Dodging the falling sky

September 17, 1997

I've been having a hell of a time getting a Rumor Mill written - the events at Apple in the last month are DIRECTLY relevant to the Amiga's future (at least where the PowerPC is concerned) and every time I wrote 75% of an article, Steve Jobs would do something else stupid and I'd have to start over.

So I've given up trying to keep up. Ric Ford's MacInTouch is a good resource for Mac gossip - it should more accurately be titled Ric Ford's Mac Page & Rumor Mill because it's the Mac equivalent to what I do here. I will summarize the broad strokes of what's been going on, though.

When Apple first licensed the Mac OS to Mac clone makers like Power Computing, the agreement was for the various versions of Mac System 7. What recently shipped as Mac OS 8 could just as easily have been called 7.7 or 7.9 - but it wasn't. End result: all the clone makers needed to renegotiate. They agreed to sickeningly inflated OS prices up to 12 times what they were paying for 7.x, the agreement was finalized, all that remained was for Apple to finish their side of the paperwork - and then Gil Amelio went bye-bye.

Steve Jobs hates Mac clones. He sees nothing wrong with Microsoft swiping every last Mac innovation and not doing a particularly good job of it - but for some reason, the notion of selling OSes, ROMs, and motherboard designs to other companies to expand Mac OS market share grates on Steve's nerves. But then, he probably still shudders at the notion of a Mac with expansion slots, or of a Mac whose case can be opened without having to order a special screwdriver from Apple.

Anyway, it wasn't enough that Apple just let the negotiations die. No. Like idiots they thought Power Computing and others would sit and let it happen - and they were wrong. Power, for its part, rallied the Mac world behind them and tried to get Apple to see the light - only to have Apple just buy them.

I watched this and wondered which clone maker would be next. This Power Computing is NOT the European company who makes Amiga floppy drives, btw. I liked these guys - the Power 120 here at work is a dream machine, nice roomy A2000-like case with eight SIMM sockets, three front-facing drive bays, three NuBus slots, and a total of three fans to keep an airflow inside the case. It cost about $2000 less than a similarly powered Apple system would have cost at the time. Apple has done itself - and the world - a disservice by ending that company. But I figured maybe it was just because it was Power, and its outspoken (former) CEO Joel Kocher. Apple knows better than to piss off its other licensees, like Motorola and IBM, right? Kinda hard to make Macs without chips these days, y'know - Exponential's long gone, and making Macs out of the Signetics 68070 (a 68010 clone) is a leap backward.

Then Apple pissed off Motorola. First they dumped their CHRP plans - leaving Motorola with an unreleased and largely unsellable CHRP system. Then they bought out Motorola's Mac OS license - thus ending the Motorola StarMax line of Mac clones. Then Steve called Motorola and said "We want cheaper chips. To the tune of, let's say, half off."

I fully expect Motorola to cut Apple dry.

IBM is taking a similar stance. Moto and IBM have announced they've reaffirmed their commitment to embedded PPCs - which are bigger business than Apple ever was anyway. (The MPC801, for example, is the chip that drives the WebTV. Your new sportscar is likely to have Motorola Inside and have higher integer benchmarks than a Pentium.) But all indications from inside and outside the companies is that they'll be throttling back to an unknown extent both production and development of the 600 and 700 series PPC chips used in the Power Mac.

You're following this, I hope. One reason the Amiga is trying to move to another CPU from the 68000 series is that Motorola no longer develops those chips for desktop computer applications. The ColdFire and 683x0 chips run printers and cars and stereos just fine but don't work so well in Amigas. Now it's EXTREMELY LIKELY (emphasis on this) that the PowerPC may find itself in a similar rut before the end of the year.

"But what about Exponential" I hear you say.

Time to rant. I've mentioned Exponential's death here before. It's been ALL OVER the papers and the Web. And yet nobody's heard it.

The story is this: Apple was Exponential's biggest investor. Exponential makes this groovy bipolar/CMOS 500MHz PPC compatible monsterchip. They finally get it ready for production - it doesn't quite hit the 500MHz mark (but it comes close, at 433MHz) - and Apple says "nope. Sorry." The "G3" chips from IBM and Motorola, the next generation PPC chips, approach those speeds at similar cost. Apple thus refused to buy Exponential chips, make Exponential-based Macintoshes, and most importantly, provide Mac ROMs patched to work with the Exponential chips, so no Mac cloners could either. This was several months ago.

The very next day Exponential closed its doors.

Last month someone bought Exponential's patents at an auction. Among those patents was one for a technique of sharing registers across two instruction sets on a single CPU. Some possibilities: PPC and 680x0? (Why? Emulated 68K code on an Exponential chip would have been faster than a real 68K anyway.) PPC and Intel X86? Far more likely, given who attended that auction, are either: nothing (Intel could have bought them to bury the technology), HP PA-RISC and Intel X86 (AMD or Cyrix could use such technology to make a clone of the upcoming Intel/HP joint Merced project, which runs both Intel and PA-RISC code), or Digital Alpha and X86 (Digital is reported to have been at that auction).

In other news Exponential's survivors are suing Apple for what happened. I think it'll be entertaining.

So let's review.

• The Exponential superfast PPC clones won't happen. Ever.
• CHRP won't happen.
• Motorola and IBM may punish Apple for its actions by suspending the PPC itself.

The PowerPC is no longer a valid choice for a future Amiga CPU. It's that simple. This is what it looks like today. Tomorrow it may get even worse. The PPC is falling from the sky and I don't think the Amiga should be under it. Argue with me if you like.

The only remaining reasons for the Amiga to go PPC are:

• Power Shapeshifter or Power Emplant/Fusion. Apple may very well take offense to this if it achieves any kind of market share - and it might not even technically be possible. You thought swiping 512K Mac ROMs was a challenge - just try swiping a 4MB Power Mac ROM set. Power Computing died for the rights to those ROMs.
• Phase 5. Wow. They make great 680x0 cards. Do they set the Amiga's direction? Only if we let them. Does anyone really think the A\Box is going to be compatible with any future Amiga? Does Phase 5 consider it a priority?
• CU Amiga says so. Last month's big write-up on the Power Amiga I found appalling. It was loaded with inaccuracies, obsolete information, distortions, and outright fabrications. The same crappy distorted journalism so prevalent in the PC world has found its way to the Amiga.

I almost referred to Apple the other day as "Cupertino Infinite Loop." Nothing wrong with that - it's the address of their corporate headquarters, 1 Infinite Loop, Cupertino. But look at the initials. CIL. Commodore International, LTD?

One way or another I think Apple should move its headquarters to 1200 Wilson Drive, West Chester Pennsylvania, and complete its transformation.

I've seen "bare" Alpha systems - 166MHz 21064, no RAM, no drive, video card included, case, power supply, keyboard and mouse - auctioned off at $300. I've also seen Pentium motherboards, board and chip and RAM, brand new go for $500. Those are the prices you'd expect to pay for used A2000s. Find me a Mac for that price that's newer than 1989. Consumer-level, consumer-priced systems in the sub-$500 range are possible even without the PPC. Intel ain't the best choice in this arena - but it is a choice. Alpha is a great high-end chip but not such a hot performer at the low end - but that could change, if Digital sees a demand for a low-cost, low-power 32-bit consumer CPU that runs Alpha code, they'd probably build one. Another possibility is the Digital/Acorn StrongARM - the chip at the heart of the Apple Newton Messagepad 2000. 167MHz that lasts eight hours on two AAs.

My only concern is that high-end Amigas and low-end Amigas should run the same code. I don't want to see a segmented Amiga platform - low-end StrongARM machines and high-end Alphas that don't meet in the middle. The two primary markets for Amigas are home tinkerers (sub-$600) and video production (above $2000) - but by having both systems run the same code, the potential to make a dent in the remaining "middle ground" now owned by Intel is there and mustn't be ignored.

One final note. We ran a log analyzer against the Rumor Mill logs and found that the #6 most active organization visiting my site was - of all places - Microsoft Corporation. Now, I've known for years there are Amiga fans inside Microsoft - some of whom arrived with Blue Ribbon Soundworks, for example - my hat goes off to you guys. And I've long known someone at Microsoft visits here on a regular basis. But I find it a little eerie that Microsoft proxy addresses are clocking more hits here than users on Netcom and PSInet and other large nationwide ISPs. I've joked on occasion about "Bill Gates read on the Rumor Mill that this particular tactic wasn't working so he changed it" - maybe I hit the mark a little closer than I thought.

It may be nothing - but it's worth noting I have never seen an apple.com proxy address in the logs. Are these visitors friend or foe?

Bill? Is that you?