Dave Haynie Interview with Amiga Blast
AB - Hi, and thank you for giving us a bit of your precious time.
Sure thing. I've always found that a bit more closeness to the end-user than you typically get on "other" platforms is extremely useful. This was something Fred Bowen, I, and a few others did our best at back on the Commodore 128, and tried to maintain throughout the life of the Amiga at Commodore. It's hard sometimes, when there are two or three full-time jobs to finish in a day, but I try :-)
AB - Please, introduce yourself and your tasks in your company.
I'm Dave Haynie. I worked at Commodore 11.5 years, first on some of the 8-bit Commodore systems like the C128, then on the Amiga. I was on the A500 for about a month, then took over the A2000. I worked on both A2500 models (the accelerator cards), I designed the Zorro III bus and bits of the A3000, I did the first AA and AAA prototype systems, etc. and so forth.
Since Commodore, I've worked days at Scala, which to some extent was the "Ex-Commodore East", given about 80% of the Engineering office is ex-Commodore, with the remainder somehow Amiga associated (GVP, independent development) (FYI: The 3DO Company is "Ex-Commodore West"). I've done various consulting projects in hardware as well, most ill-fated (silly me, trying to make money in the Amiga biz as it's shrinking).
Last November, Amiga Technologies hired me to consult on the Power Amiga project. I wrote a proposal, they liked it, and we started trying to do something. As time went on, it became clear that ESCOM didn't have the deep pockets we expected, and the project shut down before it really began. Shortly thereafter, much of Amiga Technologies was closed down. Out of the ashes of that arose PIOS. Our mission: to become successful as a PowerPC hardware vendor, and do what we can to deliver on the promise of the PowerAmiga ideas we developed.
AB - What are you and your company exactly developing?
Our goal is to develop "Amiga-like" PowerPC machines. At this point in time, there's little sense in re-inventing PowerPC systems from the high-end. Most of these machines are based on good, solid ideas, some from the PC Clone hardware market, and as in the PClone market, it's extremely difficult to improve upon what's being done and still remain competitive. So at this level, PIOS is a system integrator, using what's available to deliver on the idea of a range of PowerPC, multimedia-friendly machines.
However, since the Amiga's decline, the industry as a whole seems to have forgotten the low-end. The PC Clone business serves the low-end primarly buy dumping out "obsolete" machines at discounted prices (anything below US$1000). We feel that it's possible to deliver a PowerPC system, based on current technologies, that will sell in this price range, picking up where the A500 and A1200 left off. That's not to say we're going to build anything that PHYSICALLY looks like an A1200, for example. But I think our PIOS ONE system, in development now, will be something of the spiritual heir to this whole idea. I've always loved the idea of the home computer, and think they should be affordable to kids, schools, and in general, regular people who don't necessarily have US$1500-US$3000 to spend on a system.
AB - When should we see the first results?
We will show a system at CeBit in the Spring. Very possibly before then.
AB - How much is your machine Amiga compatible?
Well, being a PowerPC machine, it's not, right off the bat. We've investigated Amiga emulation technologies, and believe a commercialized version is definitely an option. However, the ultimate level of "Amiga compatibility" can only be reached though an AmigaOS port to the PowerPC. At Amiga Technologies, Andy Finkel outlined a plan to achieve this, including a plan to define a Hardware Abstraction Layer (HAL) for the PowerAmigaOS. This is the means by which a single binary OS release can run on any reasonable PowerPC machine, including both new systems, Mac compatibles, and upgraded Amigas. A PowerAmigaOS would also, by definition, include efficient emulation of the 68K binaries and environment, much as the MacOS does today.
AB - And what can you say about "binary-compatibility", shall we expect to be able to run our software on your new machines?
Not yet. PIOS is primarily a hardware company, and as such, we're not developing an operating system. We have offered to help VIScorp, or any new owner of the AmigaOS, in a PowerPC port, but so far, this project has not started. Assuming VIScorp gets the Amiga technologies (today is the deadline), this may change. Until there's a PowerAmigaOS, you will have some options for Amiga emulation on PIOS machines, but they're fairly slow at the moment.
PIOS' plan for an improved emulation is really contingent on the PowerAmigaOS project. We see this as a bridge to the new OS, not an end in itself. It would be wrong to call ourselves "Amiga compatible" if there is no real, native AmigaOS coming in the future. After all, you can, for example, run UAE on the MacOS today (PIOS is shipping MacOS on the first PPC systems), but few would claim this gives you a workable Amiga.
We do expect to be able to run a variety of operating systems on the same hardware, so users will have options. The MacOS is popular, but we don't see it as a viable replacement for the AmigaOS.
AB - Can you tell us some hardware specifications of your new machine?
We're using PowerPC, naturally. The first systems, OEMed, are based on the PPC604 and PPC604e. We expect to have some PPC603ev systems later on, also OEMed. The PIOS ONE will also use the PPC603 family, at least that's the plan -- as with Amigas, we expect a fairly reasonable way to change the CPU. We're also endorsing the PCI bus as the primary expansion bus in these systems. It's not quite settled whether the PIOS ONE is using PCI technologies or not, but as a small form system, PCI slots are likely to be an expansion option. Again, some details are not yet settled.
[FOLLOWIUNG QUESTION _ONLY_ IF YOU ARE USING A MOTOROLA RISC]
AB - Is your machine Mac PPC compatible too?
The first PIOS systems are. It's not certain if our low-end system will run MacOS off-the-shelf, though it's likely to be available as an upgrade, possibly even via a ShapeShifter-type system.
AB - What about developer support? Can we write native software for your new machine right now?
As I said, we're not currently in the OS business, and look upon OS development primarily as a strategy to support our hardware. So at present, you develop for the PPC OS you want, run it on our hardware. Since our low-end system won't be a "clone" of something standard, we will certainly have more details on it available to developers as time progresses. Especially of interest are OS ports.
AB - Is there any Amiga RISC cross compiler avaible at the moment? Do you plan to release one of your own?
No. And we don't currently plan to make our own compilers.
AB - Where do "Copper" and "Planar" concepts end in your new architecture?
Current systems use standard SVGA-type graphics, which support planar displays from 1-4 bits in depth. Though for the most part, there's no great reason to run planar graphics. Even most of the graphics cards on 68K Amigas use chunky pixels for 8-bit display modes, and beyond.
AB - Will your machine have a MIDI port by default?
Not all of them, though early machines do have the Mac high-speed serial port, which hooks to a wide variety of MIDI interfaces. Some of our systems are very likely to have MIDI built-in, we have a good bit of musical interest at PIOS.
AB - What is your target of the market? And your ideal user?
We're targeting the general market to some extent, but I think the emphasis will be on multimedia systems, for specialized use at the high end, for consumers at the low end. In other words, the Amiga's traditional strengths -- we understand these markets.
AB - Is your machine easily expandible/upgradable?
The OEMed machines have normal PCI slots, as you would expect of any desktop system these days. The PIOS ONE will have some form of expansion, it's still being discussed. It'll be at least as expandable as the A1200, hopefully much moreso. AB - Will your machine have internet/intranet/net support by default?
Yes, we plan internet support eventually across the line, including bundled net software, Java, etc.
AB - What about the OS of your machine?
Currently MacOS, because its there. We're investigating alternatives, and still hope a PowerAmigaOS will come to market as soon as possible.
AB - How much will your machine cost?
Current systems are at the high-end, in the US$3000-US$5000 range or thereabouts, depending on what's in them. Ultimately, we'll have complete systems well below US$1000.
AB - Can we have multi-processors mainboard on your machine?
Right now, we don't have any processors on the mainboard. Possibly we will, but we're definite fans of the Amiga's modular CPU card concept, which I helped to pioneer on the A2000. The designs to come will definitely support multiprocessor modules; assuming an OS drives them properly, this is a great thing.
AB - Is there any coprocessor?
All systems will of course have blitters of some kind. We are looking into various types of alternate processors, but it becomes true that, in most cases, a second CPU buys you more use than anything but a very low cost or very specialized auxilary processor. So support of multiprocessing is high on the list. We belive this is even possible under the AmigaOS, at least once ported to the PowerPC.
AB - Is there anything you wish to say to Amiga Blast readers?
We're trying to what we can to keep the Amiga, and the ideal it represents, alive. That's not been easy, especially for a company like PIOS, which started last May based on investment money. We have no choice but to ship systems now, or go out of business, and unfortunately, the MacOS is the only game in town. Some Amiga users have accused us of "selling out", or "promoting the enemy", but this is simply not true. First of all, the MacOS isn't the enemy (you know who is), and secondly, we're simply doing what we can do to remain in business, and hopefully grow. We will have other OSs, incluing the AmigaOS if at all possible, when it's in our power to deliver on this.
V.P. Hardware Engineering
"...no RISC, no fun"