Apple to Intel: A Software Q&A

Well, it’s true. Weeks of speculation, capped by a furious weekend of blogging, emails and forum activity culminated in today’s announcement that Apple is indeed moving from Power to IBM.

Several weeks ago, when the Wall St Journal first hinted that Apple was considering migrating to Intel based chips, my parents emailed me to ask what I thought of the notion. Setting aside the fact that as anyone who’s dealt with me professionally can attest, I assiduously avoid any semblance of hardware coverage (unless it’s gear I can test out ;), I offered the following off-the-cuff response:

Unlikely, but would be a good move on their part. IBM’s set either way because three of their chips power the next generation XBox [not to mention other media devices], but i still doubt that Apple’s going to go to Intel chips.

I take some small consolation in the fact that I was in good company. I feel safe saying that the news of confirmation – broken last Friday by the enterprising Stephen Shankland of News.com – comes as a surprise even to those who follow Apple far more closely than I do.

Indeed, the groupmind was hard at work all weekend coming up with an amazing array of alternative explanations ranging from the highly plausible to the outlandish. In the end, however, Jobs did what people ardently believed he wouldn’t: he got Apple on the Intel bandwagon. Given that I’m not a hardware guy, I’ll leave the deciphering of the motivation(s) for the decision to others more knowledgeable in these areas. Instead, I’ll try and provide answers to the questions that I, as a software guy, had on the deal.

Q: Does this mean that OS X will run on generic x86 hardware, making it a volume platform competitor to Linux and Windows?
A: No. While I don’t have an official link for you, there seems to be general agreement on this point, and Winer’s very specific here. OS X will not be running on your Thinkpad anytime soon (or at least that’s Apple’s plan).

Q: What about the applications? The ISVs? What’s the impact on them?
A: This is the $64,000 question to me. The potential disruption to them was apparently the primary reason Apple didn’t make the transition years ago, and was the basis for my belief that we wouldn’t see Apple running on Intel. Jobs indicated this morning that many applications would run on the new platform unchanged, and others would run with a minimum of alterations.

For the major ISVs, the impact is likely neglible. Adobe and Microsoft have already promised to support the new platform, and I have little doubt that virtually every major ISV will follow suit.

The real burden will be on the little guys, folks like Plasq‘s atariboy. The new developer pack is apparently $1K, and they may be faced with lengthy recompiles or abandoning code entirely. Here’s how Darrin Cardani put it on a forum that atariboy pointed me at this morning:

All the high-speed AltiVec code I’ve written over the past 5 years now has to be rewritten for SSE/SSE2 which, frankly, isn’t as good…In the long run, it probably won’t make a whole lot of difference, but in the short term, it sure is a big pain in the ass.

I’ve seen a lot of similar comments. So to put it lightly, I’d view this morning’s claims of simple porting skeptically, at the very least.

Q: Will Apple lose developers over this?
A: Not so far as I can tell. As I considered the implications of the deal over the weekend, I thought it at least possible that this transition might be an opportunity for the Linux or Microsoft platforms to pick up new blood. But virtually everyone I’ve spoken with or track is still on board with Apple. They may be happy about the news, they may be disappointed, but almost universally they expect to continue developing for OS X.

This does not mean, however, that competitors will not attempt to leverage the transition, probably via the injection of FUD. The opportunities there are clear: anyone think your average OS X user wants to worry about “universal binaries”? Me neither.

Q: What does this mean for competing platforms?
A: Not as much as you might think. For all that it’s a profound departure for Apple, if it’s going to be limited to Apple only hardware I don’t expect it to have a hugely transformative effect on the industry. It will (perhaps, I’m not entirely convinced of this) make Apple’s hardware, and therefore its software, more competitive over the longer term. But short term, it’s still a closed platform, and will compete with Linux and Windows as before. If they’d released OS X to generic x86 gear, that would be an overnight game changer. But they haven’t. Doesn’t mean they couldn’t (and I’d be willing to bet that the folks from Redmond are thinking about this right now), but they haven’t.

Q: That’s OS X on other hardware; what about Linux or Windows on Apple hardware?
A: Well, Linux can already do this so I don’t see the possibility as all that exciting. But Winer does report that it’s possible that Windows will run on Apple’s hardware. While this is likely to excite legions of Windows users who drool over gorgeous new Powerbooks, however, I don’t see it as a huge channel for either Windows or Apple. An interesting corner case, but little more. From a hardware availability standpoint (Apple’s very restrictive with their gear) alone, it doesn’t make a ton of sense.

Q: Is there anybody else who might stand to benefit from the transition?
A: Desktop Java advocates are likely to use the opportunity to try and highlight the value of cross-platform compatability to an audience slightly more receptive than normal. I also expect desktop Linux suppliers to place some calls to the Power chip crowd (IBM, Freescale, etc), as by ’07 Linux will be the only desktop OS capable of running on both Power and x86 (unless Apple plans on maintaining its Power compatability indefinitely). It’s possible that Linux projects that maintain Power ports (like my own distro, Gentoo) might opportunistically target technically savvy Apple users that feel abandoned by Jobs’ decision.

Q: What about users? What should they expect?
A: Bit better hardware performance in the short term, minor complications in application availability and compatability, but probably not much else. If you’re looking for eBay investment opportunities, however, I’d start buying G4/G5 gear right now, b/c many Apple hardware owners are suffering from a case of Intel envy and/or buyer’s remorse.

Q: What’s your biggest question on the deal?
A: There are two, actually:

First, how is what Apple/Intel will be delivering that different than generic x86? I don’t understand this yet. We know that OS X will not run on OS X hardware, but what impact will the platform delta have from an application perspective? Will emulation applications, for example, experience difficulty because of the uniqueness of Apple’s platform?

Second, why Intel, and not AMD? I’m as mentioned not a hardware guy, but it seems to me that developers are more hopped up by AMD than Intel these days, and I have to think that AMD would be more attentive to Apple than Intel will. But it’s not my area; maybe it’s something like this.