a cool hack.
So the RIAA is fighting the future. Their client face a serious loss of power in a world in which artists and listeners can have a direct relationship. And really, find a way for those musicians to get paid 25-50 cents per download, and the record companies will have no more place in this world than the noble apatosaurus. They're fighthing Napster especially, because they currently have name-enough to start doing something on the up and up, which they couldn't fight.
That's not the only place they're fighting. You may have notice the current rise in producer-created bands like the aforementioned N'Sync. It's no coincidence -- a group built that may well be owned by the record companies. Contracts have been building up to give the record companies dramatically increasing power. Go buy that new Clarke/Baxter book, and you'll see their names on the copyright page. Go buy just about any CD, and you'll see (C)2000 Sony, Inc. or some-such.
And it gets worse. If you're a consultant, you may have noticed that a major "work for hire" bill passed a few years back, effectively making it harder for you to work for a company and still claim what you did for them is an independent work, and not theirs. Well, the RIAA & co. managed to sneak what's in effect a rider to that bill through, piggy-backed on some cable TV bill, which make "recording artist" just as subject to the new work-for-hire laws as "computer engineer". So unless they had particularly good lawyers a few years back, there's a fair chance than many of the major music acts out today are effectively just employees of their record companies. Which, if nothing else, lets the dinosaurs have serious control of much of today's popular music catalog. Making it all the more difficult for an independent, legal, and micropayment-based distribution system to gain momentum.
Water is certainly what WE need for life. So when you're talkin' about a place like Mars, the existance of water fills two needs. First of all, there's a real desire in some circles to change the role of astronaut back from "glorified truck driver" to "explorer". The only way to be an explorer is to go somewhere new. The only way to get any major government or world consortium to pay for it is to prove it has some specific validity, scientifically and probably economically.
So the chance of life, as we know it, goes hand in hand with the discovery of water on a planet. But more to the point, if there's water there, reasonably accessable, we can stay there indefinitely. If not, we'll have no permanent presence with our level of technology.
AIM is the typical "we're not Intel or Microsoft" story in this industry. Rather than let Intel actually need to punch out a few teeth, Apple basically had them all surgically removed while Mot and IBM were, apparently, stoned on nitrous. Or at least their lawyers...
The only way PPC systems would match x86 in price would be to have a large enough open systems market rivalling the PClone market. I was doing that in '97 when Apple pulled the plug. And while moving into the set-top box business from there has our stock some 20x higher than in the Maclone days, I'm still bitter. Not just because Apple proved to be the dicks everyone kept telling me they'd be (no, they've really changed, I met with the CHRP group myself), but simply that the PPC's good enough. It deserves success.
After that fell blow, neither Morotola nor IBM has felt particularly compelled to race with Intel just for Apple's sake. When phone or internet switches or other high-end embedded gear need an upgrade, they'll crank out a faster chip (like IBM's forthcoming PPC750 update, starting out at 700MHz). With all that copper and SOI and all, it's not like IBM couldn't deliver if they wanted to.
Love the quote, man -- and I don't come to/. for egosurfing (though at least I'm on a genuine AMD/Linux machine this time). So far, I think the relative improvements from Pentium to Pentium are damn near logarithmic. If they really changed the microarchitecture for the first time since the "pro", maybe this should be, I dunno, the Pentium 6 or something. Or maybe it actually will be a yawnner by then, and is appropriately named.
For one, the "game console" was never the intention of the Hi-Toro/Amiga gang. It was simply a way to attract venture capital, because in '82/'83, consoles were a big deal.
Metacomco didn't create AmigaOS. Most of the AmigaOS was created in-house. The DOS subsystem, one fairly small piece of the whole OS, was not done being completed in time, so Tim King and company ported the DOS subsystem from TriPos (sort of the UK's answer to UNIX) to the existing Amiga OS.
There NEVER was a real Amiga 1500. The thing sold as an Amiga 1500, only in the UK, was simply a bog standard Amiga 2000 bundled with two floppy disc drives, rather then usual single floppy or floppy and hard drive.
The Amiga 1200 and Amiga 4000 shipped in the fall of 1992, a full year-and-a-half before Commodore bit the big one. They were actually somewhat bastardized versions of things that had been in the works in 1991, with the intention of shipping in early '92. Commodore had gone though a management change, and their original idea was to drag feet on the Amiga, in an attempt to kill it, while bolstering the PClone program. The flaw in that formula, as any first year business school dropout, you, me, and my cat could tell you, was the simple fact Commodore made virtually all their money on the Amiga. By the end of '92, all of the PC engineers had left or been assigned to Amiga projects. The VP of Engineering responsible for these screwups had also left, but the root problems in upper managements remained, and led directly to the end of Commodore.
It looks like technical problems in retrospect, but basically, when managers at a company like Commodore (peaked at about 1 billion in sales, and there may have been some monkey business even in those days) take higher salaries than the bossmen at Apple, IBM, and Compaq, and spend less on engineering than plain old PC companies (who often do no actual engineering, period), well, you just can't maintain technological advances, even with some of the best engineering people in the business.
The New Amiga is really only Amiga in spirit, if that -- they have yet to show it, one way or another. I'll with-hold judgement until they're actually doing something Amiga-like.
-Dave (systems engineer on A2000, A2500, Zorro III, A3000, etc)
From a read of the patents, from other more open work at other companies, it's fairly clear that if Transmeta is really making a CPU (they have tried to deny it without really denying it, I think they are), what they have is a VLIW engine of some kind with lots of hardware support designed to allow it to run not just a popular instruction set (say, x86) as fast as you'd expect a modern CPU to run this, but also emulate the hardware subsystem. So they could release a thing that looks and behaves like a PC, but actually doesn't have either x86 or a real PC architecture living down below. You all know this is possible; x86 emulators have lived on 68K, PPC, SPARC, etc. machines for years. The Transmeta stuff would make it fast, say, 50% or so of the native CPU speed.
The Linux connection (strongly hinted at by Jim Collas, former President of Gateway's Amiga division, when I spoke to him about the various AmigaNG rumors) is, of course, Linus himself. But the coolness factor is this: here we have a CPU, running emulated instructions as fast as some version of the Pentium II/III, and doing it cheaper and with much less power. Only the first one is new work -- it's easy to find CPUs that run faster than PII/IIIs using less power (PowerPC, for example). The thing is, the x86 code runs at 1/2 or so of the native Transmeta CPU core speed, perhaps based on the limits of their dynamic recompiler, the loss of abstraction in binary code, etc. What about native code?
The modern trick in all this, same thing Sun's doing with MAJC, is to make the idea of a VLIW processor legit by never tying a system to native binaries. You run x86 or Java Byte Code or whatever through a translator, and when the machine architecture, and thus instruction set, change, you build a new translator, everything's hunky-dorey code wise, and you get to forget entirely about hardware legacy. But Linux and other open source stuff doesn't have a binary legacy problem, period. So there's no problem in coding Linux native, and if you did this, Linux would run twice as fast, relatively, as any closed source OS on this platform. The need for Linus on this would be building a mixed code manager for Linux, so that the kernel could cooperate actively with the dynamic translator and run your choice of x86, native, or other code modules (full-speed Java, etc).
This is the last layer of HW abstration. Some day, all CPUs may be built this way.
Gateway has dropped all AMD systems, true. But this, when taken with other recent Gateway actions, is more likely to scare Gatway shareholders than non-Intel fans, as it's a fairly clear indication of troubles within Gateway.
The most visible of these moves is that they have dessimated their customer/tech support group. In restricting their products to Intel-only, it's likely they're shipping pretty much the same thing in every box, CPU/motherboard wise, with all motherboards made by Intel. This will definitely cut support costs, simply because support staff needs only know one or two PCB variations.
It's also a retreat back to what Gateway's confortable with. They were one of the last major companies to start using non-Intel products, and they're one of those with the least internal development around (which they have also cut back recently).
This isn't to say that fear of Intel, or even Intel themselves, hasn't hurt the Athlon introduction some. But that's not the only explanation; the Super7 introductions were plauged with chip and software issues for some time. Some vendors will want to believe that's not an Athlon problem before they jump on board.
It's also a standard cost-benefit thing; is Athlon compelling enough to jump now versus later, or as a flagship, or whatever. Maybe? If the Transmeta fulfills its design goal of delivering a 4x cost/power advantage over some comparable state-of-the-art chip, it'll catch on fast. Especially on laptops, which, being fairly well self-contained and made by larger companies, are less succeptable to fear of Intel.
It was the Saturn that competed with the Sony PSX, not the Genesis. The Genesis was the original Sega machine, the 68000-based unit that competed against the Nintendo "family computer" and SuperNintendo (based on a reverse-engineered 6502 and a 65C816, respectively).
The Saturn was technically resonable, spec-wise. But it had major problems, and that's why it was so easily curshed by Sony. For one, the dual H2 processor setup was hard for programmers no used to writing games on a multiprocessor system. And the machine was just too fat -- way too much hardware in there for what it did. As well as the dual H2s, there was an H1, a 68000, and a Yamaha DSP chip (the 68K, Yamaha, and 512K of private memory did their MIDI/Audio section, usually a tiny piece of one custom chip in the other machines). They simply couldn't compete on price with Sony or Nintendo. Look inside an N64 -- four main chips. The Saturn had more CPUs than that!
Naturally, marketing made a big difference, but that's only part of it. For one, you're marking primarily to kids, which for the most part means specs don't matter a great deal. This is the main reason companies follow the very successful Nintendo formula and establish recurring characters (practically every game written by Nintendo is based around characters who are from the original arcarde Donkey Kong, or characters introduced in earlier spinoffs).
Nintendo fell behind in that generation primarily by being a year late. Sega couldn't get software written fast enough to keep up their initial marketing push. They have learned from most of their mistakes this time around, and as well have a ~16 month lead over Nintendo and Sony, at least in the USA.
But I do agree -- you can take this market with th e right product at the right time sold the right way. Part of that is having zillions of dollars to put behind it. Back when the N64 was launched, they spent around $50 million just on the Christmas ad campaign. Today, it's probably more than double that. Microsoft can certainly play in that sandbox.
Re:Why Linux Supporters Should Be Excited About X-
on
More on the MS "X-Box"
·
· Score: 1
The X-Box, should it actually happen, can't be open. Here's why...
Sega, Sony, Nintendo, etc. are selling game consoles that range from fairly well (Sega) to highly (the others) integrated systems, with customized CPUs, etc. This is how these machines sell for $99.95-$199.95 today, which is rarely if every making any of them a dime in profit. Even with high integration, you're not making a PClone-ish system any cheaper. And in fact, the emerging specs make it sound much more expensive, but we'll see -- they could and very well might roll special versions of the chips, to keep them out of PCs.
So you have to follow the money. Money is made on software. Sony makes about 1/4 of their total profits on PSX software, even though total income from that market is fairly small compared to the many other businesses they have. Microsoft no doubt wants to make money on the software, and in order for YOU, as a developer, to pay MS for the luxery of supporting the X-Box with a game, you have to license something from Microsoft, or they have no hold over you. For the 3DO box, you had to licence the OS, which was always loaded from disc (probably what MS will do; this lowers the HW costs and eliminates OS compatibility issues). At Commodore, the CDTV and CD-32 had the OS in ROM, and simply required a C= Copyrighted file at a certain location on the CD.
That's not to say you couldn't hack it, just that such a hack wouldn't be legal to distribute.
Of course, if MS is the only one making money, this won't fly. They could adopt a WebTV model -- WebTV is sold roughly at cost, but Microsoft pays vendors a bonus for every unit sold, so they do make a profit (if they sell them, of course). However, this is illegal in many countries. Clearly, vendors can't compete if they have to make a profit on the box, that's what killed the 3DO box, even before Sony and Nintendo came along to bury the body.
MS could manufacture the box themselves, but they don't usually like to get involved in expensive hardware. Look at all the folks who have lost money on WinCE handheld machines -- Microsoft made money in every case, because they sold the software. I suspect they'll want the same thing now.
The X-Box should be well supported for a few reasons. For one, this is PC hardware as PC authors know it, more or less. They'll have to license the Windows or WinCE or whatever you put on disc, but it'll be pretty close to what they're doing in Win9x today. The CPU won't necessarily be fast by PC standards, but then again, the target is the 720x512 or less space of a TV set, not your 1280x1024 monitor. It'll be easy to port Windows-only games to this platform, and quite a few of the older games, not selling for the regular PC, could be recycled here.
On the downside, if you have to use Windows, the system won't kick butt at the same level as the Sega (which can run WinCE, but rarely does) or new Sony/Nintendo machines. Looks aren't everything, though, and much of this market is driven by hype and clever marketing, not just graphics horsepower. Microsoft is a master of this, so far.
Well, kind of. For situation (1), the Grand Caymans don't much care if you're a US Citizen or not -- it's not their job to administer US tax laws. It also doesn't much matter if you become a Cayman citizen or not, since its the US government you have to convince of your non-citizenship, or as far as they're concerned, you're still paying those taxes.
Now, there are two ways to go about this. If you're a criminal type, of course, you don't much care what the US government thinks. And you're probably not getting too many of those drug, prostitution, and illicit gambling charges on VISA, these are in cash. Which, of course, is smuggled out of the country hidden in dolls or coffins or whatever, at least according to most of the gangster movies I've seen.
The other thing is to actually become a citizen of a tax haven and remain a non-resident of other countries. An actual case of this is Irving Gould, onetime principle owner and Chairman of the Board at Commodore. At some point in time, he moved to the Bahamas, real house an everything, and I assume became a citizen. But he had to be careful about being percieved as a resident of any country. So he actually had people who kept track of how many days he spent in any given country. For example, it was not uncommon for the PET Jet to circle until after 12AM, when old "Uncle Irv" was coming into the US, just so that fraction-of-a-day couldn't be held against him in US tax court. In other words, maybe a bit shady, but all perfectly legal.
I don't supposed you would do this until you're making enough money to have that kind of lifestyle, but that kind of goes without saying. Most offshore accounts offer little interest, charge high fees for transactions, require opening deposits in the $10-$25 range, etc. It's not likely to attract the typical working stiff, even if you're a contractor or some-such getting lots of payments in cash. On the other hand, a millionaire with lots of foreign income and fewer scruples could well go for such a thing, profitably.
It'll probably only get easier, with the e-banking and e-trading. There are already US-based companies specializing in setting up foreign investors in US stocks and bonds, and it probably won't be long until US folks have the same online shot at some foreign securities. Now, when you're doing the electronic transfer to your bank in New Jersey for that killing you just made on the Newer Markt, ole Uncle Sam will be there to congradulate you. But all you'll hear from that Caymans' bank is the $50 (or whatever) deposit fee.
Which just goes to illustrate the old point that the rich do, in fact, get richer.
1) room for much more of the expensive L2 cache for full speed operation and modules up to 2MB.
2) redesign to the AGTL+ bus, which allows four processors to run, even using modules (versus the normal P6 bus, delivering two way SMP with the module, four way with the older socket).
3) more money. Intel knows that people who need 4-way or better SMP systems will pay for this, often to a foolish extent. Intel loves to milk some sector for high margins, and you can guess this won't happen when there's a direct (or near so) replacement like K6 or Athlon.
Ok, so now consider PIII Xeon with only 256KB of L2 cache. Certainly, this is the same chip as you get when you buy a PIII-regular. But of course, it's on the Slot 2 module, and if you want a four processor (or better) SMP system, you have no other choice. Intel basically has you, and they like it that way.
Ok, first of all, forget completely about "Slot". That's just a hookup, a physical connector, it doesn't tell very much of the story.
Slot 1 is the second physical delivery of Intel's P6 bus; Socket 8 (for Pentium Pro) was the first; Socket 370 is the third, and there's every indication Intel will market a Socket 426-or-thereabouts to handle some extra pins on a socket-based Coppermine chip. The reason you can't clone this without a license is simple: Intel has patents on the P6 bus. They originally didn't license, but now they are licensing, at least to chipsets (SiS, ALi, VIA) if not CPUs.
AMD extended Socket 7 (which is once of the physical conventions for the non-patented P5 bus) to 100MHz, and ran their K6 family there. But they realized that K6 would have its limits. In part of their IP settlement with Intel, they promised not to make clones of things like the P6 bus, so legalities for everyone else aside, that wasn't an issue.
Enter DEC (now part of Compaq). DEC designed the EV6 bus for the Alpha 21264, as a remedy for conventional CPU buses. In their earlier Alpha systems, even with L1..3 cache and all, they were so dependent on memory speed, the typical bus sharing in an SMP system (very important to DEC) was a problem. Especially when you wanted to make this bus (the cricital CPU to rest-of-the-world link) extremely fast. EV6 is a point-to-point bus; all you have is a CPU and some system chip on it, never anything else. This allows them to run EV6 very fast. DEC openly licenses EV6, AMD adopted it for their CPU. Since the PC market demands a socket or slot, they created Slot A, which (for cost reasons) uses the Slot 1 connector turned around. AMD runs Slot A at 200MHz now, 266MHz in the forseeable future. DEC runs EV6 systems up to 400MHz. This is a data rate; the bus runs DDR (a new data event on every clock edge, not just every clock cycle), which really doesn't matter; the data rate is indicative of performance.
The important thing to realize about today's Athlon systems is that they're something like the first Pentium systems shipped out, retrofitted '486 systems much as the first Athlons are running based on modified Super7 chipsets. So memory is stuck at 100MHz, half the CPU interconnect's speed. You won't see Athlon reach its performance potential until 200MHz, or better, memory systems are delivered.
And I do mean "or better" because of the EV6 architecture, there's no shared CPU bus. So a system chip (North Bridge) can actually use memory faster than the CPU can deal with. With switched interconnects, proper buffering, and super fast memory, one could imagine PCI, AGPx4, and multiple Athlons all rompin' along, each at virtually full speed. This can't be done efficiently within the current P6 architecture. That's why I find Athlon interesting. I hope AMD lasts long enough for it to realize its potential.
In essence, I agree. People want this for fun. I still use a number of Canon and Leica rangefinder cameras, even in this day of digitals and SLRs smarter than any human. Because they're fun, because there's a degree of excellence there not available in today's "wonderbricks". AmigaOS is very much the same way.
Your details are wrong, though. AmigaOS was NOT "based on Tripos". The AmigaOS was largely original work. In order to get to market in time, Amiga had the DOS subsystem (a tiny piece of the greater whole) from Tripos ported to the existing AmigaOS. That's the only Tripos connection.
The chipset isn't particularly interesting anymore, in this day of graphics chips running 600x faster than poor old Agnus/Alice, Paula, and Denise/Lisa, audio chips with more DSP performance than the fastest Amiga ever made, etc. It's important to know that the simulataneous screen resolution display was kind of hack, or chip limitation. Yes, there was a nice display list coprocessor that switched the screens. But the reason you could mix resolutions was the fact that the pixel clock was fixed at around 35ns. So every display ran 35ns, 70ns, or 140ns pixels. In a modern graphics card, you're mucking around with a PLL to independently vary refresh and resolution, and PLLs inherently take a more-than-scanline while to sync up to new frequencies. The cancelled Amiga Advanced Architecture project supported this trick by allowing up to four independent PLLs to exist in a system. But sure -- you can make an Amiga card for the PC. This was actually proposed twice while I was at Commodore, the first time it even made some sense (I did an engineering analysis). In practice? Look here: http://www.siamese.co.uk/pci-amiga-crew.html. The point, really, is "why bother". If you run an Amiga emulator on a reasonably fast PC, using the graphics API targeted to Windows or X, rather than Amiga chip emulation, you'll have an Amiga faster than any one that has ever existed in the real world. The OS, on the other hand, is still pretty damn cool. Some of you may recall a time when hacking was for fun, not profit. I think lots of people would love to hack the AmigaOS. It would be good, really, for an industry so caught up in ancient architectures like Windows and UNIX, to learn what AmigaOS has to teach. It's probably too late to make it anything that takes the world by storm, but that doesn't render it useless. In fact, most of the lessons of AmigaOS seem to have not made it into the PC mainsteam, whereas all of the hardware tricks did long ago, and then were soundly exceeded. As far as making new chips, a really sharp guy called Mick Tinker claims to have reimplemented the Amiga chips; don't know if he's using an FPGA or a true gate array (or possibly just simulating in VHDL or something). Of course, there's some extensions in there too; Amiga chips as they were implemeneted aren't terribly interesting today.
Slashdot Mirror
If you think a RealAudio server on a C64 is weird, this will totally cook you're noodle. There's a whole new C64 in development, designed in programmable logic by Jeri Ellsworth (http://www.geocities.com/cm_easy). This is absolutely for real -- I saw a prototype last year, and the first PCB last month at the Amiga Expo show near Baltimore. Jeri's brain-picked many if not most of the original C128 team on the details -- we were a little more accessible than the C64 people.
Crazier yet, it hooks to a PS/2 mouse and keyboard (the keyboard interface can emulate the C64/C128 port scan, something I actually worked out and proposed for the C128D, since the hardware to serialize this cost less than the cable we used to run the whole scanning matrix to the dumb keyboard, but it wasn't done). It hooks to an ATA hard drive, and has one PCI slot (I/O mode only, intented it seems for Ethernet support).
The C128 would last longer due to fast mode AND the extra memory.
Anyone with a C128 should type this one:
SYS 32800,123,45,6.
Man, when I started out (1973), you had to be really, really dedicated to use a computer. Everyone who used one knew how to program (ok, mostly because they didn't do much else). When I hacked into my first UNIX system (1975, one of the hundreds at Bell Labs in Holmdel), half the users didn't even have passwords. Didn't need 'em, UNIX (System III, maybe something even earlier) was just to hard to use.
Now anyone can learn a computer. My Mom knows three operatings systems (AmigaOS, MacOS, Windows). What's this world coming to? They were hard to create, they should be hard to use. In fact, most of you shouldn't be on a computer, either. I think the test should be "can I design one from scratch". In case there's no silicon FAB in your backyard, you can use FPGAs...
As for "professional", unless you're involved in video/MPEG-2 rendering, realtime digital audio processing, or heavy-duty CAD rendering, there's a fine chance that any CPU still made will be just dandy for professional use. Most professionals spend their time in compilers, word processors, spreadsheets, or other non-CPU-intensive apps.
Most people find Internet just dandy on anything made in the last 5, maybe 10 years. Despite all of Intel's consumer ads, most of the time you're bandwidth limited... and I'm saying this from the other end of a serious optical pipe, not a 56K dialup.
As for games, sure, some of those kiddie demand the fastest performance they can get their hands on. So they're running 1GHz Athlons if they're dedicated. Compared to most computers in most regular consumer homes, the Duron is very fast. After all, it's mainstream, and it's relatively new. You might expect Intel's next major Celeron upgrade to run a little faster than the Duron, if they can swing it.
So, if programmers are sorcerers, what does that make us computer hardware engineers? And, in understanding how the computer operates, does that may me mad, relative to normal life?
Much of this is based on the OS design. The original RTLinux people looked at implementing realtime in plain old Linux, but decided it couldn't be done without massive work and risk to the API integrity. Of course, they were after hard realtime. Mainstream Linux would benefit greatly from "as good or better than Windows" realtime; hard realtime is only truely necessary for certain classes of embedded apps.
Another thing they don't mention, but should, is the need for a reentrant kernel and real, actual, useful, and working threads in Linux. This goes hand in hand with well written realtime/near realtime applications. Being preemptable is a whole lot more useful when the multitasking is far more fine grained than you have in today's Linux. Also, kernel-mode threads (as you have in BeOS and NT) are particularly useful for realtime-sensitive data.
Ummm... not to be a Microsoft supporter or anything, but NT/Win2K, while not a hard realtime system, is dramatically better at functional realtime than any mainstream version of Linux. Not in the BeOS or QNX league, but good enough for things like hard core Digital Audio Workstation applications (when you have 64 audio tracks and 128 MIDI channels running low latency and glitch free under Linux, Linux will have a real chance in this market, but it's challenged enough today playing back a single MP3 without glitching).
However, the patent office is notoriously bad at investigating prior art. They basically just look at existing patents, not "what's out there in the real world". Second problem is that there's no requirement that the patent examiner of record actually be "skilled in the art". The PTO was issuing software patents for 10-15 years before they had any actual examiners with software engineering credentials, and seem to be even further over the edge on business practice patents.
Here's one heinous example I know about. I was a technical consultant to an employer's legal team going up against an IBM suit -- basically, they wanted access to all our technology, so they came after us with a stack of patents. One of these, filed in _1984_!, claimed to patent nothing more or less complex than cut and paste of text between multiple buffers in a text editor. Since our computer shipped with MicroEmacs, we violated this "invention". Never mind that the exact same sequence of keystrokes would have achieved the exact same effect in TECO Emacs, which I was using at CMU back in '79 (derived from earlier versions done at MIT...).
Sure, you can fight a patent with prior art, but the presumption is usually that the patent was researched fully. Clearly an increasingly bad assumption, but that's what you'll find in the system.
So yes, there is some danger here.
But there are issues. In some cases, we are bundling services with the system -- these are not yet something you would likely buy for home use if you already had a PC. In some installations, you're getting our box via another service provider, such as a cable TV company, and may very well be able to share the connection with a PC.
TiVo and more specific purpose units are, and sure, there's no really reason no to allow alternate hookups even there, except perhaps the support issue. They want to sell an appliance -- every unit is basically the same unit, and not offering an Ethernet hookup may save on the support side without really hurting their acceptance among advanced users. After all, I bought my TiVo anyway; I would have preferred an ethernet hookup.
Of course, TiVo may also be taking a shortcut and not running over a full TCP/IP link, which is another issue.
In time, this will change. The phone thing works for the first appliance, gets annoying for the second (I also have a satellite box in the same room), unmanageable much beyond that. As connected appliances grow, they will at first support and later demand some kind of home network. Hopefully, that'll be IP6 based with some choice in options, though there's certainly a chance the consumer electronics giants will embrace one thing (maybe bluetooth, maybe not) so completely that computer users and smaller manufacturers like me will have not choice but to follow suit.
We're entering the transition phase now -- give it a few years.
Back in 1985, Amiga hit so strongly in the sound and graphics areas because Jay was a genius, but also because no one else really tried to do it at such a level. Commodore could have made their own CPU, like in the 8-bit days, but it made no sense: you couldn't compete, in a market of maybe a million units a year, with a whole CPU family sold in the 10's-100's of millions.
The graphics companies today are like the CPU companies back then -- they're locked in a constantant battle of one-upmanship with each other, working on very complex designs which sell in the 10's of millions for pretty little cash. The only possible way Amiga or anyone else can compete is to do something very different. If your target is still the desktop, that would be a hard one, especially given all the graphics chip talent these guys have sucked up from the workstation world.
You don't think all this was motivated by a few kids downloading MP3s, do you? Sure, it's wrong, even if it isn't necessarily hurting anyone (I've used Napster twice; each time it cost me over $100 in CDs I just HAD to get after listening to the MP3.. but that's just me).
I have an article on ZDnet (so kill me) about this very thing: TalkBack Central: RIAA vs. Artist: Will the real victim please stand up. Basically, the RIAA represents the huge record companies, not the artists. And they're being forced to contend with an unpleasant fact: they're dinosaurs, and the mammals are looking about ready for an uprising.
Lemme 'splain. There was a time, long ago, when a record company was instrumental in the recording process. They found musicians, hooked them up with other musicians, hired songwriters, producers, etc. That still happens to an extent: groups like "N'Sync" were manufactured this way. But they're the exception these days.
So now enter computer technology, which lets the artists work in their own private home studios. This was inevitable from the days in the 60s when folks like the Beatles first used the studio as part of the creative process, not just a place to lay down a few tracks. I can get better quality in my home studio than "the lads" every managed (technical, folks -- the talent just ain't there).
So, what DO the record companies do? Primarily, they manufacture and distribute CDs and tapes on a huge scale. Their clout gets their artists in the stores and on the radio. They advance musicians money against eventual royalties for making record and videos, but rarely actually pay for that. Aside from the whole meatspace distribution, they're functional as an engine of transfer in just one more way: up to 95% of the profits made on a record goes into their pockets.
Now enter Napster. Napster has two important things, even if it's morally wrong. First off, it demonstrated a great peer to peer technology, bettered still by GNUtella and others. What this did is fundamental: it moved the problem of electronic distribution from necessarily being a huge server-based project (look at the size of eBay's rig, for example, and they're just dishing up pictures and text) to Just An Application. The second thing they did was on the culture. After Napster, there are millions of college kids, and others, who now think electronic distribution is a fine way to listen to music, not just a cool hack.
So the RIAA is fighting the future. Their client face a serious loss of power in a world in which artists and listeners can have a direct relationship. And really, find a way for those musicians to get paid 25-50 cents per download, and the record companies will have no more place in this world than the noble apatosaurus. They're fighthing Napster especially, because they currently have name-enough to start doing something on the up and up, which they couldn't fight.
That's not the only place they're fighting. You may have notice the current rise in producer-created bands like the aforementioned N'Sync. It's no coincidence -- a group built that may well be owned by the record companies. Contracts have been building up to give the record companies dramatically increasing power. Go buy that new Clarke/Baxter book, and you'll see their names on the copyright page. Go buy just about any CD, and you'll see (C)2000 Sony, Inc. or some-such.
And it gets worse. If you're a consultant, you may have noticed that a major "work for hire" bill passed a few years back, effectively making it harder for you to work for a company and still claim what you did for them is an independent work, and not theirs. Well, the RIAA & co. managed to sneak what's in effect a rider to that bill through, piggy-backed on some cable TV bill, which make "recording artist" just as subject to the new work-for-hire laws as "computer engineer". So unless they had particularly good lawyers a few years back, there's a fair chance than many of the major music acts out today are effectively just employees of their record companies. Which, if nothing else, lets the dinosaurs have serious control of much of today's popular music catalog. Making it all the more difficult for an independent, legal, and micropayment-based distribution system to gain momentum.
So the chance of life, as we know it, goes hand in hand with the discovery of water on a planet. But more to the point, if there's water there, reasonably accessable, we can stay there indefinitely. If not, we'll have no permanent presence with our level of technology.
The only way PPC systems would match x86 in price would be to have a large enough open systems market rivalling the PClone market. I was doing that in '97 when Apple pulled the plug. And while moving into the set-top box business from there has our stock some 20x higher than in the Maclone days, I'm still bitter. Not just because Apple proved to be the dicks everyone kept telling me they'd be (no, they've really changed, I met with the CHRP group myself), but simply that the PPC's good enough. It deserves success.
After that fell blow, neither Morotola nor IBM has felt particularly compelled to race with Intel just for Apple's sake. When phone or internet switches or other high-end embedded gear need an upgrade, they'll crank out a faster chip (like IBM's forthcoming PPC750 update, starting out at 700MHz). With all that copper and SOI and all, it's not like IBM couldn't deliver if they wanted to.
Love the quote, man -- and I don't come to /. for egosurfing (though at least I'm on a genuine AMD/Linux machine this time). So far, I think the relative improvements from Pentium to Pentium are damn near logarithmic. If they really changed the microarchitecture for the first time since the "pro", maybe this should be, I dunno, the Pentium 6 or something. Or maybe it actually will be a yawnner by then, and is appropriately named.
For one, the "game console" was never the intention of the Hi-Toro/Amiga gang. It was simply a way to attract venture capital, because in '82/'83, consoles were a big deal.
Metacomco didn't create AmigaOS. Most of the AmigaOS was created in-house. The DOS subsystem, one fairly small piece of the whole OS, was not done being completed in time, so Tim King and company ported the DOS subsystem from TriPos (sort of the UK's answer to UNIX) to the existing Amiga OS.
There NEVER was a real Amiga 1500. The thing sold as an Amiga 1500, only in the UK, was simply a bog standard Amiga 2000 bundled with two floppy disc drives, rather then usual single floppy or floppy and hard drive.
The Amiga 1200 and Amiga 4000 shipped in the fall of 1992, a full year-and-a-half before Commodore bit the big one. They were actually somewhat bastardized versions of things that had been in the works in 1991, with the intention of shipping in early '92. Commodore had gone though a management change, and their original idea was to drag feet on the Amiga, in an attempt to kill it, while bolstering the PClone program. The flaw in that formula, as any first year business school dropout, you, me, and my cat could tell you, was the simple fact Commodore made virtually all their money on the Amiga. By the end of '92, all of the PC engineers had left or been assigned to Amiga projects. The VP of Engineering responsible for these screwups had also left, but the root problems in upper managements remained, and led directly to the end of Commodore.
It looks like technical problems in retrospect, but basically, when managers at a company like Commodore (peaked at about 1 billion in sales, and there may have been some monkey business even in those days) take higher salaries than the bossmen at Apple, IBM, and Compaq, and spend less on engineering than plain old PC companies (who often do no actual engineering, period), well, you just can't maintain technological advances, even with some of the best engineering people in the business.
The New Amiga is really only Amiga in spirit, if that -- they have yet to show it, one way or another. I'll with-hold judgement until they're actually doing something Amiga-like.
-Dave (systems engineer on A2000, A2500, Zorro III, A3000, etc)
The Linux connection (strongly hinted at by Jim Collas, former President of Gateway's Amiga division, when I spoke to him about the various AmigaNG rumors) is, of course, Linus himself. But the coolness factor is this: here we have a CPU, running emulated instructions as fast as some version of the Pentium II/III, and doing it cheaper and with much less power. Only the first one is new work -- it's easy to find CPUs that run faster than PII/IIIs using less power (PowerPC, for example). The thing is, the x86 code runs at 1/2 or so of the native Transmeta CPU core speed, perhaps based on the limits of their dynamic recompiler, the loss of abstraction in binary code, etc. What about native code?
The modern trick in all this, same thing Sun's doing with MAJC, is to make the idea of a VLIW processor legit by never tying a system to native binaries. You run x86 or Java Byte Code or whatever through a translator, and when the machine architecture, and thus instruction set, change, you build a new translator, everything's hunky-dorey code wise, and you get to forget entirely about hardware legacy. But Linux and other open source stuff doesn't have a binary legacy problem, period. So there's no problem in coding Linux native, and if you did this, Linux would run twice as fast, relatively, as any closed source OS on this platform. The need for Linus on this would be building a mixed code manager for Linux, so that the kernel could cooperate actively with the dynamic translator and run your choice of x86, native, or other code modules (full-speed Java, etc).
This is the last layer of HW abstration. Some day, all CPUs may be built this way.
The most visible of these moves is that they have dessimated their customer/tech support group. In restricting their products to Intel-only, it's likely they're shipping pretty much the same thing in every box, CPU/motherboard wise, with all motherboards made by Intel. This will definitely cut support costs, simply because support staff needs only know one or two PCB variations.
It's also a retreat back to what Gateway's confortable with. They were one of the last major companies to start using non-Intel products, and they're one of those with the least internal development around (which they have also cut back recently).
This isn't to say that fear of Intel, or even Intel themselves, hasn't hurt the Athlon introduction some. But that's not the only explanation; the Super7 introductions were plauged with chip and software issues for some time. Some vendors will want to believe that's not an Athlon problem before they jump on board.
It's also a standard cost-benefit thing; is Athlon compelling enough to jump now versus later, or as a flagship, or whatever. Maybe? If the Transmeta fulfills its design goal of delivering a 4x cost/power advantage over some comparable state-of-the-art chip, it'll catch on fast. Especially on laptops, which, being fairly well self-contained and made by larger companies, are less succeptable to fear of Intel.
The Saturn was technically resonable, spec-wise. But it had major problems, and that's why it was so easily curshed by Sony. For one, the dual H2 processor setup was hard for programmers no used to writing games on a multiprocessor system. And the machine was just too fat -- way too much hardware in there for what it did. As well as the dual H2s, there was an H1, a 68000, and a Yamaha DSP chip (the 68K, Yamaha, and 512K of private memory did their MIDI/Audio section, usually a tiny piece of one custom chip in the other machines). They simply couldn't compete on price with Sony or Nintendo. Look inside an N64 -- four main chips. The Saturn had more CPUs than that!
Naturally, marketing made a big difference, but that's only part of it. For one, you're marking primarily to kids, which for the most part means specs don't matter a great deal. This is the main reason companies follow the very successful Nintendo formula and establish recurring characters (practically every game written by Nintendo is based around characters who are from the original arcarde Donkey Kong, or characters introduced in earlier spinoffs).
Nintendo fell behind in that generation primarily by being a year late. Sega couldn't get software written fast enough to keep up their initial marketing push. They have learned from most of their mistakes this time around, and as well have a ~16 month lead over Nintendo and Sony, at least in the USA.
But I do agree -- you can take this market with th e right product at the right time sold the right way. Part of that is having zillions of dollars to put behind it. Back when the N64 was launched, they spent around $50 million just on the Christmas ad campaign. Today, it's probably more than double that. Microsoft can certainly play in that sandbox.
Sega, Sony, Nintendo, etc. are selling game consoles that range from fairly well (Sega) to highly (the others) integrated systems, with customized CPUs, etc. This is how these machines sell for $99.95-$199.95 today, which is rarely if every making any of them a dime in profit. Even with high integration, you're not making a PClone-ish system any cheaper. And in fact, the emerging specs make it sound much more expensive, but we'll see -- they could and very well might roll special versions of the chips, to keep them out of PCs.
So you have to follow the money. Money is made on software. Sony makes about 1/4 of their total profits on PSX software, even though total income from that market is fairly small compared to the many other businesses they have. Microsoft no doubt wants to make money on the software, and in order for YOU, as a developer, to pay MS for the luxery of supporting the X-Box with a game, you have to license something from Microsoft, or they have no hold over you. For the 3DO box, you had to licence the OS, which was always loaded from disc (probably what MS will do; this lowers the HW costs and eliminates OS compatibility issues). At Commodore, the CDTV and CD-32 had the OS in ROM, and simply required a C= Copyrighted file at a certain location on the CD.
That's not to say you couldn't hack it, just that such a hack wouldn't be legal to distribute.
Of course, if MS is the only one making money, this won't fly. They could adopt a WebTV model -- WebTV is sold roughly at cost, but Microsoft pays vendors a bonus for every unit sold, so they do make a profit (if they sell them, of course). However, this is illegal in many countries. Clearly, vendors can't compete if they have to make a profit on the box, that's what killed the 3DO box, even before Sony and Nintendo came along to bury the body.
MS could manufacture the box themselves, but they don't usually like to get involved in expensive hardware. Look at all the folks who have lost money on WinCE handheld machines -- Microsoft made money in every case, because they sold the software. I suspect they'll want the same thing now.
The X-Box should be well supported for a few reasons. For one, this is PC hardware as PC authors know it, more or less. They'll have to license the Windows or WinCE or whatever you put on disc, but it'll be pretty close to what they're doing in Win9x today. The CPU won't necessarily be fast by PC standards, but then again, the target is the 720x512 or less space of a TV set, not your 1280x1024 monitor. It'll be easy to port Windows-only games to this platform, and quite a few of the older games, not selling for the regular PC, could be recycled here.
On the downside, if you have to use Windows, the system won't kick butt at the same level as the Sega (which can run WinCE, but rarely does) or new Sony/Nintendo machines. Looks aren't everything, though, and much of this market is driven by hype and clever marketing, not just graphics horsepower. Microsoft is a master of this, so far.
Uh, that was supposed to be $10K-$25K. I could swing the $10-$25 opening minimum myself :-)
Now, there are two ways to go about this. If you're a criminal type, of course, you don't much care what the US government thinks. And you're probably not getting too many of those drug, prostitution, and illicit gambling charges on VISA, these are in cash. Which, of course, is smuggled out of the country hidden in dolls or coffins or whatever, at least according to most of the gangster movies I've seen.
The other thing is to actually become a citizen of a tax haven and remain a non-resident of other countries. An actual case of this is Irving Gould, onetime principle owner and Chairman of the Board at Commodore. At some point in time, he moved to the Bahamas, real house an everything, and I assume became a citizen. But he had to be careful about being percieved as a resident of any country. So he actually had people who kept track of how many days he spent in any given country. For example, it was not uncommon for the PET Jet to circle until after 12AM, when old "Uncle Irv" was coming into the US, just so that fraction-of-a-day couldn't be held against him in US tax court. In other words, maybe a bit shady, but all perfectly legal.
I don't supposed you would do this until you're making enough money to have that kind of lifestyle, but that kind of goes without saying. Most offshore accounts offer little interest, charge high fees for transactions, require opening deposits in the $10-$25 range, etc. It's not likely to attract the typical working stiff, even if you're a contractor or some-such getting lots of payments in cash. On the other hand, a millionaire with lots of foreign income and fewer scruples could well go for such a thing, profitably.
It'll probably only get easier, with the e-banking and e-trading. There are already US-based companies specializing in setting up foreign investors in US stocks and bonds, and it probably won't be long until US folks have the same online shot at some foreign securities. Now, when you're doing the electronic transfer to your bank in New Jersey for that killing you just made on the Newer Markt, ole Uncle Sam will be there to congradulate you. But all you'll hear from that Caymans' bank is the $50 (or whatever) deposit fee.
Which just goes to illustrate the old point that the rich do, in fact, get richer.
1) room for much more of the expensive L2 cache for full speed operation and modules up to 2MB.
2) redesign to the AGTL+ bus, which allows four processors to run, even using modules (versus the normal P6 bus, delivering two way SMP with the module, four way with the older socket).
3) more money. Intel knows that people who need 4-way or better SMP systems will pay for this, often to a foolish extent. Intel loves to milk some sector for high margins, and you can guess this won't happen when there's a direct (or near so) replacement like K6 or Athlon.
Ok, so now consider PIII Xeon with only 256KB of L2 cache. Certainly, this is the same chip as you get when you buy a PIII-regular. But of course, it's on the Slot 2 module, and if you want a four processor (or better) SMP system, you have no other choice. Intel basically has you, and they like it that way.
In fact, I'm surprised it only around $50.
Ok, first of all, forget completely about "Slot". That's just a hookup, a physical connector, it doesn't tell very much of the story.
Slot 1 is the second physical delivery of Intel's P6 bus; Socket 8 (for Pentium Pro) was the first; Socket 370 is the third, and there's every indication Intel will market a Socket 426-or-thereabouts to handle some extra pins on a socket-based Coppermine chip. The reason you can't clone this without a license is simple: Intel has patents on the P6 bus. They originally didn't license, but now they are licensing, at least to chipsets (SiS, ALi, VIA) if not CPUs.
AMD extended Socket 7 (which is once of the physical conventions for the non-patented P5 bus) to 100MHz, and ran their K6 family there. But they realized that K6 would have its limits. In part of their IP settlement with Intel, they promised not to make clones of things like the P6 bus, so legalities for everyone else aside, that wasn't an issue.
Enter DEC (now part of Compaq). DEC designed the EV6 bus for the Alpha 21264, as a remedy for conventional CPU buses. In their earlier Alpha systems, even with L1..3 cache and all, they were so dependent on memory speed, the typical bus sharing in an SMP system (very important to DEC) was a problem. Especially when you wanted to make this bus (the cricital CPU to rest-of-the-world link) extremely fast. EV6 is a point-to-point bus; all you have is a CPU and some system chip on it, never anything else. This allows them to run EV6 very fast. DEC openly licenses EV6, AMD adopted it for their CPU. Since the PC market demands a socket or slot, they created Slot A, which (for cost reasons) uses the Slot 1 connector turned around. AMD runs Slot A at 200MHz now, 266MHz in the forseeable future. DEC runs EV6 systems up to 400MHz. This is a data rate; the bus runs DDR (a new data event on every clock edge, not just every clock cycle), which really doesn't matter; the data rate is indicative of performance.
The important thing to realize about today's Athlon systems is that they're something like the first Pentium systems shipped out, retrofitted '486 systems much as the first Athlons are running based on modified Super7 chipsets. So memory is stuck at 100MHz, half the CPU interconnect's speed. You won't see Athlon reach its performance potential until 200MHz, or better, memory systems are delivered.
And I do mean "or better" because of the EV6 architecture, there's no shared CPU bus. So a system chip (North Bridge) can actually use memory faster than the CPU can deal with. With switched interconnects, proper buffering, and super fast memory, one could imagine PCI, AGPx4, and multiple Athlons all rompin' along, each at virtually full speed. This can't be done efficiently within the current P6 architecture. That's why I find Athlon interesting. I hope AMD lasts long enough for it to realize its potential.
In essence, I agree. People want this for fun. I still use a number of Canon and Leica rangefinder cameras, even in this day of digitals and SLRs smarter than any human. Because they're fun, because there's a degree of excellence there not available in today's "wonderbricks". AmigaOS is very much the same way.
Your details are wrong, though. AmigaOS was NOT "based on Tripos". The AmigaOS was largely original work. In order to get to market in time, Amiga had the DOS subsystem (a tiny piece of the greater whole) from Tripos ported to the existing AmigaOS. That's the only Tripos connection.
The chipset isn't particularly interesting anymore, in this day of graphics chips running 600x faster than poor old Agnus/Alice, Paula, and Denise/Lisa, audio chips with more DSP performance than the fastest Amiga ever made, etc. It's important to know that the simulataneous screen resolution display was kind of hack, or chip limitation. Yes, there was a nice display list coprocessor that switched the screens. But the reason you could mix resolutions was the fact that the pixel clock was fixed at around 35ns. So every display ran 35ns, 70ns, or 140ns pixels. In a modern graphics card, you're mucking around with a PLL to independently vary refresh and resolution, and PLLs inherently take a more-than-scanline while to sync up to new frequencies. The cancelled Amiga Advanced Architecture project supported this trick by allowing up to four independent PLLs to exist in a system. But sure -- you can make an Amiga card for the PC. This was actually proposed twice while I was at Commodore, the first time it even made some sense (I did an engineering analysis). In practice? Look here: http://www.siamese.co.uk/pci-amiga-crew.html. The point, really, is "why bother". If you run an Amiga emulator on a reasonably fast PC, using the graphics API targeted to Windows or X, rather than Amiga chip emulation, you'll have an Amiga faster than any one that has ever existed in the real world. The OS, on the other hand, is still pretty damn cool. Some of you may recall a time when hacking was for fun, not profit. I think lots of people would love to hack the AmigaOS. It would be good, really, for an industry so caught up in ancient architectures like Windows and UNIX, to learn what AmigaOS has to teach. It's probably too late to make it anything that takes the world by storm, but that doesn't render it useless. In fact, most of the lessons of AmigaOS seem to have not made it into the PC mainsteam, whereas all of the hardware tricks did long ago, and then were soundly exceeded. As far as making new chips, a really sharp guy called Mick Tinker claims to have reimplemented the Amiga chips; don't know if he's using an FPGA or a true gate array (or possibly just simulating in VHDL or something). Of course, there's some extensions in there too; Amiga chips as they were implemeneted aren't terribly interesting today.