So has Steve's infamous Reality Distortion Field™ worn off yet? Because, you know, it was running full blast during Tuesday's Seybold keynote address; at one point during the proceedings, we got a reading of an unprecedented 7.6 on the Jobs Scale using our patent-pending Distort-O-Meter. (No, you can't borrow it. Stop asking.) That's not to say that some of the things he said weren't true, since the RDF is often employed to further enhance the intoxicating effect of even perfectly true good news. But it does mean that it behooves us all to take a step back from the heady G4 announcement, breathe deeply, and take another look at some of the more iffy claims that were made. In particular, let's talk about the awe-inspiring performance of the new G4 processor beating at the heart of Apple's new clear, silver, and "graphite" pro machines. While you're right to be excited-- very excited-- about gigaflop performance, there are some caveats to keep in mind.

A gigaflop represents a billion floating point operations per second, whereas most common computing tasks seem to rely on integer performance instead. And from what we can piece together from various stuff floating around the 'net, the G4 has integer performance roughly on par with that of the G3, scaled for clock speed-- nothing to sneeze at, but not nearly the kind of overall performance improvement implied by Steve. That may be the point Intel plays up when the company eventually tries to counter Apple's claims that the G4 is "2.94 times faster" than the fastest Pentium III; set a 600 MHz PIII against a G4 both performing a word count on a massive Microsoft Word document, and we bet the PIII will win hands down. (We hope we're wrong.) Plus, one of the biggest speed-boosting features of the G4 chip is the "Velocity Engine," the vector processor that Motorola calls AltiVec-- but AltiVec can't do its job unless an application is written to use it. And while adding at least minor AltiVec support is as easy as recompiling (taking real advantage of the technology involves coding specifically for it), what are the odds that Microsoft is ever going to let Office run faster on the Mac than on a Wintel? Yeah, right.

"But AtAT," we hear you saying, "what about all those great demos that Steve performed live? That 600 MHz Pentium III-based Compaq system got its butt whipped." Well, yeah, but as faithful viewer Tom Kepler points out, a CNET article quotes ABN AMRO analyst David Wu as saying that "there are lies, damn lies, and demos." Those demos all involved AltiVec-enhanced applications (Photoshop 5.5) and/or processes that rely heavily on floating point operations. We're not saying that Office is everything (heck, we don't even use it ourselves), but you shouldn't expect the G4 to be three times faster than PIIIs across the board. For Photoshop? Sure. For various science and engineering applications? You bet. For mind-bending 3D games? Almost definitely, though Steve didn't show any at Seybold. But for the average user, don't expect to finish your work day in a third the time of your PC-using colleagues if you get a G4. Not that we're trying to talk you out of buying one, by any means-- in fact, we'd rather talk you into buying one. Preferably, for us. If only we had our own Reality Distortion Field to exploit...