Requiem for a Dear Friend

1995.08.08

Last updated: 2020.01.11

It's only now just hit me.

I mean, it's not like I hadn't heard all the rumors, seen all the news reports, read all the punditry. I knew intellectually what Reality was. And yet, I never really understood it, believed it, until now.

The Amiga really is dying.

Strange. I can hear you groaning even from here. If I listen carefully enough, I think I can even make out what you're saying: "Get a life!" "It died years ago!" "Wake up and smell the coffee!" "Quit dreaming!"

Dreaming? Yeah, I suppose I have been dreaming these last nine years. Dreams are usually pleasant experiences where strange and lovely things happen, and Amiga was certainly that kind of place. The Reality of the Microsoft-dominated world is, shall we say, a less remarkable experience (which is not to say I couldn't make a lot of remarks about it).

This is a lament of sadness, of frustration, of anger. It's easy enough to say, "Just let it go," but that's rather like asking a person to give up a old sweater they really like. Yeah, it's torn and has holes in it, but you know it. It's familiar to you. It's an old friend. It's not so easy to cast aside old friends, especially friends with whom you grew up.

Does any of this make any sense? I mean, it's all well and good to simply say, "It's dead. Move on." But can you imagine how depressing that path is? Imagine for a moment that the Internet suddenly became economically unviable, died, and you were suddenly faced with the Reality of choosing among one of the major commercial online services. Would you move on quietly, or would you kvetch and complain and lament the passing of That Which You Really Did Like A Whole Lot Better, even as hundreds of thousands of Prodigy users smugly told you, "Wake up, it's dead?"

Well, that's rather like how I feel about the Amiga. As much as I and about three million other people wish it weren't so, Amiga's passing is inevitable. And tragic. Amiga's failure in the marketplace is not simply a testament to the incompetence of Commodore management, but an incalculable loss to the advancement of the computing industry as a whole.

Hyperbole? Possibly; let's look at it more closely. Few would argue that, at the time of its introduction, Amiga's technology was unmatched, both in terms of hardware and software. Taken as a whole, it was the finest small computer ever designed.

How does it fare nine years later? Well, Amiga's greatest strength -- it's custom graphics chips -- became its greatest liability about four years after its introduction as new graphics cards for the PC and Mac appeared, exceeding Amiga's resolution and color depth. Amiga's sound capabilities were soon eclipsed as inexpensive sound hardware started showing up for the PC and Mac. PC and Mac moved to 1.44 Mbyte floppy drives, while Amiga stayed at 880K. Both the PC and Mac got multitasking. And of course, the available software grew steadily for the Mac, and exploded for the PC.

But look more closely: Just how well have these technological advancements been integrated into the PC? Ask any multimedia or game software developer just how easy it is to make use of these nifty features, and they'll relate to you any number of stories ranging from disquieting to nightmarish.

What version of DOS are they running? Are they on a 286, 386, 486, or Pentium? Are they using XMS or EMS memory (or both)? Are they running any astonishly weird (but extremely popular) TSRs? Do they have a graphics card of any kind installed? Is it EGA, VGA, SVGA, MCGA, Tandy, or something else? How about a sound card? Is it FM-synthesis or sound sampled? How about the disk controller/driver? Is it a clever DMA controller, or is it just an IO port with a driver that spins in a loop on the port until all the bytes are transferred? (Hey, you have to know this stuff if you're writing performance-critical multimedia applications.) This is the Reality of software development on the PC.

On the Mac, this Reality is mitigated somewhat, as Apple have spent a lot of time and resources trying to address these issues. Unfortunately, they've solved these problems mostly by throwing ever more RAM and CPU at them, which are then promptly consumed by the new system modules. And who knows whether or not that INIT your new application just installed in the System Folder is going to break something.

As for multitasking, megabytes of flames have been written about whether or not pre-emptive multitasking (where the system switches the CPU between multiple programs automatically) is better/worse than cooperative multitasking (where programs yield up the CPU voluntarily). Whether or not one is "better" than the other, the embodiment of Amiga's pre-emptive model, to my eye, appears to out-class Windows's and Finder's cooperative models.

To wit: If some exceptional event occurs on a Mac, a dialog box with a button pops up informing you of the event. Click on the button, and the box goes away. Fairly clean, except that the entire machine stops dead in its tracks while it waits for you dismiss the dialog. Oh, and you can't click outside the box to activate another application (like the Finder). I can't tell you how many times I've started some lengthy process on the Mac and walked away, only to return some time later to find the system has very thoughtfully popped up a dialog box about an AppleShare server unexpectedly shutting down, my lengthy job having been halted until I explicitly acknowledge this late breaking event.

Even given that there's very probably an INIT out there somewhere that solves this problem, the very fact that this problem exists at all is amazing to me. It is amazing to me because it is a problem that never existed on the Amiga; I've never had to deal with it before. If some exceptional event occurs on the Amiga, a system requester pops up with two gadgets; Retry and Cancel. The rest of the machine continues to run. I click in another window, deal with whatever the machine is complaining about, go back to the requester and click 'Retry,' and it goes on it's merry way. "Volume 'Empty' is full." Click in Workbench, delete an old file on Empty, 'Retry.' And it works. Any other behavior is puzzling to me. And why shouldn't I be puzzled; Amiga's method seems to make a heck of a lot more sense.

I will refrain from boring you with the dozens of other such examples. Most examples, however, fall into a largely-invisible realm seen only when one is doing system programming and software development. Understand, though, that while issues of good/bad underlying design may be invisible to the end user, they are extremely visible to the programmer, who must work with them or around them to get the job done. The worse the underlying design, the harder their job. Ultimately, the user hopefully won't be able to tell because the programmer successfully hid all the underlying design flaws under smart code and a pretty UI.

This, I believe, is why so many innovative applications were first developed for the Amiga, not the least of which was NewTek's Video Toaster. Because the Amiga had comparatively few design flaws, programmers and designers could concentrate less on fighting the machine and more on development. Amiga presented the fewest obstacles between the vision and its realization.

You probably enjoy many of these innovations today, either directly or indirectly. Deluxe Paint from Electronic Arts made its debut on the Amiga, and is still one of the most widely-used and highly-regarded paint packages available. Low-cost 3D rendering packages were available for the Amiga first, and underwent significant refinement; to this day, you'd be hard pressed to find a more formidable array of rendering packages on any other platform. The MOD files you're probably listening to in the background owe their heritage to the Amiga and the thousands of insane European programmers who invented and refined them. Full-screen animation at 30 frames per second is a familiar occurrence, as are hard disk transfer rates exceeding a megabyte per second through the filesystem.

But there's still a long way to go. Amiga is still the exclusive domain of "Screens;" independent video display buffers that can share the same physical monitor. You have no idea how useful this is for me, and it's a feature that I gravely miss. Pre-emptive multitasking is just now starting to show up in the form of Windows-NT and OS/2. But please explain to me why Windows-NT requires 12 megs of RAM and a 66 MHz Pentium to appear as fast as my old Amiga 1000, which had 256K of RAM and a 7.1 MHz 68000.

Some may counter by saying the Amiga was unreliable; that it crashed a lot. I daresay Windows and Finder have an equal, if not greater, history of unreliability, as any victim of the Unrecoverable Application Error or the "Sorry, a system error occurred" dialog box will tell you. Yet somehow, Amiga earned the reputation of being crash-prone, while Windows and Finder became paragons of dependability.

This is not to imply that Amiga didn't have warts. The most obvious among these was the DOS/filesystem, based on TriPos, developed at Cambridge University, written in the little-known language BCPL. The port of TriPos to the Amiga was very quick and very dirty. Anything more than the most superficial file operations required dealing with the non-native BCPL pointer types, which had to be converted back and forth by hand.

Amiga's user interface manager, Intuition, also lacked a certain measure of "convenience" routines to automatically set up the rather daunting number of data structures required to create windows, menus, gadgets, etc. While Intuition was extremely flexible, the price of this flexibility was the work required to fill in all the necessary details.

In spite of these shortcomings, however, Amiga was still a pleasant machine to program for, as evidenced by its having flourished as thousands of programmers wrote some of the highest-quality public domain, shareware, and commercial software anywhere. One man, Fred Fish, personally compiled a library of one thousand disks of PD software, freeware, and shareware. Thousands of crazed hackers across Europe, forming in groups with whimsical names like, "The Silents," "Phenomena," and "Red Sector," wrote graphical and musical extravaganzas whose kind are only now just starting to appear (in primitive form) for the PC.

Why am I even bothering to natter on about this at all? Because it's not apparent that anyone knows or even cares about any of this. Since the Amiga failed in the marketplace, the Common Knowledge is, "Oh, it died. There's nothing to be learned from it." This is particularly frustrating when it is readily apparent (to me, anyway) that there's tons to be learned from it. Amiga was a failure in product management and marketing, not in technology. In fact, some of Amiga's technology has yet to be equalled, much less beaten (the above examples being but a few). Yet, because it died, the entire machine appears to have been blown off as irrelevant, unremarkable, insignificant, stop-moaning-about-it-and-get-a-life, etc.

To write off the Amiga as technically insignificant would be a mistake of the highest magnitude. Ignore the specific points where it falls short and look at the philosophy behind the machine. Everything is integrated almost seamlessly, yet totally modular. The system design is a minimalist one, utilizing resources frugally and efficiently. Individual modules are very complete and orthogonal. The system is also highly predictable; twiddling a bit somewhere is likely to do exactly what you expected. Because of this minimalism and predictability, the entire system can be comprehended by one person, making it easier for one person to design or manage a product that uses the system effectively and correctly.

These are not insignificant traits, and they are traits I perceive to be lacking in other major platforms. Unfortunately, they are traits that take time to perceive; you'd have to use the Amiga for several weeks before you could begin to appreciate its finer points. As someone familiar with these fine points, and based on my initial exposure to PC's and Macs, my perception is that to move to either of those platforms would be to take a step backward in sophistication.

I mean, it's been nine years since Amiga came out, and where's the rest of the market? The Macintosh has matured relatively well, but there are still warts. Apart from the aforementioned machine-halting dialog boxes, one of the most annoying among these is the perpetual possibility that any new INIT/CDEV will put the reliability of the system at risk. Since, when the Mac dies, it does so in a most unhelpful way, figuring out which INIT is the bad one is a process of exhaustive search. While Amiga was not perfect along these lines either, the failure mode was usually better defined, and individual candidates could be tested independently without requiring a reboot.

On the PC, the conditions are downright archaic. It still looks, feels, and behaves like a CP/M system (which is not entirely coincidental, given MS-DOS's origins). Despite the advent of the 386, DOS is still living in segmented 16-bit land. Hardware installation, setup, and configuration is preposterously difficult, an almost cruel parody of the disquietingly similar steps you had to take on old S-100 systems (at least back then we didn't pretend it was easy). The kooky option flag character (forward slash). The kooky directory pathname delimiter (backslash). The worst possible end-of-line sequence (CR+LF). And it's still laboring under 8.3 filenames. Not even Windows can conceal all these anachronisms (and it even introduces a few of its own). Windows-NT is not a viable alternative for the simple reason that it's an even bigger resource hog than UNIX, without actually offering any tangible benefits over it.

Both systems consume enormous amounts of disk, RAM, and CPU for no perceivably justifiable reason. (Simply saying, "Hey, disks and RAM are cheap; what's the problem?" doesn't cut it. Merely because a resource is abundant and inexpensive is not a reason to abuse it. You don't waste water, do you?) Neither system makes efficient or intelligent use of its own multitasking facilities. The black text on white background color scheme doesn't appeal to my eyes (a point of personal taste, I'll admit, but at least Amiga let me freely change it). And of course neither system offers Amiga-like draggable Screens.

Some may regard these points as disingenious, completely curable, or even non-existent once I use the machine long enough and get used to it. (The same arguments could have just as validly been levelled against negative comments made against the Amiga environment.) Well, I've been using the Mac fairly steadily for over two years -- the PC much less so -- and I've yet to discover workarounds for my primary remaining gripes; namely, the system dialogs that halt the entire machine, and the single cluttered screen. These limitations appear to be fundamental to the design of the systems. Most of my other gripes relate to little niggly details, like the non-proportional thumbs on the scroll bars which first appeared on the Mac, and which Windows mindlessly copied (Amiga's scroll bar thumbs are sized in proportion to the amount of the document currently visible, so you know at a glance how much of the document remains unseen).

In the final analysis, however, all of my little gripes and so on are irrelevant. Amiga could have behaved like anything. What is truly irksome is that Amiga has never received fair credit for many of the innovations it pioneered which are now commonplace in the industry. Nor were Amiga's concepts and design details ever properly acknowledged as useful. When Amiga came out, its multitasking OS was proclaimed by the pundits as a liability; what else could you possibly be doing while a spreadsheet is recalculating? Then Apple released MultiFinder, and was hailed as the "inventor" of microcomputer-level multitasking. (To be honest, not even Amiga can lay claim to that title, as OS-9 was running on the TRS-80 Color Computer well before Amiga's release.) Amiga had purgeable shared libraries since day one. Then Microsoft "invents" DLLs and is regaled with praise on being an industry innovator. Today, the industry press looks to Mircosoft as the One True Source for All New Things Great And Wonderful without, apparently, even once considering that, just maybe, it's been done before, years earlier, much better.

Yes, you are indeed detecting a tone of bitterness. It stems from Commodore's failure to market the Amiga properly. It was incomprehensible to me (and, for that matter, everyone else) why Commodore couldn't see the enormous value and potential this system had, and why they weren't communicating this value to everyone else. Having put my faith and livelihood behind the Amiga, I ended up being materially screwed by Commodore's ineptitude.

This does not change the fact, however, that I still believe in the machine and the philosophy it embodies, a philosophy which seems in alarmingly short supply everywhere else. The philosophy, as I see it, is this:

"We wanted to make the world a better place. So we looked at everything else we could get our hands on, sorted out all the ideas, and thought about it a LOT. We thought long and hard about every detail, how each piece relates to itself, to every other piece, to pieces we have yet to design, and to pieces that will be designed by others. We resolved to do the absolute best job possible, because we felt anything less than our absolute best wouldn't make the world a better place. The result was this system."

Members of the Amiga design team may accuse me of overstating things but, speaking entirely for myself, that's what I sense when I look at the Amiga: A commitment to do the absolute best job possible (because why would you want to do anything else). It's a philosophy I've adopted as my own, and try to apply with great energy to my work and life.

So what suddenly made me truly realize the Amiga is dying?

Well, my Amiga died. The fan in the power supply was making hideous noises, so I took it in to be replaced. As a result of transporting the thing, however, the hard disk started taking enormous numbers of read errors. Hardly surprising, as it was an old Rodime drive with an ST-506 interface (remember those?). None of the data on it was critical, but with a non-functional drive, the machine was well nigh useless. So, how to get it fixed?

"I don't suppose you have any ST-506 drives laying around?"
"Nope."
"So how much for a SCSI controller?"
"I can sell you one from $(MANUFACTURER) for $150."
"Oooo, I don't care for them. How about the one from $(SOMEONE_ELSE)?"
"I can't get those, and haven't been able to for months."
"Well how about $(STILL_SOMEONE_ELSE)'s?"
"Nope, they stopped making them."

I wandered around the store evaluating my options at that point, which were pretty much non-existent. As I did so, I caught fragments of comments of how Amiga hardware and software were almost impossible to come by these days. The PAR (Personal Animation Recorder) from Digital Processing Systems, a truly orgasmic piece of video hardware, was in plentiful supply for PC's, but backordered for Amigas.

I took the machine back to the office. As I sat clicking furtively on the ever-reappearing 'Retry' gadget, trying to get whatever data I could off the drive, it slowly dawned on me that I was seeing how it would all end. Each of us who had believed in this machine, who had shared the same kind of enthusiasm that created it, who put significant pieces of our lives into it, would one day find themselves trying patiently but futilely to get it to perform some basic function, a function it would never perform again for want of support, service, and spare parts.

Many rumors are flying about who's going to purchase the burned-out hulk that is Commodore International, which filed for bankruptcy in mid-1994. A few bidders are lined up, hoping to purchase Commodore's technology and make a go of it. But Motorola is ramping down on future 680x0 development -- the CPU on which Amiga is based -- choosing instead to focus on PowerPC. So Amiga's bound to hit a performance dead-end on its CPU. And even in death, Commodore continues to make paralyzingly stupid decisions, as the creditors and bankruptcy attorneys stall hoping to get higher bids. Thanks to these delays, Christmas 1994 was lost to anyone who might have picked up Amiga, and with the advent of new machines from competitors scheduled for a 1995 rollout, Amiga's future outside of vertical markets is virtually non-existent. It doesn't surprise me, really. I've seen Commodore make so many utterly boneheaded moves, I've become numb to it. An amoeba could have done a better job running the company.

I now realize that the Amiga at which I happily type right now will one day wither and die (already the keyboard is starting to flake out). When that sad day comes, I'll have a plaque made, and Amy will occupy a place of honor in my home. It's the machine I grew up with, the machine I made friends with, the machine that made me feel at home. When I leave that home, if the PowerPC hasn't caught on big time, I'll likely buy some commodity Intel hardware and do my best to make it as tolerable as possible. But it will never be like my first home. And virtually no one will know or believe or even care how nice that home was. And that makes me sad.

Au revoir, mon Amy.

Postscript 1996.01.10

The remains of Commodore Business Machines, including all the Amiga technology, were purchased in May 1995 at bankruptcy auction by Escom for US$12 million. Escom has spawned a subsidiary company, Amiga Technologies GmbH, which has resumed manufacture of the Amiga 1200, released the Amiga 4000T, and plans to migrate the Amiga technology to the PowerPC architecture.

We'll see...

Postscript 1998.09.07

I found out about it when it happened, but I haven't bothered to write about it until now.

As it happened, Escom keeled over as well, and the remains of Amiga were sold to Gateway Computers. Gateway has claimed that they're going to move the Amiga technology forward, update the OS and so forth. However, it's beginning more and more to look like Gateway simply bought Amiga for the patent portfolio (as Commodore held a number of early patents on GUI techniques). This gives Gateway a bit more leverage when bargaining with the likes of Microsoft, but it basically leaves the Amiga remaining where it was in 1994, when Commodore ate itself.

Some enormously clever products and hacks are still appearing for the Amiga, but I find I'm getting a lot more use out of my 486 laptop (running Linux, of course) than out of my Amiga. And, at the risk of appearing as a traitor, I recently decommissioned my Amiga 3000, replacing it with a 400 MHz Pentium-II system, which triple-boots into Windoze (for playing games, since Windoze isn't good for much else), Linux, and BeOS.

Oh, did I mention I work for Be, Inc. now?