Here it is, the final edited version of my Amiga response to the ST. This version can be posted anywhere you please. Hope that I have edited out the most unreasonable parts, so that everybody will be impressed by my cool logic. Or not, as the case may be. ------------------------------------------------------------------- The Myth of ST Superiority by Sheldon Leemon There is a lot of controversy these days over the relative merits of the Atari ST computers and the Amiga. A particularly disturbing aspect of this controversy is that a number of misconceptions have arisen about the Amiga, and have begun to circulate as fact. Some of these misconceptions can be seen in the interviews with Leonard and Sam Tramiel that appeared in Mary Eisenhart's recent article in MicroTimes. I would like to clarify some of these points. I do not claim strict impartiality, since I personally prefer the Amiga to the ST (though the ST runs a close second in my estimate). But I have put in a lot of time with both machines (I have had my ST since June, 1985 and my Amiga since September), and I have read the developer's materials and just about everything else available for both machines. I have programmed both computers in C and machine language, and have experienced first hand as a user the advantages and disadvantages of both machines. So if the following observations, if not absolutely unbiased, are at least, I hope, not completely unfounded in fact. MYTH 1: "THE ST IS MUCH FASTER THAN THE AMIGA". Since both computers use exactly the same microprocessor, and their clock speeds are within 10% of each other, their relative processing speeds should be fairly close. The case for the Atari is based on the notion that somehow, the custom chips in the Amiga slow down the 68000, robbing it of processing time. The facts are as follows: 1) In some cases, the Amiga display coprocessor chip (copper) DOES steal cycles from the 68000. In particular, the 640x200 mode with _16 colors_ can hog a lot of processor time. BUT, there is NEVER any slowdown when the Amiga is in any of the graphics modes supported by the ST (640x400 mode with 2 colors, 640x200 with 4 colors or 320x200 with 16 colors). In fact, even when using some extended modes that are not supported by the Atari, like 320x200 with 32 colors, and 320x400 with 16 colors, the custom chips have no effect on processor speed. Sam Tramiel is quoted as saying "None of our modes steal any processing power from the CPU; that's something we weren't going to allow". The way in which Atari accomplished this, however, was to limit the display capabilities of the computer to just three graphics modes, each using exactly 32K of display memory. The Amiga offers many additional modes, some of which steal processor time, and some which don't. The choice of how to use these modes is left up to the user and the programmer. For those situations where graphics power is more important than computation speed, the Amiga can handle a 640x400 display with 16 colors that steals some cycles from the processor--the ST can't do that under ANY circumstances. But when computation and display speed is critical, you can cut the Amiga display back to 640x200 or even 320x200 with only 1 bit plane for color (2 color mode). With the Amiga you have the option of using as much as four times the screen display memory as the ST, or as little as one quarter as much. In the article, Leonard Tramiel is quoted as saying "If you're going to do a graphics but engineering-intensive program, DEPENDING ON WHAT MODE YOU'RE RUNNING ON, it'll run a factor of three times faster on an ST than on an Amiga (emphasis added). The phrase "depending on what mode" is highly misleading, since it assumes that you will be using display modes on the Amiga that are not available on the ST. Any fair comparison of the two machines would require the display modes of the two machines to be the same, and if that was the case, no slowdown would occur. Once we have eliminated the bogeyman of "graphics mode slowdown", we are left with two machines that depend on the exact same microprocessor for all of their computation. Given that situation, the only logical conclusion is that the determining factor in processing speed would be the clock speed, where, as Leonard Tramiel himself pointed out, the lead of the Atari is at best 10%. The "three times faster" claim can only be based on conjecture, since there are no benchmarks to back it up. In fact, all of the benchmarks that have been published, such as those in Byte, show at very most a tiny advantage to the ST, consistent with its slightly faster processor speed. There are simply no hard facts to show the Atari is the speed demon that Leonard Tramiel depicts, nor that the Amiga is the plodding tortoise. And hands on experience with both machines does not support that claim, either. For example, the VIP Professional, a spreadsheet that runs on both machines, does not perform noticeably better on the ST than on the Amiga. (2) The rare cases where a slowdown does occur can be completely eliminated by adding external memory. Conflict only occurs when the custom chips are using the same area of memory where the program code resides. When you add a meg or so of external memory to the Amiga, the custom chips use the internal memory, the program is put in the external memory, and even a 640x200 display with 16 colors will not slow down the processor whatsoever. Adding memory to the Amiga (up to 8 meg) is a simple matter of plugging in a board into a port on the side, and does not require soldering memory chips on top of the motherboard, or buying a new machine each time that a model with more memory comes out. 3) The most ironic part of the argument that the custom chips SLOW DOWN the computer's throughput is that the whole point behind their design is to FREE UP processor time, by performing tasks independently that ordinarily would be done by the 68000. Graphics is the prime example. Atari proponents are quick to point out the instances when the Amiga custom chips steal processor time when drawing graphics. What they fail to understand is that when the ST is drawing graphics, ONE HUNDRED PERCENT of the processor time is consumed in the process. For example, when you draw a line on the Amiga, the processor tells the blitter where the starting and endpoints of the line are, along with some other stuff, and then goes back to serious computing while the blitter itself does the actual manipulation of display memory in order to draw the line. On the Atari, the program has to use the 68000 to calculate each point of the line, and to set every memory location on the path of that line a byte at a time. While it is doing this, the processor cannot do ANYTHING else--in effect ALL of its cycles have been stolen. So even when the Amiga's custom chips steal cycles its 68000 still gets MOST of its processing time while the drawing is going on, whereas the ST's gets ABSOLUTELY NONE, since it is doing the drawing. The fast graphics capabilities provided by the Amiga's custom chips are particularly important in graphics-based systems like GEM and Intuition, since in such systems drawing goes on ALL THE TIME. Moving the mouse pointer. Moving and sizing windows. Manipulating icons. And most importantly, EVEN DISPLAYING TEXT ON THE SCREEN IS STRICTLY A GRAPHICS OPERATION. Since the ST has a bit-mapped display rather than one that is character-mapped, EACH DOT OF EACH TEXT CHARACTER has to be drawn with BY THE 68000. When you are printing or scrolling text on the ST, the processor alone bears the burden for these operations, and cannot do anything else while it performing them. On the Amiga, the blitter moves the whole block of bits for each text character, leaving the 68000 to do the job of computing, not drawing text. This makes text printing on the Amiga faster than on the ST, particularly when the text characters are not byte-aligned. On machines that are as graphic-intensive as the Amiga and ST, faster graphics translates to faster throughput. If this were not true, and a blitter chip did not really provide any speed advantages to a graphics computer, why would Atari be trying so very hard to retrofit one onto their ST series? MYTH 2: "THE ST DISPLAY IS MUCH BETTER THAN THE AMIGA'S". The problem with this assertion is that most often, it is made on the basis of comparing the ST and its color monitor with a similar picture on the Amiga and its color monitor. This of course proves nothing, since it involves two variables, the computer and the monitor. When comparing the two monitors side by side, it's natural that the Atari display looks sharper and crisper, since it's smaller (the old Mac trick), and uses less of the screen area. Try running the Amiga display to a 9" Sony hi-res monitor, though, and your opinion may change. The only reasonable way to compare the video quality of the two machines is on the same color monitor. In all fairness, those who have run both machines on the same display still give a slight edge to the Atari. This may, I think, be partially explained in terms of a design tradeoff on the part of the Amiga. The Atari displays can be matched closely to the machine, since they are not intended for use with any other machine. The Amiga, however, works with every kind of monitor (and TV set) made, and from the beginning the machine was designed with video applications in mind. In fact, unless I miss my guess, I'd say that the reason that the clock speed of the Amiga is 7.2 MHz and not 8 MHz like the ST is that the former figure is an even multiple of the 3.59 MHz color burst crystal used in color televisions, so that using that clock speed makes it easier to sync the display to a TV set. Two years ago, when I first saw a prototype Amiga, I was amazed to see that it could display clear 80-column text on an ordinary (and rather cheap) TV set. When I reported that fact, I got E-Mail from broadcast engineers, video physicists, and even some guy who claimed to have won the Nobel Prize for TV Repair, saying that it was flatly impossible. Funny how I never hear from those people since the Amiga came out, and they can see it for themselves. What difference does full NTSC video compatibility make in a computer? Well, it does make it easier to hook the thing to a projection TV for a demonstration, something that I sorely miss when speaking to Atari groups. It also makes it possible to use those great Amiga graphics for video titling on your home video productions. Or use animation packages like the superb Deluxe Video to produce animated promotional videos for advertisements and in-store displays (as many people are now doing). And with an inexpensive Genlock interface such as the one Commodore has displayed (and hopefully will soon produce), it will be possible to use the Amiga to replace video equipment that costs ten times as much, something that should be of interest to every cable TV company in the country. The enormous flexibility that the Amiga's video capabilities offer to the user are, in the opinion of many, well worth the very slight advantage in display clarity that it cedes to the Atari when it comes to analog RGB monitors. The other part of the myth of the ST's video superiority concerns its monochrome display mode. Again, quoting Sam Tramiel from the article cited above: "there's no monochrome mode on the Amiga, you can't run a 640x400 high-resolution monochrome machine. So for serious business applications, terminal applications, you just can't do it". Here we see two misconceptions expressed in as many sentences. The first misconception is that a 640x200 display is not suitable for serious business use such as "terminal applications". That, however, is the very display resolution of the IBM Color Graphics Adapter, which is the current STANDARD among business users (since the IBM monochrome adapter cannot display Lotus 1-2-3 graphics). I find it odd that ATARI would argue that IBM'S equipment is not suited for business. And those of you reading this article on-line using a 640x200 display would probably agree that it is at least marginally suitable for terminal applications. Even more damaging to that argument is the fact that the majority of ST owners (at least in the U.S.) have ONLY the color monitor, whose maximum resolution mode is the exact same as that offered as the default mode of the Amiga Workbench. For you see, in order to make use of all three resolution modes offered by the ST (as compared to about 20 offered by the Amiga), you need to buy two separate monitors. In order to switch between medium and high resolution modes on the ST, you have to unplug one monitor, plug in the new one, and reboot the computer--there is no convenient way to make the switch, though hopefully somebody will invent a switchbox soon. And some of the software written for the ST's color monitor, primarily game software, will not work on the monochrome display (though everything works on the color display). So while the Atari monochrome monitor DOES offer a nice text display, very few ST owners use it as their only monitor (Atari itself puts this figure at 20% of the U.S. market). In fact, Atari has been running a promotion lately which gives their dealers one monochrome monitor free with every 520 and disk drive that they buy, from which you may draw your own conclusions. Therefore, it seems unlikely that we will see any software in the near future that relies solely on the 640x400 resolution capabilities of the ST, since most ST owners just do not possess this capability. One last tiny, nit-picking point about the ST color display. While the 640x200 4-color mode has become the de facto standard for the ST, in one respect it is the least pleasant to use. Because of space limitations (and, some say, the unwillingness to pay DRI the extra funds required), Atari opted to include only one set of graphics data for its desktop icons. Since the 320x200 mode and the 640x400 mode have the same aspect ratio, they chose graphics that would look good in those two modes. In the 640x200 mode, however, those same icons look rather ridiculous--they are tall and skinny, half as wide as they should properly be. All in all, they give the impression of an El Greco version of GEM. The other misconception implied by Mr. Tramiel in the above quote is that the Amiga does not have a usable 640x400 display mode. Nobody will deny that with the current color Amiga monitor, it is not possible to use 640x400 interlaced text because of the the "jitter" of the low-persistence phosphor display. Even so, careful color selection and placement can be used to avoid the problem entirely. For example, the Digi-View digitizer can create black and white digitized images in 640x400 resolution with 16 grey levels, and those very life-like images do not "vibrate" in the least on the current color display. Moreover, the user can always get a separate monochrome monitor that uses a high-persistence phosphor, and enjoy 640x400 resolution without the jitter. Recent changes made by the 1.2 version of the Amiga Operating System enable the use of a 640x400 screen as the standard Workbench environment. Nor is the color 640x400 display on the Amiga totally hopeless for text work. Several analog RBG monitors with high-persistence phosphors are now on the market, and Commodore-Amiga displayed some of them at the Spring Comdex. As one who attended that show, I can tell you that the video quality of the Amiga color display in 640x400 mode on such a monitor is every bit as good as the 640x200 display on the current Amiga monitor. Moreover, the persistence of the phosphor was not high enough to cause "trails" when the screen scrolled. While it is true that most of these monitors currently sell in the $1000 price range, the same was true two years ago of the kind of monitor now used with the Amiga. There seems to be no reason why the price of such monitors cannot come down to the same level of the present Amiga monitor (one of them, the NEC Multisync, can already be purchased for as little as $500). And there are newer monitors coming onto the scene that can digitally store and combine the two frames of an interlaced display, creating a non-interlaced display from them. In short, there is every chance that in the near future we will see a single affordable monitor that can be used for ALL of the Amiga graphics modes, including 640x400 in 16 colors. In the meantime, the truly fanatic Amiga user may buy one color monitor for display resolutions of 640x200 and below, and a separate monochrome for 640x400. MYTH 3: "THE AMIGA OPERATING SYSTEM CHANGES SO OFTEN AND IS SUCH A MESS THAT NOBODY CAN PROGRAM THE MACHINE". This myth is a leftover from the pre-release days of the Amiga. The developers who had to work with the earliest versions of the Amiga had an uphill battle all right, not because the operating system was so buggy, but because half of it hadn't been written yet! Don't forget, according to MetaComCo, it was not until FEBRUARY 1985 that they produced the very first prototype disk operating system for the Amiga. Until that point, you could not even read or write a disk file conveniently. In contrast, the ST was almost ready to ship to developers by then. Considering the amount of catching up the Amiga had to do, and the time its developers had to do it in, it's small wonder that the operating system was changing "every week". ST enthusiasts seem to be greatly amused that the internal version number of the Amiga OS is already up to version 33. As any programmer can tell you, however, every program goes through a constant series of revisions while it is being developed. For example, a piece of software for which I am currently writing documentation has gone through 11 beta versions in the past 6 weeks. Since each one represents a number of compiles, it can be seen that how high the version number goes depends mostly on how many times you stop to number the program. But no one should one conclude that developers have gotten 33 different versions of the Operating System to work with. Only a handful of developers outside of Amiga ever SAW a version lower than 27. The fact is, the vast majority of developers received the official release as their first version of the Operating System. They, like the general public, have been faced with exactly one revision of the OS since the machines release (with another one in the works). This record compares quite favorably to machines like the Mac, and even the ST, which has had a similar number OS revisions since its release. The revisions have maintained upward compatibility, so that all software that worked under the older versions will still work under the newer versions. Though the Amiga Operating System is far from perfect (and will undoubtedly be improved in the future, since it is not locked in ROM), it is a system that was designed for maximum flexibility and expandability. There are many design features that support future hardware enhancements, such as OS support for screens as large as 1024x1024 pixels. The Amiga OS even is designed to support the 68020 processor and the 68881 floating-point coprocessor. A third- party maker of 68020 boards, CSA, was able to plug their replace- ment board into the 68000 socket of the Amiga and get it to run existing Amiga software, at a much higher rate of speed, with no modification. Such forethought is evident throughout the Amiga OS. The virtues of the Amiga Operating System (and its faults) are too numerous to discuss here in detail, but we cannot move on without mentioning the one feature that the Atari community invariably overlooks--multitasking. This feature alone raises the Amiga OS to an entirely different plane than that of most other personal computers. Those who have not had the benefit of it may think that multitasking is no more than a flashy gimmick, since most of us are used to doing one thing at a time. Those who have had the chance to use this feature, however, think otherwise. For example, an Amiga owner recently told me of his amusement upon reading the complaint of an ST user about a terminal program that would not allow him to format disks while online. With the Amiga, you can run a wordprocessor, format disks, and play a game simultaneously while online! It can be a great feeling to just pull your program window to reveal the desktop, and start up another program without closing the current one. Of course, as ST owners would be quick to point out, GEM can multitask--sort of. There are always desk accessories that can be launched from your program (though most of the time, you can use either the program OR the accessory, but not both at the same time). And there are Operating System add-ons like the soon-to- be-released Micro RTX that is supposed to add a fair level of concurrency to GEM (providing that all of the programs are written in a certain way so as to conform to the requirements of the multitasking module). But these are no real replacement for multitasking built into the lowest level of the Operating System. Such a system allows ANY program to multitask with another, subject only to the limitations of display space and memory. While we are on the subject of Operating Systems bugs, it is only fair to point out that the ST has certain problems in that area itself. In the words of Russ Wetmore, a programmer who has worked with the ST as closely as anyone, and a man who is one of the leading lights in the Atari programming community, "I hate GEM. I hate GEM. I hate GEM. Did I tell you that I hate GEM?". Although Russ may have put it too strongly, the point that he is making is that from a programmers standpoint, GEM can be somewhat cumbersome to work with. But what is more disturbing is Atari's implementation of it. Or rather, lack of implementation. Some rather important elements of GEM are missing from the ST version. Things like software-loaded fonts. And virtual device drivers. Atari has been promising for months that a GDOS addition to GEM will be available Real Soon Now, but until then, a lot of the GEM calls that are listed in the Digital Research documentation just do not work with the ST. For example, currently, the ST supports Atari's own printers, and the Epson. That's it. As a result, application programs have to kludge together their own support for other printers, the very thing that virtual devices are supposed to prevent. So you have to track down a "First Word printer driver" and a "Degas printer driver". On the Amiga, there is global support for a dozen printers supplied with the system, and third-party and public- domain support for dozens more. That means that installing just ONE printer driver makes your printer work with ANY program. And that includes all of the special features too, such a bold print, italics, underlining, and custom line spacing. Even if the ST version of GEM was as complete (and bug-free) as the IBM PC version, it would still lack some of the nice "extras" found on the Amiga. Like user-definable console keymaps. A built-in speech synthesis device and text-to-speech library. Even funda- mental things like a built-in command line interpreter. But there is no reason to belabor the point. It should be enough to point out that sufficient software has been produced for the Amiga to lay to rest the charge that it is impossible to program the machine because of its OS. MYTH 4: "THERE'S A LOT MORE SOFTWARE FOR THE ST THAN THE AMIGA" To find out how much software is available for a particular computer, the traditional test is to stack up all of the diskettes one on top of the other to see how high the pile reaches. Remember when IBM did just that to show us how much software there was for the PCjr? By this measure, the contest is very close. If you look at the catalogues of available software put out by Amiga and ST for their dealers, you will find a similar number of listing for each. And the number of ads in magazines like AmigaWorld, Amazing Computing, Antic and STart are pretty similar (if anything, there appear to be more ads in the Amiga mags). Sheer quantity, however, is not everything. I remember a period when every time I would walk into an Apple dealer's showroom, the salesman would start in on how there were over 20,000 software packages available for the machine. I would look around, and see only two or three dozen of those packages on the shelf. What happened to the other 19,964 I would wonder? Are they storing them in the back room, or are they just too dreadful (or useless) to sell? When you come right down to it, a handful of packages account for most of the software sales. My local Amiga dealer has his two or three dozen packages on his wall, just like the ST and Apple dealers. Once it's been established that each computer has a number of software titles in the categories of word processing, spreadsheets, data bases, accounting packages, computer aided design, drawing and painting, music, education, and games, the question is no longer which has more software, but which has more good software. When I started to write this, I thought that the best strategy would be to compare the most important packages in each category for both machines. In trying to compare specific packages head to head, however, I kept getting reminded by both sides of favorite software that I had forgotten to include in the evaluation, or had not ever seen. I also came to the conclusion that there was a heck of a lot of software in every category for each machine, far too much for me to pass on the individual merits of each, with new titles appearing literally every day. Moreover, even if I could evaluate each, it would really only be my opinion, and I have been trying to avoid judgments that involve ONLY my own opinion (though I am sure that many would disagree). After all, though Mac software is different than IBM software, is it possible to say that one is clearly better than the other, when there are users who prefer each? Therefore, I will quickly summarize my SUBJECTIVE impressions, and leave the "my software can beat up your software" argument to others. First, each machine has got usable software in the main categories of word processing, spreadsheets, data bases, and accounting software. This software seems to fall into two types: simple programs that make use of the windowing environment to good effect, and more full-featured programs that don't. But none of the current generation of software in these areas are going to throw a scare into IBM or Apple. Both machines are clearly capable of supporting much better, and much better software is on the way for both. Secondly, each machine has a number of interesting software packages in specialty fields that are NOT covered as well by the PC's and Macintoshes of the world, and there will undoubt- edly be a lot more of these for each. And finally, there is getting to be a fair overlap of titles on the two machines, a trend which will probably continue. When many of the best packages run on each machine, the comparisons will become rather easier. MYTH 5: "THE AMIGA IS JUST A 'GAME MACHINE' " No one would argue that the Amiga's superior graphics capabilities do not lend themselves well to computer games. The misconception here is that these capabilities somehow overshadow or even negate the features that make it a powerful general-purpose personal computer. You would think that the words "game machine" would be never be spoken at the new Atari, since no company has suffered more (and undeservedly so) from the misguided notion that a computer that can play games can't also be a powerful business and scientific tool. And yet, in the interview cited above, Sam Tramiel is quoted as saying "The Amiga's a great game machine... It's a great, fast, low-end game machine". It seems hard to believe, based on the Amiga's specifications, that anybody could dismiss it as nothing more than a toy. If you told a PC user that IBM was going to come out with a true 16-bit computer that ran as fast as the AT, came with a built-in 880K 3.5" disk drive, a half meg of RAM, expandable to 8 1/2 meg, a fast color graphics display, built-in serial and parallel ports, 25 DMA channels, and a true multitasking Operating System he would probably say "What a great business machine!". But apparently some people hear the same specifications, specifications remarkably similar to ST, and think "You could play some really BIG game programs with 8 meg of memory. And you could store TONS of games on an 880K floppy or 40 meg hard disk. And that multitasking is really great for playing a WHOLE LOT of games at the SAME TIME". An interesting question to ask is why does the ability to play great videogames on the Amiga make it unsuitable for business purposes, when the same ability does not seem to have a similar effect on the ST? After all, there are probably more video games for the ST right now than there are for the Amiga, games like Major Motion, Time Bandit, and Atari's own Joust, Millepede, Battle Zone, and Star Raiders. In fact, the first piece of third-party software that I could obtain for the ST was "Mudpies". Yet Atari does not categorize its own computer as a "game machine". Instead, Atari officials wisely emphasize that a powerful computer can perform a wide variety of tasks. One would think that bit of wisdom would apply to other computers as well. There are a number of people who believe that a computer's ability to run complex visual simulations (sounds better than games, eh?) bears no relation whatever to its overall computational ability. But the fact is that graphics objects that appear on the screen are represented internally as numbers, a lot of numbers. Manipulating these numeric objects in real time requires the exact same type of mathematical calculations as the most "serious" business appli- cations. Often, it requires a more intense series of calculations in a much shorter time than is required by "serious" business applications. Therefore, a computer's ability to play great games can be an indication of its overall computing power, and its in- ability to play those games an indication of its lack of power. Take the IBM PC for example. It is extremely difficult to program arcade games on such a machine, which combines a relatively slow processor with slow graphics. That same sluggishness makes GEM run at about half the speed on the PC than it does on the ST. And that same lack of raw computing muscle has sent the PC power users run- ning to the AT, because they are tired of waiting for really big spreadsheets to recalculate. The key point here is that the Amiga's fast graphics, which make videogaming on the machine so easy, in no way detract from the fact that the computer has all of the requisites for a "serious" business machine--a fast 68000 processor, multitasking OS, lots of RAM, lots of mass storage, an 80x25 character display, and a full keyboard with function keys, cursor pad, and numeric keypad. The special abilities afforded by its custom VLSI chip set are all IN ADDITION to those other features, and complement them, making the computer more powerful. It seems silly to say that the extra power offered by those chips somehow disqualify the computer from any business use. That's like saying the Lotus is a nice car for frivolous sports like racing, but that you couldn't possibly use it for "serious" driving chores like going to the store to buy some milk, or delivering pizzas. CONCLUSION. From the above you might have gotten the impression that I think that the Amiga is the greatest computer past, present, or future, and that the Atari ST is a worthless piece of junk. NOT SO! I think that the Amiga has its share of very real problems, and I also think that a rational case could be made for preferring the ST to the Amiga. To somebody who paid $850 for a 16K Atari 800 just five years ago, it is almost unbelievable that you can now get a full-blown 68000 system with half a meg of memory, a disk drive, and an analog RGB monitor for less money, using inflated dollars to boot. It is quite reasonable to argue that the extra features of the Amiga, such as its graphics and multitasking do not add up to extra functionality, and that when you get down to it, just about anything that you can do with an Amiga you can do with an ST for a whole lot less money. Or that the ST's low price is bound to generate a huge installed base, which in turn will assure massive third-party hardware and software support for the machine, while the relatively high cost of the Amiga will restrict it to a much smaller specialty market. Or that the ST has a clean, simple design that is easy to work with, while the Amiga is unnecessarily complex, and represents "overkill" for a computer its size. I would not mind hearing any of these arguments put forth on the Atari side. Nor do I mind hearing people criticize the Amiga for ACTUAL faults, like its slow and rather eccentric DOS (watch out Atari fans, the same zany Brits that brought us AmigaDOS are working on a new, slow, BASIC interpreter for the ST). With so many rational and reasonable arguments, then, why invent fairy tales like an 8 MHz 68000 computer that runs three times as fast as a 7.2MHz 68000 computer? Or try to convince people that it is possible for a 68000 computer that can handle up to 8 megabytes of RAM to be only a "game machine"? The fact is that the Atari is a "plain-vanilla" 68000 system at a sensational low price, and that the Amiga is that same system dressed up a little fancier. Which you prefer is mostly a matter of personal taste, or your financial resources, not a question of the clear and overwhelming superiority of one machine over the other. One thing is for certain, though. Either machine beats the hell out the boredom induced by the illegitimate offspring of IBM and Intel.