So this one’s going to have a short lifespan, given that come Monday’s WWDC keynote, all of this weekend speculation is mootBut, come on, how can I resist?
Late Friday — 5 pm Pacific — CNet published an article by reporter Stephen Shankland, and the title pretty much said it all: “苹果放弃IBM，转而使用英特尔芯片”。
It’s essential to note that neither CNet nor the Journal are reporting this as a rumorTwo weeks ago, when the Journal first published something about this, they left themselves several outs: the only thing they reported as having happened for certain is that Apple had met with Intel and was considering the switchThey explicitly mentioned that talks could fall through; thus, regardless if it panned out, they could claim the article was accurate.
This time, however, both CNet and the Journal are reporting this as a fait accompliBoth reports flatly state that Apple is moving the entire Mac platform to Intel processors, and Jobs is announcing it Monday at the WWDC keynote.
Apple has used IBM’s PowerPC processors since 1994, but will begin a phased transition to Intel’s chips, sources familiar with the situation saidApple plans to move lower-end computers such as the Mac Mini to Intel chips in mid-2006 and higher-end models such as the Power Mac in mid-2007, sources said.
Apple Computer Inc.is expected to announce Monday that it will begin shifting its Macintosh computer line next year to Intel Corpchips, people familiar with the situation said.
The move is a major change in strategy by Apple, a high-profile win for Intel, and a potential blow to International Business Machines Corpand Freescale Semiconductor Inc., suppliers of the PowerPC chips that Apple has long used in its Macintosh systems.
这意味着什么是非凡的必须occur: either Apple is going to announce the biggest hardware transition in company history (and, arguably,行业history), or, two well-respected news organizations have committed enormous blunders.
Some of you may be willing to chalk this up to the latter, on the grounds that you see all sorts of bullshit get published that doesn’t pan outThat’s true, but not in CNet, which has a pretty good track record, and certainly not in the Wall Street Journal, which many consider the most-respected newspaper in the world (editorial page politics being a wholly separate issue).
如果没有发生，这是一个灾难性的报道错误There is no “maybe” in either report.
不管怎样我认为不太可能that Apple is about to move the Mac to Intel processors, I think it’s even减likely that the Wall Street Journal would publish a story this wrong.
The only way this makes any sense is that there’s something else. Something bigNot that CNet and the Journal have the story wrong, but that they only have part of the story — and the part they don’t have is what’s going to knock our socks off.
If there’s nothing more to the story — if it’s just a case of Apple switching the Mac from PowerPC to x86 Intel processors — it makes no senseThere is nothing in these reports that addresses any of the issues that make this switch so improbable.
For one thing, if they announce tomorrow that they’re moving to Intel processors, and but that high-end PowerMacs won’t make the transition until “mid-2007” — as reported by CNet — how will Apple avoid the奥斯本效应, wherein by pre-announcing future hardware, sales of current hardware evaporate?
(And why would Apple transition lower-end machines like the Mac Mini first? Apple has always introduced new processors at the high end, and they trickle down the product line over the course of a few years.)
它不仅仅是因为他们正在转换为踢It’s not performance — sure, G5 processor speeds haven’t progressed as quickly as Apple had hoped (and as Jobs had predicted publicly at WWDC two years ago), but that’s a problem that has stricken the entire semiconductor industry正如2005年5月30日的期刊所述MDJ, since the day of the G5 introduction at WWDC 2003, the top-of-the-line G5 has seen a larger percentage increase in clock rate (2.0 to 2.7 GHz) than Intel’s Pentium 4 (3.2 to 3.8 GHz).
Some argue that it’s about cost, but I have yet to see a single such argument backed by actual numbers — the cost per unit Apple would save after switching.
I can’t think of a single scenario, no matter how contrived, that both fits with what CNet and the Journal have reported and makes sense for Apple and its customers and developersBut I can think of a few scenarios wherein CNet and the Journal are mostly or sort-of right.
一个想法就是Apple是切换到英特尔处理器，但是不到x86Instead, perhaps, Intel is set to begin production of PowerPC processorsThis sounds unlikely, but is it any less likely than Apple switching to x86?
Licensing-wise, I believe (but I could definitely be wrong) this could happen, because of the rights Apple holds to the PowerPC platformIt would certainly make sense for Apple, because it wouldn’t involve a transition to a different CPU architectureIt wouldn’t be a switch; it’d just be a new addition to the PowerPC familyMac developers wouldn’t need to recompile or rewrite their software; Apple wouldn’t need to worry about people trying to get Mac OS X running on commodity x86 PC hardware.
它也可能对英特尔有意义The Mac is no longer the only major PowerPC platformAll three next-generation video game consoles — PlayStation 3, Xbox 360, and Nintendo Revolution — are based on PowerPC processorsGetting into the PowerPC business would not necessarily mean only producing chips for Apple.
This would further explain why these stories have only centered around Intel, and haven’t mentioned AMD.
但是：华尔街日报报道explicitly states that Apple is switching the Mac to x86:
Apple’s decision, which comes after years of industry speculation and behind-the-scenes lobbying by Intel, could cause disruptions for users of the MacintoshAmong other things, application programs will have to be adapted to run on Intel’s x86 chips, the calculating engines used in most personal computers that run Microsoft Corp.’s Windows operating system.
Software emulation of the PowerPC instruction set on x86 processors, with acceptable performance, still strikes me as completely unfeasible但是，如果我错了怎么办？
And, as pointed out by Christopher Ong on the MacJournals-Talk mailing list, Transitive’s board chairman is Peter van Cuylenburg, who — according to his传递网站上的传记— was the president and COO of a certain company called NeXT Computer in 1992.
With a fast enough emulator, the transition to x86 could be just as smooth as the 680 × 0-to-PowerPC transition a decade ago.