I'm curious whether or not the higher-ups at VA Linux think it's possible to develop the company enough to be worth its stratospheric valuation.
Obviously, a high valuation is partially a self-fulfilling prophecy, because it gives that company the ability to buy others with its high-priced stock. But I would be interested in any other strategies Mr. Augustin has in mind to develop VA Linux into the enormously valuable company the stock price suggests, and how it will affect its relationship with the Open Source world.
I suspect that creating vast companies like Microsoft, based on Linux, is simply impossible. An enormous service-based organization might be possible, but it seems to me that it would take many years to build one. The GPL is designed to prevent the monopoly, hardball shenanigans that Microsoft pulls. Can building hardware and providing expertise instead generate enough revenue?
For years, it's been impossible to conduct anything like a rational public -policy discussion about the dissemination of sexual information[....]
It's been impossible for years to have anything like a rational public-policy discussion about anything, much less anything controversial, which you have eloquently pointed out in prior essays. The sex-repression crap is just the worst example of a serious problem. We do not talk to each other anymore.
You know, the closest thing I've seen to a public policy discussion lately is probably Slashdot. I'm both mildly amazed and mildly appalled. This is the best it gets???
Here, we are libertarian geeks preaching to the converted. I wonder how much of an effect these discussions can have on a wider world?
Does anyone but geeks really read these forums?
And when was the last time you changed your mind about a political view based on a discussion you had online?
"Most historic moment of our lifetime"... bah humbug. History is something that happened; a date is just a date.
It was just another day. There really wasn't anything special about that particular day. The only extra value it had was what we humans invested it with. It happens to be (roughly) the second millenium since the time when we believe a particular religious figure was born. It's not even especially accurate.
Tonight, when you go home and see your loved ones, it's just as important to love them and be with them as it was 4 days ago. Tonight is no less important than the Millenium was. The calendar is something we made up; it's a fantasy that most of us happen to agree about. The simple fact that everyone bought into it doesn't make it any less a fantasy. The media hype is just herdthink.
It's not worth that much heartache and pain for something that is so imaginary.
I don't know about you, but with my mid-month paycheck I'm going to go make an earthquake kit. I figure emergency supplies should be the cheapest they'll be for another 100 years.:)
I suspect that the hype prevented the disaster
on
Apocalypse Not
·
· Score: 1
The hype strikes me, now, as a negative-fulfillment feedback loop. It's a lot like the virus scares. In general (not always), the more noise you hear about a given virus BEFORE it goes off, the less damage it will do.
The really devastating failures are the ones that take you by surprise. If we hadn't done anything -- if everyone had poopoohed the idea of a Y2K failure, it's likely that there would have been catastrophic failures, and people would have died.
However, as I was telling people all the way along, the distribution networks wouldn't break down. If necessary, most places can still do things on paper if they have to. There's no way that Safeway (a grocery store out West here) is going to let their shelves get bare. It's not like the goods and products suddenly ceased to be manufactured. And if manufacture of some goods were to stop for a few days, it wouldn't create more than a ripple -- a bit of economic damage, perhaps, but nobody was going to starve to death. The one intelligent thing I heard suggested was to stockpile 10 gallons of water per person, as there were some last-minute worries about the water supply. Water is something you need NOW. Anyone can live without food for a week. It's a bit uncomfortable, but you won't die.
That said, the Y2037 bug in Unix has me worried. In 37 years, there won't be very many people left that really understand the fundamental architectures of today's machines, and you can bet that plenty of today's machines will still be in service, faithfully performing the jobs they were designed to do. By then, our networks will be very very complex, interrelated to an astonishing degree, and individual failures will be able to have far greater ripple effects than they can today. Y2037 might well be a disaster. It certainly wouldn't destroy civilization (unless it was tottering on the brink to begin with), but it could take years to recover from the systems crashes.
2037, for those of you who are non-Unix folks, is when the Unix counter for time overflows on 32-bit machines. Unix machines measure time in seconds elapsed since 1970, using a 32-bit counter, meaning they can only go to 4 billion seconds or so. 2037 is when 4 billion seconds will have passed, and an awful lot of today's machines are going to be VERY confused at that rollover. 64-bit machines shouldn't be affected for 4 billion times longer than that -- in other words, by using 64 bits to track seconds with, a machine could outlive our Sun without getting confused about the date. Just shooting from the hip, I think 64 bits' worth of seconds ought to be about 240 billion years. That's quite awhile.:)
Personally, if I am still alive, I plan to come out of retirement in about 2034 to start fixing all those old machines, and I plan to make a bloody mint.:)
"In 1967, you thought we were a bunch of clueless propeller heads. In 1997, you learned better.
"Y2K Consultants, Inc: It's Time to Pay Up (tm)"
(stolen from Entrepeneur, an older-but-fun game at www.stardock.com.)
Folks, not all downtime is created equal. If you are down at 12:01 on New Years' Eve, a holiday, and Saturday... big deal. Nobody is shopping then anyway.
I'd rather be down for awhile early Saturday morning than at 6PM Monday evening.
Planned downtime is a lot better than unplanned downtime. All the assertions I've seen to the contrary have so far been completely stupid.
Boy, it's nice to see people with actual experience posting on this topic. It amazes me that more people don't flame Bruce et al for such uneducated opinions. It is utterly obvious that they have very little real-world experience in actually making systems work every day, being in the hot seat if they fail. But they still Make Pronouncements From On High as if they were experts in the field.
Argh. Readers, take anything you see on slashdot with a whole container of salt: there are real experts here, and there are those who claim to be, and differentiating them without significant experience in the field yourself can be difficult. Repeating as gospel anything you read here is very foolish -- check, double-check, and triple-check before you take anything on/. as Revealed Truth. High moderation points are often just mob mentality, and have little bearing on actual expertise or good thinking.
I'll just make one comment: any attempt to divide any group of people into two groups, based on some fantasy projection like "how they approach technology" (as opposed to provable things like male/female) is irrational.
Actually, that's not quite true. It's hyper-rational. It is using the rational mind's ability to divide, pattern-match, and differentiate in a way that does not fit the tool well.
Pointing the rational mind at people and trying to explain their motivations is highly similar to pointing a shotgun at a butterfly to understand it. It causes similar problems in both cases.
I haven't used Novell servers heavily since NW 3.12, but one thing I always appreciated about Netware was that they really thought their problems through.
It often took awhile to wrap your head around why they solved a given problem in a given way, but once you understood it, it was like lightning striking: of COURSE you would model this problem that way, because then you can do X, Y, and Z easily. I found this to be true over and over, wherever I looked. Every solution was arcane but incredibly powerful once you understood it.
In comparison, NT has been quite disappointing in this area. It's got a shiny graphical interface, but given the choice between ease of use and really solving a problem, Microsoft seems to opt for ease of use. On servers, I don't mind arcane syntax and weird models if they are good ones: I'm a paid professional, I can take the time to learn it if the job gets done better.
Linux and Netware both share that last trait in spades -- but comparing linux-of-now with Netware-in-3.12, on the whole I'd call Netware's solutions superior. They're not really in the same problem space, but I haven't often been impressed with the clarity of thinking about real-world problems with Linux solutions. It can almost always be made to do what you want, and is undoubtedly more flexible than Netware, but Netware 3.12 was in my opinion a better file-and-print server than Linux is today. Linux is a far better application server and general-purpose computing platform, assuming you're willing to invest the time you will need in it.
I haven't used NW5, and I presume it's quite a lot better -- they've had, what, six more years to work on it?:) If they have thought their other services through as well as they nailed file-and-print in 3.12, I'd consider it a very, very strong contender for your modern network.
Did it ever occur to you that, for many solutions, Microsoft software is a perfectly valid, acceptable choice?
In our company, they implemented NT because there wasn't much expertise in-house at the time. They eventually brought in an IS staff, but the core system design choices had already been made.
And, ya know what? If you have experienced, savvy admins, you can run a nicely stable NT network. Ours is very good. Admittedly, on average, our NT servers seem to need a reboot about once every two to three months, but that's acceptable at this point.
Now, I would prefer to be deploying Unix, and we plan to gradually move toward Unixy solutions (not too much Linux yet, we don't trust ext2 enough to put critical data on it), as we plan to scale some of our system services past what NT boxes can handle. This does not invalidate the original choice of NT which, given the design constraints, was really quite good. Three years ago, when the fundamental work was being done, Linux wasn't ready yet, Novell didn't do Web stuff very well, and the 'real' Unixes required more expertise than was available.
Your blanket assertions that IS staff who wanted to shut down over New Years' are incompetent, and then above that any IS manager who implements Microsoft software must be doing it 'because everyone else does it', are false. Egregiously so. You are showing remarkably poor thinking here.
I would suggest that you contain your blanket assertions about why people do things and their competence level. It is possible to make a decision completely different than the one you would have made and still be competent. Believe it or not, there are actually people in the world who know more than you do about how to make systems work.
In the real world, the IS Director's job is to get the job done. It is not to conform to some standard of ideals. If getting the job done requires standing on one's head and counting backward from 100 thrice daily, then that is what you do. IS is there to make people's lives easy, not difficult. Imposing artificial, intellectual standards about what software is to be used is a good way of putting one's company out of business.
Suffice it to say I would never in a million years hire you to run my network. I don't need evangelists -- I need brilliant people who find the best solution for the job at hand. Sometimes that's free software, and sometimes it isn't.
Netware is just as stable as Linux, maybe even more so. It's a rock-solid system. Its interface is arcane, but certainly no harder than UNIX.
You have a very good solution sitting there under your nose -- don't disregard it because lots of people like Linux. Netware has been around awhile and is really, really good. Apparently it did have some Y2K issues, but, hey... nobody's perfect.:)
Depends on what it costs you to be down. In our case, very little -- maybe a touch of prestige. What business-related email is coming in at 12:01 on January 1, 2000?
I consider this part of my job not to be one of prestige, but of avoiding risk. If I take a needless risk with my company's equipment, data, or even my time merely to satisfy my own vanity, I am doing them a disservice.
Vanity I save for new implementations and new projects -- for administration and upkeep, I'm perfectly willing to lose some prestige if that's what the job requires.
(If you didn't see my other comments on this topic, we did leave our systems up, FYI... seemed the least risky option.)
In the real world, where most of us live, there is a lot of Microsoft software. It has not been shown to be especially reliable, and I can't look at the code or hire someone else to look at the code for me. I have no idea what bugs lurk there -- and I don't think Microsoft does either, to be frankly honest.
Personally, I was in favor of taking our systems down overnight, simply to prevent date subtraction bugs. Someone else pointed out that this was making a change right before a major event, and that this probably wouldn't be wise -- a compelling argument, to which I acceded.
Regardless, claiming that I am somehow incompetent because I wanted to shut down systems over NY is flat stupid. Computers are not magic; they are highly predictable devices. However, the software that runs on a large fraction of them is not well understood by anyone. Trusting it unconditionally is foolish.
Consider that Microsoft was still releasing patches as of December 15.
Strikes me that you have a mighty strong opinion about how to run large networks, when it appears your expertise is not in that area. From what I can tell, you are a programmer, and a very good one. That's wonderful, but does not qualify you to make pronouncements about system administration. You probably don't deal, every day, with the stupid bugs and problems caused by unforeseen interactions in closed-source software. You live in a tightly controlled world of your own code. I don't have that luxury.
I don't presume to tell you how to do your job, and expect the same respect in return. And it strikes me that making public pronouncements on the competence of people working, every day, in an area you don't is not just arrogant, it's foolhardy.
You can trust I won't value your opinion as much in the future.
I work for an ecommerce-ish company. We debated about it somewhat heatedly. I was in favor of disconnecting and shutting everything down, because any likely corruption bugs are probably going to hit at the rollover. The most likely equations to break are the ones involved with subtracting some earlier date from now; with the numbers on both sides of the rollover, that's where problems are most likely.
However, after we talked about it further, we decided that the risk in shutting down would be greater. We are leaving things alone, running exactly as usual, so we know any problems are Y2K related.. If we were to make any significant changes and then had trouble, we wouldn't necessarily know where to look.
That argument was compelling enough to me to change my mind and agree to it; we'll be watching to see what happens. We're really not expecting anything major, but we definitely are expecting minor glitches. And we are planning to continue to devote some time to Y2K issues for the next six months or so -- we figure they will keep cropping up for awhile.
We are under no illusions that Jan 1, 2000 is the miraculous End of All Date Bugs.:)
You probably don't want to do RAMBUS; it's expensive and doesn't really offer all that much, performance-wise. Wait a few months for DDR SDRAM -- you'll get similar performance for a much, much lower price.
I do not think you can buy, *anywhere*, a motherboard that will take 1GB of RAMBUS anyway. The design spec is for 3 RIMM slots, providing 768MB, but that nasty bug in the chipset has reduced most boards to 2 slots, or 512MB. I do not know if the chipset can handle chips of greater than 256MB. Even if it can, you will pay such an enormous amount of money for them that you might as well buy Octanes anyway. (256MB chips are running about $12 PER MEGABYTE -- $3000 EACH.) I shudder to think of what a 512MB chip would cost.
If you need to buy right away, you might actually check to see what SGI can offer you in the high end. They're doing a lot of good work with Linux. If anyone can set you up with a seriously kick-ass machine, they should be able to. They have recently released some patches to the kernel that will allow database access directly to the hard drives, for speed -- I would think that would be a likely big win for a high-end, data-intensive environment, assuming you write your own, custom software.
This guy is absolutely wrong, and he's probably knowingly wrong. (ie, fraudulent).
I've seen this kind of thing many times. After awhile you start to spot the frauds. Consider:
1. He's appealing to the 'underdog' effect. "Yeah, scientists everywhere hate me, and I'm going to prove them all wrong!" This elicits sympathy. 2. He is promising way, way too much. Any one of these inventions would make him a wealthy man. He wouldn't need investors the way he has gotten them; he would make a material sample, show it to some big corporation, and get bought for a couple hundred mil.
Consider: he's promising about a dozen revolutionary advances all at the same time, for one low, low monthly payment. It's just not likely that he would simultaneously overthrow all of established physics in so many areas at once.
There may very well be breakthroughs of this magnitude, but they won't happen like this: all promises up front with no delivery. Instead, a scientist will announce a revolutionary discovery and probably will make a whole lot of money. Then, a lot of OTHER people will make further progress based on the original breakthrough, and they will make tons of money too.
There's not enough wattage in anyone's head to throw back the frontiers of science in so many directions simultaneously. If this were real, his head would be totally wrapped around making the energy release work. That alone would make him enough money to buy small countries. The only reason to make claims in so many areas at once is to get investment money. I can't imagine of a surer way of showing that his basic breakthrough -- the power generation -- has no substance. If it were real, he would already be demonstrating a machine that worked.
This reminds me almost exactly of a claim by a company that was local to where I lived, about ten years ago. They claimed to have sped up the AT compatible machines of the day by a factor of 100 using off-the-shelf components.. Of course, it was a fraud, but it had a lot of people excited and I believe the owner made a lot of money. I don't know what happened to him -- hopefully he is in jail. Intel spent many billions to make computers 100 times faster than an AT. If it could have been done with off-the-shelf components, you can bet Intel would have done that. They're not stupid.
As a culture, we like to believe in the myth of a single person seeing the brilliant insight that nobody else is capable of seeing. But the stuff he's talking about here is too basic. There's no way that millions of scientists missed it. Just like the AT-machine that was 100 times faster, this 'invention' will disappear, and most of the investors will lose their money.
The poster's comment above about using UAE as a virtual machine actually has some real merit. The Amiga is a very clean environment, and you can emulate one very nicely on a P2/400 or so.
Even if everyone thinks using UAE is stupid, I think it's a good proof of concept. The Amiga is a ridiculously complex design, amazingly hard to emulate, and yet the developers of UAE have managed to get an emulation running pretty damn fast. At one time, it was widely believed to be an impossible feat.
If they can make a virtual Amiga run as well as it does, it seems likely that a clean, simple virtual machine design would run very quickly indeed. The 68000-series architecture wouldn't be a bad choice. Even a virtual Atari ST might work pretty well, if that abomination of a desktop was replaced with something else.
After thinking about this a few minutes, I think the Amiga would be a better choice. Strip out the custom chip emulation layers and do a pure-CPU Amiga OS in a virtual machine. It wouldn't have memory protection, but wouldn't need it; if you want to run more than one VM program, start more VMs. It would be multithreaded and highly efficient. The total footprint of the device would probably be under 2MB... including 512K of RAM for the Amiga OS to work with.
It seems like such an obvious solution that I'm wondering why it hasn't been done. Ideas?
Oops! Yes, you are right, I forgot about the ROM issue. I have both an A500 and an A2000 sitting in my closet (broken), so ROMs weren't an issue for me.
If you are going to emulate the Amiga, you should probably purchase the ROMS, at least if you want to be legal. I also encourage you to register WinUAE(I did!).
However, the people in the Amiga community moaning and griping about sites like Lazarus are just dumb. The Amiga is *dead*. It won't come back. It is just too outmoded to be worth salvaging anymore. Might as well put up those old programs so that everyone can see them -- if some of those useful old ideas are re-introduced into computing, we will all benefit.
Obviously that doesn't include the (12?) or so programs that people are still trying to make money from... but fer chrissake, who cares if someone copies Lemmings 1 anymore? *sigh*
Agreed wholeheartedly. I actually met Carl Sassenrath, the guy who wrote Exec, at a party once. Really nice guy. As I mentioned sideways to another Amiga person there, "Wow. What do you say to a demigod?":-)
IIRC, in 1994, a "good" machine would have been something along the lines of a 486-66; I don't think the Pentium shipped until 1995. At the time, it would have cost around $3,000, fully loaded.
I think you're right that the comparison is a bit specious. A 1985 A1000 cost about what a 486-66 would have cost in 1994. However, I didn't buy the A1000, my parents did. My first machine was an A500, and I eventually scraped up the $1200 or so to get into an A2000. Even by 1994, my income hadn't gone up that much (was still young), and there's no way I could have saved $3K for a 486-66. Realistically, a span of about 7 years is fairer: 1987 to 1994, comparing a cost range of $500 or so.
Even if I had had $3k in 1994, a 486-66 still wouldn't be as fast as an Amiga for multimedia apps, at least under Windows 3.1. I suspect that a custom program written in DOS on one of those machines could have come pretty close. It wouldn't have had four-channel sound, but you can fake that with 16-bit sound. (Amiga has 4 8-bit digital channels.) And the 256-color modes really did look a *lot* nicer than the standard 16-color Amiga hi-res palette. For still pictures and strategy games, the PC was much better than the Amiga in 1994 -- but it still wasn't as fast at animation.
Another way to look at it: In 1985, a PC that cost around the same amount as an Amiga would have been an 8Mhz 286 (2nd generation AT), with 16-color CGA graphics, and usually 128K of RAM. That may give you a better idea of how far ahead the Amiga really was.
And you are right that there was no MMU protection in Amiga's multitasking. This made the machine very unstable indeed for the first year or so. By the time the A500 shipped, people were getting pretty good with the very new concepts in the OS, like message-passing, memory pointers, and semaphores. A single bad program could take down the whole machine, but that's also true of every consumer OS until NT. (and Linux, of course.) Because of this, the overall programming quality was very high. Everyone knew their program could crash your computer easily, so everyone was really careful not to do that.
Overall system stability with the Amiga was at least as good as with Win98, even without hardware MMU support. And it was blazing fast; it could task switch in something like 50 clock cycles. Motorola actually used the core of the Amiga OS, Exec, as an example of how to do fast task switching on the 68000 chip.
A hardware MMU would probably have been 'better', but in actual practice it didn't seem to matter too much once the programmers got up to speed.
Amiga was developed by a company that thought they were making the next great console. Their engineers, probably the finest design team that had ever been assembled to that point, had other ideas: they were going to make it the ultimate personal computer. Most of their development was done semi-secretly -- the people bankrolling the initial project didn't realize the feature set had expanded so enormously.
Obviously, I'm not up on all the details here, but that original seed company ran into financial trouble. Atari lent them money, knowing that they probably wouldn't be able to pay it back and they would get this incredible technology very cheaply when the little company defaulted. (I don't know if the little company was called Amiga or not.)
Well, sure enough, they hadn't quite completed their design and the loan was due -- Atari was about to move in and shut them down. Commodore rode in as a white knight and bought the company lock, stock, and barrel. They repaid Atari and took the technology for themselves. This infuriated Jack Tramiel of Atari.
To get revenge, Atari slapped together what was essentially a cheap piece of shit that, on paper, appeared to have many of the same technical merits the Amiga did. In actual fact, it was a 68000 with some RAM, I/O ports, and very simple graphic and sound chips. TOS was their single-tasking operating system. The desktop was called GEM, and was ugly and sluggish. But it had a 68000, like the Amiga, and it actually clocked the CPU faster (8Mhz) instead of the Amiga's 7.14Mhz. It also included a built-in MIDI port, which is probably the only really cool thing about it.
The Atari STs had three big advantages over the Amiga. They were cheaper, were easier to program, and were backed by a big, sleazy advertising campaign. They ran 'comparison' ads all the time which were horribly slanted. They actually did pretty well with it for quite some time. The infighting between the Amiga and ST weakened both computers, and eventually destroyed both: the PC juggernaut overwhelmed them, though the Amiga lasted a lot longer. The Amiga/Atari wars were worse than any of the distro wars you see now. But ultimately, while the Atari died a lot sooner, they both still died.
It took ten years, however, for the PC to catch up to the Amiga completely. Out of the box, it could do 4,096 colors (32 at once in low res, 4096 at once in a special HAM mode that was really hard to program for). It had sprites, four-channel digital sound, and an array of (for the time) incredibly powerful custom chips that offloaded almost all of the graphic and sound work from the CPU.
The operating system multitasked with an incredibly light overhead. You could seriously expect to run several smaller programs in 256K. 512K was quite usable, and when you expanded the machine to 2.5MB of RAM you had a really kick-butt machine.
It didn't have the concept of memory-to-disk paging, but that's probably just as well. Disk I/O was always bad on the Amigas; their filesystem really wasn't very good, and their floppies weren't especially reliable. It wasn't until you added a hard drive that they really started to sing, and most consumers couldn't afford hard drives back then.
It wasn't until 1994 that I could really multitask on my PC (with an early Linux,.8 or.9 or thereabouts) the same way that I had done on the Amiga. Linux was, in many respects, a bit like coming home again. Linux didn't really do graphics well (in a way it still doesn't: SVGALib just isn't that hot), and X was slower than dirt on the 386-16 I was running. It wasn't an Amiga replacement quite yet, but it was sure closer than Win 3.1.
My personal PC didn't rival my original Amiga for actual useful power until about 1996: considering our family bought our Amiga 1000 in Christmas of 1985, I think that's just amazing legs. I have a friend who is still using an A2000 (1987 or so) and absolutely swears by it.
At this point the original technology is hopelessly primitive and probably not worth saving, IMO. There are just too many features missing that we are all used to. The BeOS is, at present, the closest you can get to the Amiga. If you want to get back into that type of technology again, I'd suggest BeOS on a dual-CPU PC instead.
They did finally emulate the Amiga in software. There is a commercial package with ready-to-run binaries that will allow you to do almost anything you could do with a real Amiga. Check Cloanto for details. You can also, if you wish, download and assemble the pieces separately without paying for them, a la Linux.
It's worth a look. There were some cool ideas back then. The Amiga was the most technologically brilliant personal computer ever created. There are a lot of us older geeks out there who have very fond memories indeed.:-)
As (I believe) ESR pointed out, in the digital age, we don't deal with an economy based on scarcity.
If this were 1950 and you wanted a copy of a piece of music, you would have to steal a record. This would deprive the shop owner of the ability to sell the record to someone else. (or if you stole it from a home, the prior owner could no longer listen to it.)
However, in the Digital Age, making a copy of something doesn't destroy the original. The BSA and RIAA want you to believe that they have 'lost' something for every copy of software or mp3, but they haven't. They still have the original: they may have lost a potential sale, but losing a potential sale is a lot different than losing a physical record. It doesn't directly cost them anything.
I'm not going to argue whether or not it's "right" -- I suspect that's probably on a case-by-case basis. All I'm really pointing out is that applying thinking that comes from scarcity economics to a digital economy doesn't quite work. The issue is nowhere near as black and white as some posters in the thread would have you believe.
The cognitive dissonance on this issue is really causing trouble.... as evidenced by the RIAA and its outrageous actions. We need to come up with new rules for this stuff. The old ones don't fit..
We have 8-port rack-mount Raritans at work, and I'm quite happy with them. They're pretty expensive (around $800), but on the whole quite good. I run 1024x768@85Hz through them and they look fine. We use a cheaper model as desktop switches for some of the engineers, and I believe they successfully run 1600x1200@75Hz through them.
Note that the more expensive models use a special custom cable that is KVM on one end, and a male DB25 on the other. The switchbox has DB25 ports on the back. This cuts your cable clutter down quite a lot, and the DB25s have thumbscrews to hold them in place.
At home I am using a slightly older 2-port Raritan, and presently am running 1024x758@85Hz as well. I do see a tiny bit of resolution drop with the switchbox in the way, but it's barely noticeable. I can easily work with it. The newer ones seem clearer.
I have seen one glitch. With NT SP6, the Raritans, and the keyboard we use in the server room, we have a problem with the mouse driver for the built-in touchpad on the keyboard. It jumps all over the place and can't be used at all. If I replace the mouse driver with the one off the NT CD, it returns to normal behavior. I have no idea why this happens.
We do not use wheel mice with these switchers, so be sure to check compatibility if you are going to run a wheel mouse. I don't know if it will work properly.
We have been very happy with them. They fool the machines into thinking they have a live keyboard and mouse all the time, so the PS2 mouse port is correctly enabled at boot, and you don't have trouble as you switch between machines. They have adapters so you can use an AT keyboard, PS/2 keyboard, serial mouse, or PS/2 mouse, and the box will do all the correct protocol switching for you to work with different machines. Basically, they are set-and-forget solutions.
The other thread on the Atens looks pretty interesting. They're quite a bit cheaper than the Raritans, and look like they may actually be better. However, I can tell you from quite a bit of exposure that the Raritans work well and won't give you trouble.... I'd consider them a thoroughly proven solution. I cannot personally make any such guarantee about the Atens.
Probably some of you have seen the years-long flamewar on comp.os.ibm.pc.games.strategic with Derek Smart. This fellow is eerily reminiscent.
He doesn't answer questions straight, is always 'under attack' by 'those who are out to get him', and implies that he is somehow noble for being hated.
Ol' Mister Smart turned out to be 99% bullshit. He threatened legal action on any number of occasions for things which were obviously inactionable. He repeatedly failed to deliver even 10% of what he promised. Even his PhD, which he so proudly proclaims at the bottom of every.sig, is unaccredited.
Now, I don't know that John here is the same way, but I've seen this style of not really answering-questions and implying that questioners are part of the Grand Conspiracy Against Him before. Several times, in fact. And every time, that person has turned out to be 99% bullshit.
The worst thing you can do to Vranesevich is to ignore him -- I suggest you do so. He is most likely all flash, no substance.
Slashdot Mirror
I'm curious whether or not the higher-ups at VA Linux think it's possible to develop the company enough to be worth its stratospheric valuation.
Obviously, a high valuation is partially a self-fulfilling prophecy, because it gives that company the ability to buy others with its high-priced stock. But I would be interested in any other strategies Mr. Augustin has in mind to develop VA Linux into the enormously valuable company the stock price suggests, and how it will affect its relationship with the Open Source world.
I suspect that creating vast companies like Microsoft, based on Linux, is simply impossible. An enormous service-based organization might be possible, but it seems to me that it would take many years to build one. The GPL is designed to prevent the monopoly, hardball shenanigans that Microsoft pulls. Can building hardware and providing expertise instead generate enough revenue?
It's been impossible for years to have anything like a rational public-policy discussion about anything, much less anything controversial, which you have eloquently pointed out in prior essays. The sex-repression crap is just the worst example of a serious problem. We do not talk to each other anymore.
You know, the closest thing I've seen to a public policy discussion lately is probably Slashdot. I'm both mildly amazed and mildly appalled. This is the best it gets???
Here, we are libertarian geeks preaching to the converted. I wonder how much of an effect these discussions can have on a wider world?
Does anyone but geeks really read these forums?
And when was the last time you changed your mind about a political view based on a discussion you had online?
"Most historic moment of our lifetime"... bah humbug. History is something that happened; a date is just a date.
It was just another day. There really wasn't anything special about that particular day. The only extra value it had was what we humans invested it with. It happens to be (roughly) the second millenium since the time when we believe a particular religious figure was born. It's not even especially accurate.
Tonight, when you go home and see your loved ones, it's just as important to love them and be with them as it was 4 days ago. Tonight is no less important than the Millenium was. The calendar is something we made up; it's a fantasy that most of us happen to agree about. The simple fact that everyone bought into it doesn't make it any less a fantasy. The media hype is just herdthink.
It's not worth that much heartache and pain for something that is so imaginary.
I don't know about you, but with my mid-month paycheck I'm going to go make an earthquake kit. I figure emergency supplies should be the cheapest they'll be for another 100 years. :)
The hype strikes me, now, as a negative-fulfillment feedback loop. It's a lot like the virus scares. In general (not always), the more noise you hear about a given virus BEFORE it goes off, the less damage it will do.
:)
:)
The really devastating failures are the ones that take you by surprise. If we hadn't done anything -- if everyone had poopoohed the idea of a Y2K failure, it's likely that there would have been catastrophic failures, and people would have died.
However, as I was telling people all the way along, the distribution networks wouldn't break down. If necessary, most places can still do things on paper if they have to. There's no way that Safeway (a grocery store out West here) is going to let their shelves get bare. It's not like the goods and products suddenly ceased to be manufactured. And if manufacture of some goods were to stop for a few days, it wouldn't create more than a ripple -- a bit of economic damage, perhaps, but nobody was going to starve to death. The one intelligent thing I heard suggested was to stockpile 10 gallons of water per person, as there were some last-minute worries about the water supply. Water is something you need NOW. Anyone can live without food for a week. It's a bit uncomfortable, but you won't die.
That said, the Y2037 bug in Unix has me worried. In 37 years, there won't be very many people left that really understand the fundamental architectures of today's machines, and you can bet that plenty of today's machines will still be in service, faithfully performing the jobs they were designed to do. By then, our networks will be very very complex, interrelated to an astonishing degree, and individual failures will be able to have far greater ripple effects than they can today. Y2037 might well be a disaster. It certainly wouldn't destroy civilization (unless it was tottering on the brink to begin with), but it could take years to recover from the systems crashes.
2037, for those of you who are non-Unix folks, is when the Unix counter for time overflows on 32-bit machines. Unix machines measure time in seconds elapsed since 1970, using a 32-bit counter, meaning they can only go to 4 billion seconds or so. 2037 is when 4 billion seconds will have passed, and an awful lot of today's machines are going to be VERY confused at that rollover. 64-bit machines shouldn't be affected for 4 billion times longer than that -- in other words, by using 64 bits to track seconds with, a machine could outlive our Sun without getting confused about the date. Just shooting from the hip, I think 64 bits' worth of seconds ought to be about 240 billion years. That's quite awhile.
Personally, if I am still alive, I plan to come out of retirement in about 2034 to start fixing all those old machines, and I plan to make a bloody mint.
"In 1967, you thought we were a bunch of clueless propeller heads. In 1997, you learned better.
"Y2K Consultants, Inc: It's Time to Pay Up (tm)"
(stolen from Entrepeneur, an older-but-fun game at www.stardock.com.)
Folks, not all downtime is created equal. If you are down at 12:01 on New Years' Eve, a holiday, and Saturday... big deal. Nobody is shopping then anyway.
I'd rather be down for awhile early Saturday morning than at 6PM Monday evening.
Planned downtime is a lot better than unplanned downtime. All the assertions I've seen to the contrary have so far been completely stupid.
Boy, it's nice to see people with actual experience posting on this topic. It amazes me that more people don't flame Bruce et al for such uneducated opinions. It is utterly obvious that they have very little real-world experience in actually making systems work every day, being in the hot seat if they fail. But they still Make Pronouncements From On High as if they were experts in the field.
/. as Revealed Truth. High moderation points are often just mob mentality, and have little bearing on actual expertise or good thinking.
Argh. Readers, take anything you see on slashdot with a whole container of salt: there are real experts here, and there are those who claim to be, and differentiating them without significant experience in the field yourself can be difficult. Repeating as gospel anything you read here is very foolish -- check, double-check, and triple-check before you take anything on
I'll just make one comment: any attempt to divide any group of people into two groups, based on some fantasy projection like "how they approach technology" (as opposed to provable things like male/female) is irrational.
Actually, that's not quite true. It's hyper-rational. It is using the rational mind's ability to divide, pattern-match, and differentiate in a way that does not fit the tool well.
Pointing the rational mind at people and trying to explain their motivations is highly similar to pointing a shotgun at a butterfly to understand it. It causes similar problems in both cases.
I haven't used Novell servers heavily since NW 3.12, but one thing I always appreciated about Netware was that they really thought their problems through.
:) If they have thought their other services through as well as they nailed file-and-print in 3.12, I'd consider it a very, very strong contender for your modern network.
It often took awhile to wrap your head around why they solved a given problem in a given way, but once you understood it, it was like lightning striking: of COURSE you would model this problem that way, because then you can do X, Y, and Z easily. I found this to be true over and over, wherever I looked. Every solution was arcane but incredibly powerful once you understood it.
In comparison, NT has been quite disappointing in this area. It's got a shiny graphical interface, but given the choice between ease of use and really solving a problem, Microsoft seems to opt for ease of use. On servers, I don't mind arcane syntax and weird models if they are good ones: I'm a paid professional, I can take the time to learn it if the job gets done better.
Linux and Netware both share that last trait in spades -- but comparing linux-of-now with Netware-in-3.12, on the whole I'd call Netware's solutions superior. They're not really in the same problem space, but I haven't often been impressed with the clarity of thinking about real-world problems with Linux solutions. It can almost always be made to do what you want, and is undoubtedly more flexible than Netware, but Netware 3.12 was in my opinion a better file-and-print server than Linux is today. Linux is a far better application server and general-purpose computing platform, assuming you're willing to invest the time you will need in it.
I haven't used NW5, and I presume it's quite a lot better -- they've had, what, six more years to work on it?
Did it ever occur to you that, for many solutions, Microsoft software is a perfectly valid, acceptable choice?
In our company, they implemented NT because there wasn't much expertise in-house at the time. They eventually brought in an IS staff, but the core system design choices had already been made.
And, ya know what? If you have experienced, savvy admins, you can run a nicely stable NT network. Ours is very good. Admittedly, on average, our NT servers seem to need a reboot about once every two to three months, but that's acceptable at this point.
Now, I would prefer to be deploying Unix, and we plan to gradually move toward Unixy solutions (not too much Linux yet, we don't trust ext2 enough to put critical data on it), as we plan to scale some of our system services past what NT boxes can handle. This does not invalidate the original choice of NT which, given the design constraints, was really quite good. Three years ago, when the fundamental work was being done, Linux wasn't ready yet, Novell didn't do Web stuff very well, and the 'real' Unixes required more expertise than was available.
Your blanket assertions that IS staff who wanted to shut down over New Years' are incompetent, and then above that any IS manager who implements Microsoft software must be doing it 'because everyone else does it', are false. Egregiously so. You are showing remarkably poor thinking here.
I would suggest that you contain your blanket assertions about why people do things and their competence level. It is possible to make a decision completely different than the one you would have made and still be competent. Believe it or not, there are actually people in the world who know more than you do about how to make systems work.
In the real world, the IS Director's job is to get the job done. It is not to conform to some standard of ideals. If getting the job done requires standing on one's head and counting backward from 100 thrice daily, then that is what you do. IS is there to make people's lives easy, not difficult. Imposing artificial, intellectual standards about what software is to be used is a good way of putting one's company out of business.
Suffice it to say I would never in a million years hire you to run my network. I don't need evangelists -- I need brilliant people who find the best solution for the job at hand. Sometimes that's free software, and sometimes it isn't.
Netware is just as stable as Linux, maybe even more so. It's a rock-solid system. Its interface is arcane, but certainly no harder than UNIX.
:)
You have a very good solution sitting there under your nose -- don't disregard it because lots of people like Linux. Netware has been around awhile and is really, really good. Apparently it did have some Y2K issues, but, hey... nobody's perfect.
Depends on what it costs you to be down. In our case, very little -- maybe a touch of prestige. What business-related email is coming in at 12:01 on January 1, 2000?
I consider this part of my job not to be one of prestige, but of avoiding risk. If I take a needless risk with my company's equipment, data, or even my time merely to satisfy my own vanity, I am doing them a disservice.
Vanity I save for new implementations and new projects -- for administration and upkeep, I'm perfectly willing to lose some prestige if that's what the job requires.
(If you didn't see my other comments on this topic, we did leave our systems up, FYI... seemed the least risky option.)
In the real world, where most of us live, there is a lot of Microsoft software. It has not been shown to be especially reliable, and I can't look at the code or hire someone else to look at the code for me. I have no idea what bugs lurk there -- and I don't think Microsoft does either, to be frankly honest.
Personally, I was in favor of taking our systems down overnight, simply to prevent date subtraction bugs. Someone else pointed out that this was making a change right before a major event, and that this probably wouldn't be wise -- a compelling argument, to which I acceded.
Regardless, claiming that I am somehow incompetent because I wanted to shut down systems over NY is flat stupid. Computers are not magic; they are highly predictable devices. However, the software that runs on a large fraction of them is not well understood by anyone. Trusting it unconditionally is foolish.
Consider that Microsoft was still releasing patches as of December 15.
Strikes me that you have a mighty strong opinion about how to run large networks, when it appears your expertise is not in that area. From what I can tell, you are a programmer, and a very good one. That's wonderful, but does not qualify you to make pronouncements about system administration. You probably don't deal, every day, with the stupid bugs and problems caused by unforeseen interactions in closed-source software. You live in a tightly controlled world of your own code. I don't have that luxury.
I don't presume to tell you how to do your job, and expect the same respect in return. And it strikes me that making public pronouncements on the competence of people working, every day, in an area you don't is not just arrogant, it's foolhardy.
You can trust I won't value your opinion as much in the future.
None of this is related to Y2K issues. There is no reason to bring up XiG here, nor deadlines, nor X itself.
Had I not already posted on this thread, I would have moderated the parent comment down myself.
I work for an ecommerce-ish company. We debated about it somewhat heatedly. I was in favor of disconnecting and shutting everything down, because any likely corruption bugs are probably going to hit at the rollover. The most likely equations to break are the ones involved with subtracting some earlier date from now; with the numbers on both sides of the rollover, that's where problems are most likely.
:)
However, after we talked about it further, we decided that the risk in shutting down would be greater. We are leaving things alone, running exactly as usual, so we know any problems are Y2K related.. If we were to make any significant changes and then had trouble, we wouldn't necessarily know where to look.
That argument was compelling enough to me to change my mind and agree to it; we'll be watching to see what happens. We're really not expecting anything major, but we definitely are expecting minor glitches. And we are planning to continue to devote some time to Y2K issues for the next six months or so -- we figure they will keep cropping up for awhile.
We are under no illusions that Jan 1, 2000 is the miraculous End of All Date Bugs.
You probably don't want to do RAMBUS; it's expensive and doesn't really offer all that much, performance-wise. Wait a few months for DDR SDRAM -- you'll get similar performance for a much, much lower price.
:)
I do not think you can buy, *anywhere*, a motherboard that will take 1GB of RAMBUS anyway. The design spec is for 3 RIMM slots, providing 768MB, but that nasty bug in the chipset has reduced most boards to 2 slots, or 512MB. I do not know if the chipset can handle chips of greater than 256MB. Even if it can, you will pay such an enormous amount of money for them that you might as well buy Octanes anyway. (256MB chips are running about $12 PER MEGABYTE -- $3000 EACH.) I shudder to think of what a 512MB chip would cost.
If you need to buy right away, you might actually check to see what SGI can offer you in the high end. They're doing a lot of good work with Linux. If anyone can set you up with a seriously kick-ass machine, they should be able to. They have recently released some patches to the kernel that will allow database access directly to the hard drives, for speed -- I would think that would be a likely big win for a high-end, data-intensive environment, assuming you write your own, custom software.
Good luck.
This guy is absolutely wrong, and he's probably knowingly wrong. (ie, fraudulent).
I've seen this kind of thing many times. After awhile you start to spot the frauds. Consider:
1. He's appealing to the 'underdog' effect. "Yeah, scientists everywhere hate me, and I'm going to prove them all wrong!" This elicits sympathy.
2. He is promising way, way too much. Any one of these inventions would make him a wealthy man. He wouldn't need investors the way he has gotten them; he would make a material sample, show it to some big corporation, and get bought for a couple hundred mil.
Consider: he's promising about a dozen revolutionary advances all at the same time, for one low, low monthly payment. It's just not likely that he would simultaneously overthrow all of established physics in so many areas at once.
There may very well be breakthroughs of this magnitude, but they won't happen like this: all promises up front with no delivery. Instead, a scientist will announce a revolutionary discovery and probably will make a whole lot of money. Then, a lot of OTHER people will make further progress based on the original breakthrough, and they will make tons of money too.
There's not enough wattage in anyone's head to throw back the frontiers of science in so many directions simultaneously. If this were real, his head would be totally wrapped around making the energy release work. That alone would make him enough money to buy small countries. The only reason to make claims in so many areas at once is to get investment money. I can't imagine of a surer way of showing that his basic breakthrough -- the power generation -- has no substance. If it were real, he would already be demonstrating a machine that worked.
This reminds me almost exactly of a claim by a company that was local to where I lived, about ten years ago. They claimed to have sped up the AT compatible machines of the day by a factor of 100 using off-the-shelf components.. Of course, it was a fraud, but it had a lot of people excited and I believe the owner made a lot of money. I don't know what happened to him -- hopefully he is in jail. Intel spent many billions to make computers 100 times faster than an AT. If it could have been done with off-the-shelf components, you can bet Intel would have done that. They're not stupid.
As a culture, we like to believe in the myth of a single person seeing the brilliant insight that nobody else is capable of seeing. But the stuff he's talking about here is too basic. There's no way that millions of scientists missed it. Just like the AT-machine that was 100 times faster, this 'invention' will disappear, and most of the investors will lose their money.
My $0.02.
The poster's comment above about using UAE as a virtual machine actually has some real merit. The Amiga is a very clean environment, and you can emulate one very nicely on a P2/400 or so.
Even if everyone thinks using UAE is stupid, I think it's a good proof of concept. The Amiga is a ridiculously complex design, amazingly hard to emulate, and yet the developers of UAE have managed to get an emulation running pretty damn fast. At one time, it was widely believed to be an impossible feat.
If they can make a virtual Amiga run as well as it does, it seems likely that a clean, simple virtual machine design would run very quickly indeed. The 68000-series architecture wouldn't be a bad choice. Even a virtual Atari ST might work pretty well, if that abomination of a desktop was replaced with something else.
After thinking about this a few minutes, I think the Amiga would be a better choice. Strip out the custom chip emulation layers and do a pure-CPU Amiga OS in a virtual machine. It wouldn't have memory protection, but wouldn't need it; if you want to run more than one VM program, start more VMs. It would be multithreaded and highly efficient. The total footprint of the device would probably be under 2MB... including 512K of RAM for the Amiga OS to work with.
It seems like such an obvious solution that I'm wondering why it hasn't been done. Ideas?
Oops! Yes, you are right, I forgot about the ROM issue. I have both an A500 and an A2000 sitting in my closet (broken), so ROMs weren't an issue for me.
If you are going to emulate the Amiga, you should probably purchase the ROMS, at least if you want to be legal. I also encourage you to register WinUAE(I did!).
However, the people in the Amiga community moaning and griping about sites like Lazarus are just dumb. The Amiga is *dead*. It won't come back. It is just too outmoded to be worth salvaging anymore. Might as well put up those old programs so that everyone can see them -- if some of those useful old ideas are re-introduced into computing, we will all benefit.
Obviously that doesn't include the (12?) or so programs that people are still trying to make money from... but fer chrissake, who cares if someone copies Lemmings 1 anymore? *sigh*
Agreed wholeheartedly. I actually met Carl Sassenrath, the guy who wrote Exec, at a party once. Really nice guy. As I mentioned sideways to another Amiga person there, "Wow. What do you say to a demigod?" :-)
>
IIRC, in 1994, a "good" machine would have been something along the lines of a 486-66; I don't think the Pentium shipped until 1995. At the time, it would have cost around $3,000, fully loaded.
I think you're right that the comparison is a bit specious. A 1985 A1000 cost about what a 486-66 would have cost in 1994. However, I didn't buy the A1000, my parents did. My first machine was an A500, and I eventually scraped up the $1200 or so to get into an A2000. Even by 1994, my income hadn't gone up that much (was still young), and there's no way I could have saved $3K for a 486-66. Realistically, a span of about 7 years is fairer: 1987 to 1994, comparing a cost range of $500 or so.
Even if I had had $3k in 1994, a 486-66 still wouldn't be as fast as an Amiga for multimedia apps, at least under Windows 3.1. I suspect that a custom program written in DOS on one of those machines could have come pretty close. It wouldn't have had four-channel sound, but you can fake that with 16-bit sound. (Amiga has 4 8-bit digital channels.) And the 256-color modes really did look a *lot* nicer than the standard 16-color Amiga hi-res palette. For still pictures and strategy games, the PC was much better than the Amiga in 1994 -- but it still wasn't as fast at animation.
Another way to look at it: In 1985, a PC that cost around the same amount as an Amiga would have been an 8Mhz 286 (2nd generation AT), with 16-color CGA graphics, and usually 128K of RAM. That may give you a better idea of how far ahead the Amiga really was.
And you are right that there was no MMU protection in Amiga's multitasking. This made the machine very unstable indeed for the first year or so. By the time the A500 shipped, people were getting pretty good with the very new concepts in the OS, like message-passing, memory pointers, and semaphores. A single bad program could take down the whole machine, but that's also true of every consumer OS until NT. (and Linux, of course.) Because of this, the overall programming quality was very high. Everyone knew their program could crash your computer easily, so everyone was really careful not to do that.
Overall system stability with the Amiga was at least as good as with Win98, even without hardware MMU support. And it was blazing fast; it could task switch in something like 50 clock cycles. Motorola actually used the core of the Amiga OS, Exec, as an example of how to do fast task switching on the 68000 chip.
A hardware MMU would probably have been 'better', but in actual practice it didn't seem to matter too much once the programmers got up to speed.
Amiga was developed by a company that thought they were making the next great console. Their engineers, probably the finest design team that had ever been assembled to that point, had other ideas: they were going to make it the ultimate personal computer. Most of their development was done semi-secretly -- the people bankrolling the initial project didn't realize the feature set had expanded so enormously.
.8 or .9 or thereabouts) the same way that I had done on the Amiga. Linux was, in many respects, a bit like coming home again. Linux didn't really do graphics well (in a way it still doesn't: SVGALib just isn't that hot), and X was slower than dirt on the 386-16 I was running. It wasn't an Amiga replacement quite yet, but it was sure closer than Win 3.1.
:-)
Obviously, I'm not up on all the details here, but that original seed company ran into financial trouble. Atari lent them money, knowing that they probably wouldn't be able to pay it back and they would get this incredible technology very cheaply when the little company defaulted. (I don't know if the little company was called Amiga or not.)
Well, sure enough, they hadn't quite completed their design and the loan was due -- Atari was about to move in and shut them down. Commodore rode in as a white knight and bought the company lock, stock, and barrel. They repaid Atari and took the technology for themselves. This infuriated Jack Tramiel of Atari.
To get revenge, Atari slapped together what was essentially a cheap piece of shit that, on paper, appeared to have many of the same technical merits the Amiga did. In actual fact, it was a 68000 with some RAM, I/O ports, and very simple graphic and sound chips. TOS was their single-tasking operating system. The desktop was called GEM, and was ugly and sluggish. But it had a 68000, like the Amiga, and it actually clocked the CPU faster (8Mhz) instead of the Amiga's 7.14Mhz. It also included a built-in MIDI port, which is probably the only really cool thing about it.
The Atari STs had three big advantages over the Amiga. They were cheaper, were easier to program, and were backed by a big, sleazy advertising campaign. They ran 'comparison' ads all the time which were horribly slanted. They actually did pretty well with it for quite some time. The infighting between the Amiga and ST weakened both computers, and eventually destroyed both: the PC juggernaut overwhelmed them, though the Amiga lasted a lot longer. The Amiga/Atari wars were worse than any of the distro wars you see now. But ultimately, while the Atari died a lot sooner, they both still died.
It took ten years, however, for the PC to catch up to the Amiga completely. Out of the box, it could do 4,096 colors (32 at once in low res, 4096 at once in a special HAM mode that was really hard to program for). It had sprites, four-channel digital sound, and an array of (for the time) incredibly powerful custom chips that offloaded almost all of the graphic and sound work from the CPU.
The operating system multitasked with an incredibly light overhead. You could seriously expect to run several smaller programs in 256K. 512K was quite usable, and when you expanded the machine to 2.5MB of RAM you had a really kick-butt machine.
It didn't have the concept of memory-to-disk paging, but that's probably just as well. Disk I/O was always bad on the Amigas; their filesystem really wasn't very good, and their floppies weren't especially reliable. It wasn't until you added a hard drive that they really started to sing, and most consumers couldn't afford hard drives back then.
It wasn't until 1994 that I could really multitask on my PC (with an early Linux,
My personal PC didn't rival my original Amiga for actual useful power until about 1996: considering our family bought our Amiga 1000 in Christmas of 1985, I think that's just amazing legs. I have a friend who is still using an A2000 (1987 or so) and absolutely swears by it.
At this point the original technology is hopelessly primitive and probably not worth saving, IMO. There are just too many features missing that we are all used to. The BeOS is, at present, the closest you can get to the Amiga. If you want to get back into that type of technology again, I'd suggest BeOS on a dual-CPU PC instead.
They did finally emulate the Amiga in software. There is a commercial package with ready-to-run binaries that will allow you to do almost anything you could do with a real Amiga. Check Cloanto for details. You can also, if you wish, download and assemble the pieces separately without paying for them, a la Linux.
It's worth a look. There were some cool ideas back then. The Amiga was the most technologically brilliant personal computer ever created. There are a lot of us older geeks out there who have very fond memories indeed.
As (I believe) ESR pointed out, in the digital age, we don't deal with an economy based on scarcity.
If this were 1950 and you wanted a copy of a piece of music, you would have to steal a record. This would deprive the shop owner of the ability to sell the record to someone else. (or if you stole it from a home, the prior owner could no longer listen to it.)
However, in the Digital Age, making a copy of something doesn't destroy the original. The BSA and RIAA want you to believe that they have 'lost' something for every copy of software or mp3, but they haven't. They still have the original: they may have lost a potential sale, but losing a potential sale is a lot different than losing a physical record. It doesn't directly cost them anything.
I'm not going to argue whether or not it's "right" -- I suspect that's probably on a case-by-case basis. All I'm really pointing out is that applying thinking that comes from scarcity economics to a digital economy doesn't quite work. The issue is nowhere near as black and white as some posters in the thread would have you believe.
The cognitive dissonance on this issue is really causing trouble.... as evidenced by the RIAA and its outrageous actions. We need to come up with new rules for this stuff. The old ones don't fit..
We have 8-port rack-mount Raritans at work, and I'm quite happy with them. They're pretty expensive (around $800), but on the whole quite good. I run 1024x768@85Hz through them and they look fine. We use a cheaper model as desktop switches for some of the engineers, and I believe they successfully run 1600x1200@75Hz through them.
Note that the more expensive models use a special custom cable that is KVM on one end, and a male DB25 on the other. The switchbox has DB25 ports on the back. This cuts your cable clutter down quite a lot, and the DB25s have thumbscrews to hold them in place.
At home I am using a slightly older 2-port Raritan, and presently am running 1024x758@85Hz as well. I do see a tiny bit of resolution drop with the switchbox in the way, but it's barely noticeable. I can easily work with it. The newer ones seem clearer.
I have seen one glitch. With NT SP6, the Raritans, and the keyboard we use in the server room, we have a problem with the mouse driver for the built-in touchpad on the keyboard. It jumps all over the place and can't be used at all. If I replace the mouse driver with the one off the NT CD, it returns to normal behavior. I have no idea why this happens.
We do not use wheel mice with these switchers, so be sure to check compatibility if you are going to run a wheel mouse. I don't know if it will work properly.
We have been very happy with them. They fool the machines into thinking they have a live keyboard and mouse all the time, so the PS2 mouse port is correctly enabled at boot, and you don't have trouble as you switch between machines. They have adapters so you can use an AT keyboard, PS/2 keyboard, serial mouse, or PS/2 mouse, and the box will do all the correct protocol switching for you to work with different machines. Basically, they are set-and-forget solutions.
The other thread on the Atens looks pretty interesting. They're quite a bit cheaper than the Raritans, and look like they may actually be better. However, I can tell you from quite a bit of exposure that the Raritans work well and won't give you trouble.... I'd consider them a thoroughly proven solution. I cannot personally make any such guarantee about the Atens.
Probably some of you have seen the years-long flamewar on comp.os.ibm.pc.games.strategic with Derek Smart. This fellow is eerily reminiscent.
.sig, is unaccredited.
He doesn't answer questions straight, is always 'under attack' by 'those who are out to get him', and implies that he is somehow noble for being hated.
Ol' Mister Smart turned out to be 99% bullshit. He threatened legal action on any number of occasions for things which were obviously inactionable. He repeatedly failed to deliver even 10% of what he promised. Even his PhD, which he so proudly proclaims at the bottom of every
Now, I don't know that John here is the same way, but I've seen this style of not really answering-questions and implying that questioners are part of the Grand Conspiracy Against Him before. Several times, in fact. And every time, that person has turned out to be 99% bullshit.
The worst thing you can do to Vranesevich is to ignore him -- I suggest you do so. He is most likely all flash, no substance.