We hope you packed your hip waders, people, because it seems we're at least knee-deep in bull hockey and the level's rising fast. Remember a couple of weeks ago when the Mac world was all abuzz over this CherryOS thing? It's allegedly a PowerPC emulator for x86 that's capable of running Mac OS X on Wintel hardware just like PearPC can, except that while PearPC currently runs "about 15 times slower" than the x86 on which it runs, Maui X-Stream claims that CherryOS can run as fast as 80 percent of the hardware's native speed. This is understandably exciting stuff-- except that no one's actually seen it work. The 1.0 version that was (maybe) initially posted for download has since evaporated and the product has allegedly gone back into "beta release for selected users."

Suspicious? You should be, especially since it appears that nobody who has managed to get a copy of CherryOS has successfully gotten Mac OS X running at all, let alone at the insanely high relative speeds it's supposedly capable of achieving. More to the point, the software has a single developer (one Arben Kryeziu) who claims to have pounded this thing out entirely on his own from scratch in (cough) four months. Considering that Microsoft evidently considered the task of writing emulation software to be tricky enough that it was better to buy all of Connectix than to assign a few of its gazillion programmers to spend three or four weeks developing their own emulator, well, yeah, we're a little skeptical.

And don't forget Arben's initial insistence that he wrote the product himself and "did not use any PearPC code"; it wasn't long before somebody noticed that CherryOS clearly uses a variable name identical to one buried way down in the depths of PearPC's source code. Whoops. Even more suspiciously, Arben's original explanation for the variable name was that "names are going to be similar or identical because there are only certain ways to do things"-- but the name in question is "SPIRO MULTIMAX 3000," a jolly bit o' nonsense that has nothing whatsoever to do with the variable's function.

Confronted with that fact (and other instances of obvious PearPCery in CherryOS, like the fact that both have emulated IDE controllers named "Ein Gebuesch Media" in supposed reference to a Monty Python sketch), Arben changed his tune; according to another WIRED article pointed out by faithful viewer xxx, Arben now admits that there's PearPC code in the version of CherryOS that people are looking at, but claims that the version in question "is not CherryOS. It is a premature CherryOS. Basically, it was a very bad version."

Well, okay, that may explain why no one's gotten it to work, but it still doesn't explain what the PearPC guts are doing in there. So here's the Gospel According to Arben: "the inclusion of PearPC code was the fault of one of his programmers," and Arben "fired his ass" for having snuck it in there. Interesting. But if this superfast PowerPC emulator for x86 was written by one programmer from scratch in four months, who's this other developer? If he was one of the "couple of programmers" that Maui hired to create a standalone version of CherryOS that can boot without Windows, why would he have replaced entire chunks of code of the allegedly-working and superfast CherryOS with vast wads of PearPC when PearPC ran more than an order of magnitude slower?

More to the point, if the version of CherryOS that Arben now admits contains PearPC code is a "premature CherryOS" (which, we assume, means one that predates the 1.0 release that appeared on and then vanished from his company's web site), how can it be the fault of another programmer that was hired to continue development after Arben allegedly wrote the 1.0 version of the software from scratch? Time machine?

We'll be perfectly happy if we're proven wrong, folks, but Arben contradicts himself over and over again, and there's still no "real" product to back up any of his claims, so we can't see how this is anything but a hoax. We'll know one way or the other in a month, though; Maui now says that a free trial version and the "real" 1.0 software will be available for download one month from today. Any bets on another "unexpected delay"?