I think it's reasonably accurate to say that we're entering a post-PC world, but I also think arguing about labels misses a couple of key points.

One point Harry touched on is that phones and tablets are "intensely personal." People carry them everywhere and develop almost emotional attachments to them. They become personal in a way that stationary PCs don't.

But arguing over device labels also misses a dramatic power shift that's been happening slowly. People have always bought "home PCs" considering interoperability with work or school systems. The enterprise IT department set up servers and gave you a machine that would work with them when you arrived for your first day on the job. If you wanted your own PC at home to work with enterprise servers, you'd better start with a similar setup.

The power shift is that employees are now arriving for the first day of work with their PCs (smartphones, tablets, laptops) expecting to use them with enterprise servers. I've seen brand new company-issued PCs sitting in boxes when savvy new employees bring their own PCs.

There have always been separate home and enterprise PC markets and this separation will continue. In the past, the policies of the enterprise often drove buying decisions at home. The power shift is that home preferences are starting to drive enterprise IT policies.

I've written a couple of posts with more details on my blog: http://bobvan.tumblr.com/

A great way to visualize the data. I can't think of a better way to rate company performance. I think Edward Tufte would approve of this visualization.

"He [Steve Jobs] has to compete if a free market, thus his customers will primarily be the insecure and zealots (and those that just want stuff to work well) that he has today."

I might be ashamed of my insecurities. I could be ashamed if I let zealotry get ahead of rational thought. But who would feel shame in just wanting stuff to work well? What's wrong with that?

I've soldered CPUs in the early hobbyist days. I've written my own compilers, operating systems, GUIs, and applications. Just because I could build and maintain my own computing infrastructure doesn't mean that I want to spend my energies doing that. Been there; done that. I'd rather have a stable and reliable platform on which to build things that haven't already been built.

I have nothing against people who want to tinker. I'm a strong supporter of open source development models and have contributed to a couple of projects (precisely because I wanted them to just work well too). There's no shame in wanting things to just work well.

I see people relating to their mobile devices in ways they wouldn't relate to a desktop or even a laptop. They have an emotional connection of sorts with the device. They carry it everywhere. They depend on it for reliable communication with other people. This is why dropped calls and data performance are such a constant battle among cellular providers.

I think this shift from a detached relationship with a desktop PC to a much closer relationship with mobile devices is driving the demand for mobile devices that just work.

I think today's iOS devices score better on "just works well" than Android. My crystal ball says they'll maintain that edge for structural reasons that are hard for Android to change.

I think there's a big difference between zero-sum games and winner-take-all markets.

As long as most people buy only one smartphone, the smartphone market is a zero-sum game. Market size may increase as feature phone users upgrade to smartphones and this can increase unit sales over time. But as long as each consumer buys only one smartphone, then market share is a zero-sum game.

The PC OS market was winner-takes-all because standards did not exist to allow competition and innovation. Also, the PC market was very small in the early days when competitors stood more of a chance. Cell phones now out ship PCs 3x or 4x to one. The standards needed for cell phone competition have been in place from the start, wisely mandated by the governments that licensed the scarce spectrum.

So this is one of those rare times when I have to disagree with Harry. I don't see this as the end of the zero-sum game. The cell phone and smartphone markets have never been winner-take-all markets and they never will be. But they have been and will continue to be zero-sum games among the vendors.

I think it just looks like the end of the zero-sum game because of the decreasing importance of PCs and the increasing importance of mobile devices like smartphones and iPads. This thankfully represents a shift from a winner-take-all PC market to a mobile device market with healthy, ongoing competition and innovation.

I think "resolution" is the key word to consider here. The user interfaces that work best for me take into account screen resolution, but perhaps more importantly, pointing resolution. It's easy to take mice for granted, but it's pretty amazing that I can point to any one of the three million pixels on my desktop in less than a second. You can achieve the same resolution with just finger pointing if you zoom and pan a lot, but it takes much longer.

I haven't been a fan of desktop interfaces scaled down for mobile use. I suspect that at least part of the reason is that design assumptions about screen and pointing resolution are hard to change.

Apple has the Cocoa Touch interface for finger resolution and the Cocoa interface for mouse resolution. Cocoa Touch buttons are sized for finger pointing while Cocoa buttons are sized for mouse pointing. If you've ever edited text on an iPhone, then you know how Cocoa Touch uses a magnifying glass metaphor to get local mouse-like pointing resolution from a finger, but you also know how long it takes and how frustrating it is if used too much.

I suppose we should also consider the FrontRow user interface for media playback that's already part of Mac OS X. You could argue that it's an interface with no pointing device at all. It's based on a remote with six buttons. Four of those buttons give directional control over an on-screen cursor that shows the focus of the action.

Hardware for an iMac with touch would be the easy part. Apple has already shown with the iOS versions of iMovie, Keynote, Pages, and Numbers that common applications can get a lot done with a touch UI, but all of these leave behind some desktop functionality. Also, these are separate products from their desktop brethren. I'd guess that Apple will hold off on an iMac with touch until they have a way to deliver them in some sort of merged form that moves smoothly between mouse-based and touch-based UIs as the screen is tilted. That's certainly possible, but I think it's the harder part of the job.

So now I can finally get to Harry's question about how much Mac OS X has to change. With FrontRow, Apple showed that they already know how to add task-specific interfaces when needed. But FrontRow is what I'd call "application modal" meaning that you only get the FrontRow interface when the FrontRow app is running.

I can imagine an iMac that shifted to a different, touch-based interface when tilted, but what apps would run in tilted mode? Would you expect a tilted iMac to continue running the same Mac OS X applications from mouse mode or would you expect it to be application modal where a whole new set of applications became available in tilted mode? And would you want a curated app store for desktop applications? I can see pros and cons.

So far, Apple has been building two pretty distinct worlds along many dimensions. There's the mobile world with low-res touch interfaces, a curated app store, and very light system administration burden (e.g. backups automatically done for you through iTunes, no file system access, etc.). This is commonly said to be targeted at content consumption. Then there's the desktop/laptop world with high-res pointing interfaces, an open application platform, and more system administration work (e.g. you do your own backups, manage your own file system, etc.). This is commonly said to be targeted at content creation.

The recent trend has been to see more and more ways to sync content across these distinct worlds. Dropbox, GoodReader, Instapaper, and Apple's iDisk are just a few examples. The need for sync between worlds shows the need for content to move across worlds, but does that mean the worlds will ever merge? Sync also goes to the cloud, so where does Apple's cloud strategy fit into all of this?

Steve Jobs recently pointed out that trucks were the first widely used vehicles while cars came later. The arrival of cars didn't make trucks go away, they just made personal transportation more convenient. He likened the desktop/laptop computers of today to trucks while he likened iOS devices to cars.

So is an iMac with a touch interface an SUV? A minivan? I don't know, but my crystal ball doesn't show Apple rushing to build one soon because the market for iOS devices is still far from mature. If/when they do build it, I think they'll have thought though through not just the OS, but the user interface needed to build unified applications that work with both high-res and low-res pointing devices.

Bob

I would try leveraging the inverse square law (http://en.wikipedia.org/wiki/Inverse-square_law). Put an access point IN the podium. The presenter with a handheld device can look natural and never need an RF path of more than a meter or so. Competing signals from the audience will be at least 10s of meters away and so attenuated 100x or more relative to the presenter. I don't know if the spread spectrum modulation techniques of WiFi exhibit the capture effect of FM, but it's worth a shot. Worst case, you just have to build a transparent Farady cage for the presenter and the access point :-)