Having spent some days in preparation, ARN's Matthew JC. Powell guarantees a splendid time for allYou can't help feeling sorry for microprocessor manufacturers, can you? OK, maybe it's a trifle difficult to shed too many tears for Intel and its billions, but what I'm talking about here is the poor shmoes who name the things.
Try and follow me here. Long long ago, Intel's chips were numbered, much in the manner of sunscreens. The 80286 was succeeded by the 80386 which was in turn followed by the 80486. All was rosy and everyone knew what sort of chip was better than what. Sure, there was some confusion with DX and SX et cetera, but basically you knew the score.
Then people started reverse engineering and ripping off -- um, I mean, emulating, Intel's chips. These were also allowed to be called 80386 and 80486 because, of course, you can't prevent people using a number. Think about it: every time 80,486 people turned up for a U2 concert, the band would have to pay a royalty to Intel for quoting its crowd figures, or else give someone a free ticket.
So Intel turned to actually naming its chips. Since the next step was 80586 (the difference being the five), the logical name was Pentium. Everyone knows the "pent" bit means five -- pentagon, pentagram, pentateuch -- so we're on safe ground. Intel invested buckets of money into building equity in the word Pentium, and it paid off. Other manufacturers continued to produce chips called 80586, but there was no longer any risk of customers or dealers mistaking these for the real McCoy. Even a vastly over-hyped division bug and botched attempt at cover-up did not significantly diminish the strength of the Pentium name.
But what then? The next generation is of course 80686. Six in Latin is sex. Sexium? Find me a reseller who could sell a product called Sexium with a straight face. You couldn't exactly sell it into schools either. Switching to Greek, six is hex. A hex is a type of curse. Let's not go there.
Next generation up; seven is sept in Latin and hept in Greek. Quite apart from the fact that Heptium sounds like a nasty skin disease, Septium sounds decidedly smelly. And imagine how your business would be affected by thousands of "deviated Septium" jokes. Someone just didn't think this thing through.
Once we reach eight -- Octium -- things are OK again. All the rev-heads are happy because it sounds sort of like "octane" so it's gotta be fast, right? Fine, but you'd be skipping a fair bit to get there, and people would ask why.
So we stick with Pentium, at least for now. Addenda to the Pentium technology have been named "Pro" (can't go wrong with Pro, but does this mean earlier Pentiums were amateurs?) and "II". Sylvester Stallone found long ago that roman numerals will do nicely in lieu of a clever title, and he's yet to be proved wrong. The only complicating factor is "MMX", which apparently stands for "Multi Media Xtensions". Memo to Intel: Multimedia is one word and Extension starts with an E.
That aside, there's a great deal of name value in Pentium, so Intel's on a good thing by simply extending that family rather than moving on in generations. The new complication, and the one which got me thinking about this column, is the push to low-cost CPUs. This push, apparently driven by consumer demand for PCs that cost next to nothing and can hardly do diddly, is driving Intel away from innovation and into scaling its technology down to a price.
I've seen no such demand, but Intel apparently has. It's pulled the cache out of the Pentium II to make the same chip, but a little cheaper and quite a bit slower.
I'm told it runs slower than a standard 200MHz Pentium MMX and costs more, but people will buy it because it's a Pentium II.
But here's the thing: it isn't. It's called -- wait for it -- Celeron. That's right, Celeron. Its supposed to make you think of acceleration but, frankly, it doesn't. Somehow, it gives me the munchies. I can't wait for Tomaton, Cucumberon and Lettucon. Then I can make a salad.
Off with a fair ease
Thanks to all who responded to my informal survey on the chances of speech recognition taking off in a big way by the year 2001 (see ARN, March 4, page 64).
I have started planning extravagant ways to spend my $10 winnings thanks to the overwhelming response. One hundred per cent of respondents reckon that three years is just not long enough for the obstacles to be overcome. Those who suggested an alternative time frame seemed to agree roughly on around 2005, while others indicated it wouldn't happen in this lifetime.
To paraphrase one respondent, it seems only Surfers Paradise developers know how to wreck a nice beach.