The Fork in the Road

The decisions we face as the Amiga prepares to head into the future

By Michael Webb, Editor-in-Chief, MikeWebb@CompuServe.COM

In the years since Commodore disintegrated, people have looked back and wondered what went wrong. Some of it was obvious, while other aspects are most certainly open to interpretation.

Today, the Amiga has overcome inconceivable odds and survived, as a viable platform, in the intervening period since the bankruptcy, with little hardware development, aside from the availability of numerous 68060 accelerators and the introduction of some newer graphics boards. Much of this is doubtless due to how superior and capable a machine it was (and is), but clearly this momentum will not last forever. We therefore must prepare to move the Amiga forward if it is to continue to survive.

Not surprisingly, there is difference of opinion within the Amiga community as to how the Amiga could best be adapted for the future. Perhaps surprising is how dramatic the difference is, in some cases causing deep rifts within the community.

In Focus

In this editorial, I am going to concentrate on the issue of "standardization," or, for example, whether, and how much, the Amiga should be changed to use more "industry-standard" technology. The motivation behind my use of all the quotation marks is, to be perfectly honest, a rather disdainful view of what the definition of "industry-standard" has been warped to mean in our situation.

Don't let anybody fool you. The Amiga's technology has not hurt its success in the least over the years. Very few people, when asked, will say they dislike Amigas; many have fond memories of them from either their initial successes of the late 1980's or their widespread use in the professional graphics and video industry from the early 1990's, and the rest will simply ask you, inquisitively, "What's an Omega?" The only people that really gripe about it, in general, are the ones that have been, and are using it, and for one reason or another, are continuing to do so. No, the Amiga's technology only helped it; what hurt it was a lack of both marketing and decent management.

Also, keep in mind that this entire concept of "standardization" is right in phase with the trend being perpetuated by Microsoft, Intel, and PC manufacturers that may eventually eliminate all alternative platforms from the market, leaving only PC's, and their compatibles. While the Amiga definitely needs to be updated, taking the whole thing too far, and turning it into just another generic PC-in-a-box, will certainly defeat our purpose. Yes, the computer industry is different today from the way it was years ago when the Amiga was young and times were good, but contrary to what some think, I believe the time is better than ever for the Amiga to return embodying its original philosophy of intelligent, efficient, powerful computing. Didn't we all more or less agree that despite some deficiencies, the Amiga was, and still is, the best computer platform? Didn't they always say the Amiga was ahead of its time?

The Custom Chipset

One of the biggest ongoing debates concerns custom chipset versus 3rd-party hardware. For a time, the most vocal people were those who favored the latter, but more recently, greater numbers of custom chipset advocates have begun to voice their opinions. If you have been following the AM Feedback column in recent months, you have probably seen the issue raised; and although it is clear what my opinions are, I endeavor to give representation to all points of view.

Some say the entire concept of the custom chipset is outdated, and would greatly impair a future Amiga's capabilities; the solution, to them, is to replace the chipset with hardware developed and produced by the independent graphics hardware manufacturers who have been developing for PC's for some time now. The idea I, and others like me, have supported is that the custom chipset has always been one of the greatest aspects of the Amiga, not only due to performance considerations, but also to those of versatility and integration. We believe that it can, should, and will continue to be so, and that the current immediate solution is to implement the AAA chipset, initially developed during the twilight of Commodore's existence.

The lack of readily available information, and some misconceptions, have been the primary reasons why a number of people have lost faith in the concept of the custom chipset. I therefore reiterate that yes, the current chipset design is limited and dated. I, of all people, realize this, since I am still using an ECS-based Amiga; and although AGA is a huge improvement over that, it is still hampered somewhat by bandwidth considerations, and mostly by the lack of chunky pixel display modes (and the blitter's corresponding ability to only move bits and not bytes). What many people do not realize is that the AAA chipset would not only significantly increase the bandwidth of the whole system, but also add chunky pixel modes, and full byte-movement capability. All things considered, this would make a AAA chipset-based Amiga quite a powerful system.

Some people cite cost as a reason to discontinue the development of the custom chipset. Folks, that just ain't the case. It may be hard to remember after the rollercoaster ride we've been on, but Amigas were generally available at very reasonable prices in the Commodore years. Three years ago, if you looked in the right place, you could get a base model A4000 for a little over $1000 US, back when it was still one of the best machines on the market in all ways. The custom chipset has always been an integral part of this design, and the price has not been exorbitant. The higher prices only came more recently, and even now, they're going down again. There is no need to worry about future custom chipsets costing too much to produce or manufacture.

I don't deny that some 3rd-party graphics hardware boasts better numerical specifications than the AAA chipset. But numbers aren't everything. To help explain my views, I will include the following, a section of text from an e-mail message that I recently sent to someone:

"Some PC hardware would beat AAA for sheer numbers, but the AAA chipset also has the benefit of backwards compatibility in hardware, planar screenmodes (despite the utility of chunky pixel modes, planar modes still have their place, especially for video), and the ability to run the Amiga OS with relatively little modification (this is a very big issue right now, since it has been so long since a significantly updated Amiga system has been introduced), to name a few. These are very big advantages, especially considering the fact that the AAA chipset would be quite an able performer, more than enough for most people's needs, even though it may not be the brute-force fastest chipset around. Besides, people must remember that the Amiga has always had an intriguing ability to "defy the numbers."

...[regarding AGA], I see no cause for alarm. The AAA chipset would be tremendously more powerful than AA (AGA). And even if it's not quite up to par with that, again, the numbers aren't everything; how you implement the technology, and how useful and usable it is, weigh very heavily.

I believe that the OS is the strongest part of existing Amigas, but historically, it was always the combination of the OS with tremendously powerful and well-configured hardware such as the 680x0 processors and custom chipsets that made it so great. I believe there is the potential for this to happen again. I do not like the idea of the Amiga becoming "genericized" (whether through the use of generic chipsets or turning it completely into the otherwise excellent aforementioned Siamese System concept), because it is what has been so unique about the Amiga that has always made it so great."

Again, for those who haven't seen it, the AAA chipset would bring 64-bit architecture, true 24-bit graphics, resolutions to 1280x1024, chunky pixel displays, a newer and more powerful copper and blitter (with byte-movement capability), and 16-bit, 8-channel sound right into the Amiga's native system architecture. Couple that with the Amiga's OS and a fast processor, and you have a force to be reckoned with, as well as the ability to run any reasonably well-written older Amiga software at high speed. What we would gain from 3rd-party hardware would not equal what we would lose.

Anyway, we'll use the processor reference as a convenient segue to our next segment.

680x0 or PowerPC, or What?

Another area of great uncertainty regarding possible future Amigas is processors. For a while, it seemed as if PowerPC was the clear route, but now people are throwing other ideas into the works, like the Alpha, etc.

My general word of advice is "step back and take a deep breath." Some people are so concerned with increasing the Amiga's processing power that they are trying to bite off more than we can chew.

I believe the PowerPC is the way to go from here. It is a powerful line of processors that is being developed and advanced, and it has fairly wide acceptance, among most modern Macs, for example. The Alpha is simply overkill, and too expensive to serve as the basis for a full line of consumer-oriented computers. Plus, it doesn't take much math to see that without its Cloak of MegaHertz, it's not as astounding as it looks.

And yet, despite this, I believe the higher-end members of the 680x0 series should still be used in the next Amigas. Why? Well, take a look at the situation. Considering the awe-inspiring efficiency of the Amiga OS, few people, with the exception of those who render graphics professionally, are truly hurting for processing power with the 680x0 series. What we could all use is some improved graphics speed, as AAA would provide.

So why not use PowerPC's now? They are faster, are they not? Well, yes and no. Once the Amiga OS is ported to the PowerPC, we can really get down to business. But now, it's still in 680x0 code, and running it in emulation on a PowerPC, as early Macs did, could very well result in an OS slower than the one we run on the hardware we use today. Notice that although Phase 5 is just about ready to bring their PPC acclerators to market, they are, by their very nature, dual-processor configurations. The OS still runs on the 680x0 chip, while new applications use the PPC as they become available.

The dual-processing concept is probably a good idea as an extra-cost option for new Amigas; not only would this give the machines more brute force rendering power, but it would also encourage a gradual move to the PowerPC.

A few things to keep in mind for future Amiga CPU's:

The Amiga OS: The Champion

Over the years, Amiga users have had to give up one "point of superiority" after another, as PC's and Macintoshes caught up in various ways. The Intels got faster, graphics cards began to outperform the AGA chipset, and even in its field of greatest acclaim, graphics and video production, the Amiga began to lose out.

However, they still have not been able to produce anything to equal the Amiga OS. What may be, for the time being, our last, greatest hope, may never in fact be surpassed by the competition, simply because the trend in OS's these days seems to be completely away from everything the Amiga philosophy embodies.

Above all else, any future Amigas must build upon a foundation of that which has always made the platform great. The OS may be the strongest link of all.

There's no question the OS needs improvement (speaking of the OS in terms of all the computer's system software, not just the lowest-level routines, which are fine). The GUI could be updated and improved somewhat, the system needs better printing and networking support, and the OS could stand to have a good deal more included software and utilities. Aside from that, however, very little has to be changed.

Although you have all probably heard it a hundred times, consider what the Amiga OS is: a 32-bit, preemptively multitasking, multithreaded, nearly real-time, object-oriented operating system with a full graphical user interface that, if necessary, will run easily and quickly on a system with only a 68000 CPU, one 880K floppy drive, and 512K RAM. That is simply unheard of in the world of "modern" computing. I believe the aforementioned needed improvements (as well as others like them) can easily be added without changing these fundamental attributes. How's that? It's fairly simple: the Amiga OS is incredibly object-oriented, or modular, if you will. A new, improved Amiga OS with all kinds of software, utilities, facilities, and capabilities could be shipped on a CD-ROM or larger set of floppy disks, with the core of the OS still present and usable on just a few disks. That way, the OS could run on an absolutely base-level model, or take full advantage of the power of a higher-end machine. If you think about it, it is done that way now, to a certain extent; you only install the printers and languages that you need, for example. The same methodology could be carried to a new level; all kinds of fonts, clipart, printer drivers, graphics, animations, datatypes, icon sets, etc. name it...could be included, if desired. People with systems to handle all that could use it, while others would simply install only what they needed in order to run the computer.

Some people advocate making huge, sweeping changes to the OS, rewriting it entirely, etc. I believe that is, in a word, absurd. Why would we want to change possibly the best thing about the computer, when it is one of the few things that still truly places it head and shoulders above everything else out there? The Amiga OS is something to build upon, not something to throw away and begin anew.

Some features have been discussed for inclusion into a future version of the OS. While many of them may be beneficial, there are some that I believe we must avoid altogether:

Once again, I will use an excerpt from an e-mail message I recently wrote to emphasize the point:

"....that whole field of memory protection/crash protection/resource tracking is a real sticking point for me. I generally oppose it. The reason is not only because they do take up system resources, but also because they have become a crutch for programmers. I've heard of "highly professional" Windows 95 programs that are constantly getting shut down by the system for illegal operations. Call me an anachronism, but I believe that any given application should have full access to all the hardware's power at any given time, no matter how fast the hardware, and that programmers should be responsible for getting their software free of bugs.

Also, I've used several OS's, and had a chance to compare. While the Amiga has no memory/crash protection or resource tracking in the strictest sense, OS3.1 has shown itself to be, in my view, remarkably stable, often able to recover from a crashed application and let you save your work and then reboot. It is especially so, as I have watched the evolution from the old AmigaDOS 1.2 through today. I have used two OS's that have these features (Windows 3.1, which has a rather mild degree of it, and OS/2 Warp, whose protection is quite good), and they share two common characteristics: they are slow, and memory/disk-intensive.

Add to this something I heard from a former Commodore engineer, about how resource tracking was originally left out of the OS because the 68000 simply could not have handled it. I fear that adding this to a new version of the Amiga OS would make any new machines perform like those of yesteryear, and essentially destroy everything that has made the Amiga OS so great.

If there must be memory protection, however, then I would propose, as an alternative, such a feature in the Amiga OS that could be toggled on or off. That way, everybody could be happy, especially since I believe the OS should continue to be able to run on the oldest Amigas, at least in a stripped-down, modular form. I know I'd probably keep the feature off most of the time. But if there were to be no user control over it, we would be much better off without it, thank you. It's just very frustrating to see that the trend in OS's these days is to make everything bigger and fatter, and I hope the Amiga doesn't go down that road."

This is more or less a sampling of some of the more well-known buzzwords describing features that I believe the Amiga OS should not have. However, there are others like them. We must only look at any proposed feature and ask ourselves, will it slow down the OS? Will it take up considerably more system resources, with no way of being shut down? If the answer to either of those questions is "yes," then it's probably a feature we don't need. If I want a heavy, fat, wasteful OS, I can go down to the department store and buy a PC at a cheap price. Let's keep the Amiga OS the way it is: incredibly fast, lightweight, user-friendly, and modular; and perhaps most importantly, unique for its strengths in the entire computing industry.

All Things Considered

In this editorial, I've taken a look at only three of the many major questions surrounding any future Amigas, should they in fact exist. And while it is all well and good to talk about chipsets, processors, and OS's, what is truly key is the underlying idea. The Amiga was founded, and continues to exist upon, a philosophy of intelligent, logical, powerful, and user-friendly computing at a reasonable price. For all the blazing fast processors and graphics cards, and "feature"-laden OS's, of some other platforms, few have approached, and probably none have actually attained, this ideal on the same level as the Amiga.

To accomplish the task of bringing the Amiga to its former levels of success and beyond, again, we must build upon everything that has helped the Amiga through the years. One of my favorite expressions, in reference to the Amiga's current situation, is "Don't throw the lifeboats over the side while the ship sinks!" In effect, by planning huge changes in the Amiga's architecture, that is what we are doing. The basic design philosophy is superior, and should be carried forth as it is now. By its very nature supremely integrated and modular, the current architecture allows plenty of room for improvement without revolutionary change.

The Amiga has been down a long road, and may have a significant journey ahead of it. The path to travel is the familiar one, lain in the days before Commodore acquired, and did its best to cripple, the platform. When you have something good, you don't fundamentally change everything about it to try to bring it back from the brink; you use what works.