Is Boot Camp As Ingenious As I Think It Might Be?

I blogged this on 6/14/05:

On the other hand, if Apple ships boxes that can run OS X and Windows XP, but generic PCs are precluded from running OS X, Apple keeps (virtually) all of its existing hardware customer base, and adds to that Windows users who either a) are attracted to Apple's better hardware designs, or b) want the option of migrating to OS X later. All upside, no downside, excluding the Mac users that Apple loses to Windows, which I believe to be a very small number, and even some of them will hedge their bets with dual-purpose hardware.

Let's look at what Boot Camp has done with respect to that scenario. Basically, Boot Camp provides a quintessentially Apple approach to configuring a dual boot system: partition the disk, set up a boot loader to choose a boot volume, and install the alternate OS (Windows) on the new partition. The Apple touches: most of the installation runs in the familiar application mode (as opposed to a character-based installer), and you can re-partition your disk without destroying your existing data (it's been done before, but it's not common).

The interesting nuances: the process pretty much assumes that you are only going to install at most two OS's, and one of them is always OS X.

The bottom line is that Apple has made it very easy to buy Apple hardware and run Windows on it, but OS X always remains as a boot option. These constituencies are served:

Committed Mac users don't care; they don't use Boot Camp, and continue to use their Macs to run OS X.

People who would prefer to use OS X but need to use Windows in a pinch use Boot Camp to set up their Windows environment on their Mac hardware, and re-boot as needed to switch environments (as dual boot users do for other permutations, such as Windows/Linux).

People who would prefer to use Windows but are attracted to Apple's elegant hardware buy the hardware, configure it for dual boot, then set Windows as the default.

Apple has already made a killing: the loyal Mac base doesn't notice the change, a very small percentage of people who want to run the two OS's and are willing to deal with the complexities can do so, but the killing lies in the people who buy Apple hardware to run Windows. If you stipulate that Apple currently has 4% market share in system sales, if 1% of Windows users buy Apple hardware, that's a nearly 25% increase in Apple's hardware sales (where all the margins are). That's a compelling number all by itself, but there's more.

Remember, under normal circumstances, even those Apple machines that are running Windows all the time have retained their ability to boot into OS X, even if their owners aren't planning on doing so. The only downside for those users is that some portion of their hard-drive space is "wasted" on the OS X partition; but these days, hard drives are typically so big that this isn't likely to be an issue for a while.

Now suppose that at some point in the future, Apple introduces something akin to Wine, which enables Windows executables to be run from within the OS X environment. This is not wild speculation; Wine is already pretty robust under Linux.

At this speculative stage, we have people running Windows on Apple hardware who now have the option to boot their previously dormant OS X partitions and continue to use the applications and data that they have been using via Windows. All of a sudden, a Windows user who happens to own Apple hardware now finds it almost completely painless to boot into OS X and continue to use the Windows applications and documents on their Windows partition. These users lose (practically) none of the capabilities or data that they were using under Windows, and gain the additional (and, IMHO, attractive) capabilities of OS X. For these users, there is little or no downside to configuring their machines to boot into OS X instead of Windows, and if they find a problem, they can switch back to Windows without having lost anything.

As an admitted Mac zealot, I assume that a substantial percentage of Windows-centric Apple owners would choose to start booting into OS X, and start pointing out to their Wintel-using friends that they've achieved full compatibility on Apple hardware. And the already-beneficial trend would continue.

My Cup Runneth Over

I've had the MacBook Pro for about a month now, and it was worth every day of the wait. For most things, the perceived performance is on a par with my dual 2GHz G5, and there's no perceived difference when I run PowerPC binaries under Rosetta. (It should be noted that none of my performance-intensive work is done under Rosetta.)

In my spare time, I'm trying to get the MacBook set up to work with my HDTV:

- I've got the video connected with a DVI cable. The MacBook is capable of driving the built-in display and the HDTV separately, but FrontRow seems to require configuring the Mac to mirror the displays. Makes sense, assuming the long-term goal is to sell the Mini or something like it as a set-top box where the TV is the only display.

(By the way, what's with the price of DVI cables? Radio Shack had them listed online for about $12, and there was even an empty hook at the store so marked, but all I could find in stock at RS, Circuit City, etc. was $100+ DVI cables. Methinks someone is artificially jacking up the price as an "anti-piracy" initiative...)

- I still don't have sound figured out. The MacBook and the TV both purportedly support digital audio, but the only input jacks available on the set are what look like a normal pair of RCA jacks for analog audio. I used a cheap analog mini-plug-to-RCA-pair cable, but the sound quality is terrible, as expected.

This could become a dream setup: set-top box via Front Row, or settle in to the leather recliner with Bluetooth keyboard and mouse and a 50" display for "work"...

The Glass is Half Empty

It's not quite that bad, but I just learned that the ship date for my Mac Book Pro has been moved from February 15th to February 28th (with the processor speed increased, of course).

I'm disappointed, but the completely unrelated good news is that I was able to repair the G3 that was rapidly becoming unusable. The display had been crapping out (black was being displayed as red), and QuickTime display was illegible, but I found a replacement display on eBay for $75 (promising no red "washout"), and I was able to disassemble and reassemble the G3 in a couple of hours today and it's as good as its six-year-old new.

Obsolete before it ships?

Not really, but...

It's a running joke (or sometimes a genuine complaint) in the PC business that PCs become obsolete too fast, almost as quickly as you buy them.

I ordered a MacBook Pro a few weeks ago, and have been waiting for to ship "by February 15th", according to the estimate on the Apple Store.

Today, Apple announced that the new MacBook Pros are shipping, and a higher speed processor is available as a build-to-order option. If I had wanted the fastest possible processor, the machine that hasn't shipped yet would be less than I wanted.

On the other hand, I had opted for the slower (1.67 GHz) processor vs. the faster 1.83 GHz, and Apple's announcement includes the news that I'll be getting a 1.83 GHz model instead.

The glass is definitely half full.

Why did I opt for the slower processor? A few reasons:

My current laptop is a 400 MHz G3 (Pismo). Relatively speaking, the 1.67 GHz Core Duo would have seemed blindingly fast. (My other regularly used machine is a dual 2 GHz G5, so I think I can handle it.)

I deliberately ordered the G3 just a few weeks before it was end-of-lifed by what turned out to be the G4 Titanium PowerBook, and caught the price drop as Apple cleared the inventory. The G3 suited me perfectly, and cost me nearly $1,000 less than a new TiBook would have. Not that I didn't drool over the TiBook, but I was on a budget. I'm just not convinced you get the best bang for the buck at the very top end. (Don't ask about the G5...)

Similarly, I had economized on the Pismo and bought the 400 GHz model instead of the available 500 MHz; again, I never regretted it.

One of the best performance improvements I applied to the G3 was to replace the hard drive with a faster (and quieter) 5400 RPM drive. On the basis of that lesson, I paid for a 7200 RPM drive (vs. the now-standard 5400) for the MacBook.

I also decided that money was better spent on more RAM than GHz, so I upgraded to 1 GB on a single DIMM (to leave room for expansion). (The G3 was eventually upgraded to 640MB, a little beyond the Apple-specified "maximum" of 512MB.)

We'll see how it turns out. Soon, I hope.

b r r r e e e p o r t

Not sure why I'm playing along, but Scoble is asking for blogs containing "brrreeeport" to test blog search engines. Maybe we'll also get a sense of how many people really care...

Google and Privacy

Michael Arrington at TechCrunch notes with alarm the release of Google Desktop 3.0, which enables desktop search from across computers by maintaining a search index of your local files on Google servers. (It should be noted that use of the feature is optional).

Arrington links to a Gillmor Daily discussion about the privacy implications of this. I usually like listening to these guys, but I thought their opposing positions both missed the point, badly.

Arrington seems to be most worried about the potential for abuse by The Government, especially in the wake of Google's agreement to share information with the Big Bad Bush Administration.

Gillmor scoffs at the concern, not because he doesn't think the current administration is indeed The BogeyMan, but because (if I understand him correctly) They have more and better means to abuse our privacy anyway, and the value of storing our data in the internet cloud is so compelling that the way to deal with it is to manage how we manage ownership of that data, and specifically asks why Google's storage of indices is worse than storage of private data via online backup.

What I Find Troubling About This

To take the last point first, here's the fundamental difference from my point of view: in the case of backup, I build a package of my backup data locally, encrypt it to make in (hopefully) impenetrable, then move it into the cloud. In the case of Google Desktop, my data is moved to Google's servers in a form where they can peek into it, in the name of search.

I find this troubling. Google promises not to share my data with anyone. Fine, let's accept that and hope that Google is better at protecting it than the Boston Globe, of whose idiocy I was affected.

Being able to search my data and affirmatively determine whether it contains references to certain keywords conveys information about me, even if the content of that data is not explicitly shared. Google does not do this, as far as I know; they certainly don't let anyone to ask "who has references to Al Qaeda or Barry Manilow" in their index?" But there's a spectrum of possibilities between complete privacy and outright sharing.

As for government surveillance, challenging the government on its exercise of power is always healthy, but Arrington and Gillmor are so focused on it that they race each other to proudly label themselves as unpatriotic for wanting to retain their privacy; it's actually comical.

While it's legitimate to be vigilant to ensure that we don't become a police state over the next 20 years, I also try to make sure that many small, annoying things aren't done to me on a daily basis between now and then.

Up to now, Google has commercialized what they know about your interest as expressed by your search queries, and there's an strong element of fairness to this. If you go to Google and type "Barry Manilow" into the search box, you're transmitting your interest in the subject to Google, and they use this to deliver discreet ads with your search results that reflect your expressed interest.

But increasingly, Google has the ability to color their interactions with you based not on your expressed intentions, but on the interests implied by the content of your email (if you use GMail), and now, the content of your private files. I don't think they are doing that now, but the potential is there, and I don't see that their current privacy policy prevents that. Even if they don't publish lists that say "RetiredMidn is interested in these topics" or "RetiredMidn is one of the people interested in this one topic", my interactions with Google can now be adjusted based on the electronic equivalent of the contents of my underwear drawer.

Controlling Attention Data

Gillmor is an advocate of giving the user control of this information about themselves, as exemplified by the principles espoused by AttentionTrust.org, with which he is intimately involved. Their notion is that the user ought to have control of the information about their attention, what is included in it, and how it is shared.

In the discussion with Arrington, Gillmor seems content that users can exclude specific folders from Google's net-based indexing, and that this event will trigger the concern and discussions necessary to convince Google that they need to provide better control of this information. I'm less confident about this. Desktop search is most valuable when it can help you find stuff in obscure places, and constraining the search indexing reduces the usefulness. It also places the burden on the user to proactively identify all the places sensitive data is now or might possibly land in the future, and I don't think this is realistic. In theory, the user has the control; in practice, it's too easy not to exercise it.

Capability May Not Be Intent, But That Hasn't Stopped Us Before

Years ago, ActiveX came into being because the techies at Microsoft (and elsewhere) thought it would be cool and powerful to enable program code to be embedded in emails (and later, web pages). There were some clever ideas; when someone received an email requesting them to perform an action, the email could embed code that would perform, or support the user's performance, of the requested actions. There were all sorts of scenarios drawn up where this could be a significant productivity enhancement.

In hindsight, most people agree that this was a Bad Idea: used maliciously, this capability became a conduit, perhaps the most exploited one, for spreading malware. The consequences are that the innocent have been inconvenienced. For example, my wife is an elementary school teacher, and in the name of security she is now prevented from installing and trying out many potentially useful pieces of new software, which is what is supposed to make computers flexible and powerful.

Like Microsoft, Google is, I think, optimistically focusing on the potential value of desktop search across the network, and in their enthusiasm is blind to the potential for abuse. I hope I'm wrong.

To those who think, like Steve Gillmor, that it's alarmist to worry about the potential for abuse instead of the deed itself, I would point out that wiretapping and gun control laws are based on the same premise: good intentions aren't sufficient protection.

I'm not proposing that Google Desktop be outlawed. (When Google is outlawed, only outlaws will Google.) But I'm not installing Google Desktop (I use Macs anyway; small sacrifice), and I don't use GMail for my personal correspondence. (*)



(*) I do use GMail to collect the traffic from related public mailing lists, so I can search them in one place.

Apple with Intel Inside - The Beginning

Okay, Apple has rolled out the first two Intel-based Macs (and I've got a MacBook Pro -- that name's gonna take some getting used to -- configured in my shopping cart), but some of the real fun hasn't started yet.

Watch for solutions to run Windows Apps on Mac OS X

As Andy Ihnatko pointed out way back in June, you know somebody will devise a way to run Windows apps under OS X.

It could be a solution along the lines of VirtualPC or its predecessor, SoftPC. Although Microsoft acquired and now sells VirtualPC, both of these products were spawned from relatively small development efforts, and the technical challenge was much greater than it is now: both of these products had to run Intel binaries on non-Intel CPUs (the 680x0 and later the PowerPC). It will be much easier to build a solution where the code can run natively, not to mention much faster.

Another approach is Darwine, which is a derivative of Wine. Wine enables (many) Windows application binaries to run under Linux on x86 architectures by implementing Windows APIs. It avoids the processor emulation challenge, although I would think that keeping up with Redmond's creation of new APIs isn't all that easy, either. Most notable about Wine is that it does not require a Windows license to run Windows apps.

Darwine started out as a port of Wine to Darwin, the open source Unix foundation of Mac OS X. The project was challenged by the architectural differences between Linux and Darwin/Unix, as well as the processor emulation problem that Wine does not address. With the introduction of Intel-based Macs, some of that challenge goes away.

And, of course, there's always the option of dual-booting Macs into OS X and Windows. Unwieldy, but it may be sufficient for some people.

Oh, and then there's the possibility of buyying an iMac or MacBook just to run Windows, without OS X. There's precedent for this: a fair number of Linux users run it on Apple hardware.

A bigger market for Apple hardware

The upshot of these new possibilities, in my opinion, is that Apple has a potentially larger customer base for its hardware. The Mac loyalists should remain in the Apple fold, especially if Apple is successful in keeping OS X from running on non-Apple hardware, and there seems to be a "switcher" trend underway, perhaps from the "iPod halo effect". The new potential customers are those who are tied to Windows, but for whom the possibility of having enough access to Windows will be sufficient to get them to consider Apple hardware.

This could be huge for Apple. Let's stipulate that the sales rate for the desktop/laptop market is 95% Windows and 3% Mac. If only 1% of those Windows users buy Apple hardware, that's roughly a 30% increase in the sales rate for Apple hardware. Apple is already healthily profitable at current sales levels; any additional growth is financial gravy.

I'm sure Apple would prefer that people run OS X, but it can't hurt to get some revenue from the Windows base, and a Mac hardware/Windows software user is one (big) step closer to becoming an OS X user.

Apple and Microsoft

Lost in the noise here is the announcement of a five-year "agreement" between Apple and Microsoft wherein Microsoft has committed to continue developing Office for the Mac platform for five more years. This puts to rest the notion that Apple's move into the Intel space would alienate Microsoft and threaten Office support. This little bit of uncertainty about the Mac platform is put to rest for a while.

The unanswered (and so far unasked, as far as I can tell) is: what did Apple agree to in exchange for Microsoft's commitment? There are all sorts of possibilities, like a side deal over music and/or DRM, but I'll keep my guesses closer to home. Maybe an agreement not to develop a competitor for Office (the rumored "Numbers" application for iWork)? Or maybe to stay away from actively supporting OpenOffice? I don't know, but there has to be something, and I'd love to know what Microsoft thought they needed (or wanted) from Apple...