Dojo (HowTo)







  Easter Eggs




  Martial Arts

Exponential Article:
Riddled with mistakes

By:David K. Every
©Copyright 1999

There was a recent article about Exponentials failure in Electronic Business. While it was a pretty good article, as far as interest, it was unfortunately riddled with mistakes that reflect poorly on the journal, and of course Dean Takahashi (the author).

In defense of truth and balance, I would like to address those errors -- while, for the most part, I am going to ignore the horribly inbalanced bias that was present through-out the article.

The article begins with background on Exponential (the company that was trying to make a 533MHz PowerPC), and then has this quote.

Rick Shriner, the company's CEO, joined the engineers in a champagne toast. Then he went home to brood about why, from his perspective, Cupertino, CA-based Apple had killed his company. Exponential's vaunted chip, code named the X704, never made it to the market.

The opening sets the tone for an article, as every journalist, or even Junior High School newspaper writer, knows. The tone for this one is set around how Apple killed this poor ailing company -- while ignoring how it would have been in Apple's best interest for them to succeed.

Whenever a human being fails, the first thing most do is look for a scapegoat, someone else to blame. We are supposed to realize this, and put a "filter" on their opinions -- Journalists, even more so. I am disappointed that Dean Takahashi did not do this.

In a more bitter moment, Shriner, a former Apple vice president who has the look and manner of an aging school principal, says he still loves the Macintosh computer, but feels betrayed by Apple. "I feel badly that the Mac is dying and going away," he says. "But after experiencing this betrayal, I think, the sooner the better."

I am left to wonder, what value this whole paragraph has to the article, other than to bash Apple? Where is the journalistic responsibility for balance? You can look through the article, and see no indications that Apple is still a multi-billion dollar company with money in the bank, and many indicators that they are actually thriving (showing a profit, releasing their best products ever, and so on). But I digress -- this article was obviously not about balance or truth -- it seems to be about misappropriated blame.

The article discusses the joys of our Legal System -- the suit and counter-suit, and so on.

Other industry commentators offer a more concise explication: "It's clear that Exponential got caught in the middle as Apple took on the clones," says Linley Gwennap, editor of the Microprocessor Report in Sunnyvale.

If I was Linley Gwennap, I would either be very annoyed at this fragment quote that couldn't possibly mean (out of context) what this implies -- or else I would be terminally embarassed by the foolishness of this statement.

While there is little doubt that Exponential finally faded out with the end of cloning, there is no doubt that Exponential had already flamed out 6 months to a year before, and it had nothing to do with cloning (as a "source at Austin, TX-based clone maker Power Computing states, 'Their failure was in their own hands'"). Either Gwennap understands this and was taken out of context (in which case we should all be very annoyed with Dean Takahashi), or Linley Gwennap should be embarassed to NOT know this and be hiding under a rock.

The quick summary of what happened to Exponential is this:

  • Exponential promised to deliver (in volume) 533MHz PowerPC chips by mid to late '96.
  • They delivered a very few 410MHz PowerPC chips in early to mid '97, and they weren't ready to go into volume production for another 6-8 months. All after going way over budget, in heat, power, dollars -- while underperforming, and so on.
  • In the mean time, IBM had produced a 350MHz 604e (with a superior cache, and other things) that actually out-performed the Exponential chip (the x704). The Mach5 used far less power, and didn't require customized hardware, special power-supplies, special cooling, and so on.

When it comes to performance and computers, you have a very narrow window in which to achieve your goals, or Moore's law (the concept that performance is doubling every 18 months) will eat you alive. The fact that Apple, Motorola and IBM have actually been doubling their performance every 12 months makes this even more time sensitive. Exponential missed their window. Apple recognized it. IBM recognized it. Motorola recognized it. PowerComputing recognized it. I recognized it. And about every other semi-technical observer also recognized it. The only person who seems not to have recognized it is either Dean Takahashi (which is sad, since he is supposed to be a technology writer), or if the quote is accurate, then Linley Gwennap failed to recognize it (which would be a major embarassment for him, and Microprocessor Report which he is supposed to represent).

Apple had a lot to gain by using a chip that had higher MHz and higher performance than Intel. The article discussed how Apple brought Exponential executives with them to meetings in Hollywood, and how they not only tried to help them get funding and attention, and that Apple invested its own money in the company. Apple had every interest in Exponential technology, IF Exponential could deliver on time, and on target (performance). Exponential missed on both, and by the time they were going to be ready, it was too late.

So why a whole article devoted to "nefarious" undertones about how "dying" Apple screwed over someone who would have only helped Apple? This is not a factual article trying to document what happened, this smells more like some personal vendetta by Rick Shriner (head of Exponential) to blame someone else for his failures -- or it could be a vendetta by Dean Takahashi (or the editors at Electronic Business) to slam Apple for some past wrong, or to pay off their anti-Apple buddies. I just have a hard time beleiving that they could be that wrong or incompetent. But then they do things that show that maybe the editors and Dean ARE that incompetent -- like this quote:

The Mach 5 ran at a slow 250 MHz compared to the Exponential chip, but it included a large cache, or secondary memory, with a fast datapath into the microprocessor. That innovation actually allowed the Mach 5 to run faster than the Exponential chip because it didn't need to fetch data from main memory nearly as often. Rubenstein would later favor this chip, which Apple calls the G3 processor, but IBM could not deliver it for months.

The Mach5 was IBM's die-shrink and speed up of the 604e, which is NOT called the G3. This chip was running at 350 MHz, and did include a large cache (which the x704 did not have) -- but it did not have the "fast datapath" that I beleive the author is alluding to. He is talking about the Backside cache of the 750 chip. The 750 was also released a few months later, and it also out-performed the x704 -- but this one did it in something like 1:20th the power budget (which also related to cost, size, and so on). IBM delivered the Mach5 (604e) many months before the Exponential chip could be delivered, and the 750 was delivered around the same time that the x704 would have been released. Both (Mach5 and the 750) did not require special computer designs to handle the heat, and they didn't require as software tweaks as extensive as those for the x704, since both were modifications of chips that Apple was already using. (The Mach5 was a 604e revamp, and the 750 was the next generation 603e).

So, given how twisted the story became at the hands of Dean Takahashi, his editors and Electronic Business, maybe it all was just a bad story, written from a position of incompetence, and the multitude of errors were accidental. Had they had the slightest understanding of the technology and history, the whole premise of the article would have been different. Unless, of course, the goal of the article was not to report facts, but instead to slam Apple, and defraud the readership in the process. While I am not a conspiracy theorist, the final paragraphs are suspect:

But in the process, Exponential picked the wrong track, the Macintosh market, and bet on the wrong horse, Apple. It managed to survive for quite awhile, no small achievement since it had chosen as its habitat the collapsing ecosystem of a once-great company. But that's not how the fittest survive. Unfortunately, the business world does not hand out awards for good tries.

This ending, again, shows either incompetence, or a more insidious plot. Since, as anyone who understands the technology knows, it was a far more cost effective direction for Exponential to go with the PowerPC for the following reasons:

  1. PowerPC was RISC, and so uses less power, and is easier to design, than a bad CISC architecture (like Intel's x86 and Pentium).
  2. It would have cost Exponential probably 4 - 10 times as much to get a working Pentium based on their techonology.
  3. If Exponential had been able to get an x86 version working, they would have been dealing with a far more saturated market, and far more competition.
  4. If Exponential raised the power from 3.5w (PowerPC) to 65-85w for the PowerPC and RISC, imagine what they would have done to PentiumII's CISC that start at 16-35w. They would have created an oven -- and been trying to compete with Digital's Alpha chip in the same market.
  5. Apple not only supported them (the opposite of what would have happened with Intel), but Apple was where they got some of their talent and funding, and most of their marketing and support.

So the whole premise that Apple and the PowerPC was the wrong track, is simply stupid. Exponential had no other choices -- and they chose the right path. I am so amazed at the business incompetence of not understanding this, not to mention the electronics or technical incompetence -- and this from a company that calls themselves "Electronic Business"? That is why I am left to fear that an insidious plot or vendetta as the only reasonable explanation -- which makes me wonder if the local District Attorney should investigate "Electronic Business" for fraud (or collusion) or instead go after then for libel.

Since business doesn't hand out awards for good tries, I would hate to imagine what business will do to bad tries -- and this article's premises are all very bad tries. If this magazine can't offer a better service than this article, one will be left to wonder whether the whole magazine is just another bad try.

Created: 05/28/98
Updated: 11/09/02

Top of page

Top of Section