Cupertino Kremlinology

In preparation to try my hand at Apple prognostication, the standard disclaimers apply: I have no inside information, I am engaging in idle speculation, and I am fully prepared to eat my words. Okay. Let’s get started.

I’m not interested in what unreleased hardware will look like, since it’s safe to say anything designed by Apple will be simple and breathtakingly beautiful (as usual), and will solve the associated design challenges with kick-you-in-the-pants obviousness.¹ The interesting question I keep coming back to is whether a tablet device will run the iPhone OS, and by extension, run iPhone apps.

First of all, since the rumored tablet would presumably have more in common with an iPhone than your mom’s MacBook—multitouch screen, accelerometer—it’s safe to say it would not be running Mac OS X. Apple has already created a mobile version of OS X—the iPhone OS. Furthermore, it’s clear that Apple has designed Cocoa Touch from the beginning to have enough headroom to grow into a device with a larger screen. (Which might explain why the SDK took so long.) Evidence of this includes the UIWindow class, which exists even though iPhone applications display only one window. I bet Cocoa Touch on the tablet will allow multi-windowed applications.

I would also wager that the device would be able to run existing iPhone apps without any trouble. I’m almost positive, however, that they won’t run full screen. App developers have spent too much time optimizing their UIs so widgets are relatively finger-sized. The interfaces simply won’t scale without usability problems.

Given that size is less of a constraint in a tablet device, I expect the hardware to be powerful enough to support multitasking. This, combined with the presumption that the tablet would run iPhone apps at their native size presents an interesting UI possibility. I envision a dashboard-like layer where a limited number iPhone apps run as widgets on the tablet.

Picture this: iPhone apps launch on a translucent layer where a pair of widgets run side-by-side on the screen. A simple swipe will switch between pages of running apps that don’t fit on the first screen—just like having multiple pages open in Mobile Safari. Capping the number of simultaneously running apps is as simple as limiting the number of pages in the widget layer.

Apple would be crazy to launch a device without the momentum of the App Store behind it, and the only way hit the ground running would be to let existing iPhone apps run on a tablet. In the end though, Apple has spent far more time thinking about these problems and refining the solutions than all of the armchair product-designers and rumormongers combined. I may be wildly off predicting the app switcher, but it’s a fairly safe bet that whatever we see next Wednesday will blow us away.

There’s a Price for That

Wolfram Alpha, the hilariously useless search tool, has an API. What sets it apart from Google, aside from choking on input every once in a while, is that Alpha’s API costs money.

You can get your first 1,000 requests for as little¹ as $60 a month! It’s only eight cents per additional request beyond that! But wait! There’s more!

If you’re a “student, hobbyist, or individual developer,” 25,000 requests will set you back $1,200. (Seriously folks, that’s $0.05 for every search.) If you’re a small or medium business, though, you’d better be ready to part with two grand for the same level of access ($0.08 per request). Never has market segmentation been so blatant. They also don’t tell you how to get your nickel back if the query was a dud.²

Somehow I don’t think an API that has more in common with a casino slot machine than a scientific research tool is going to take off.

The pricing model I’m excited about is 280 North’s Atlas beta program. Here’s a slick web app for building slick web apps, and they’re not afraid to charge for a beta. No one has the balls to charge for beta these days, but these guys decided that $20 is a perfectly reasonable price to pay to test drive an online Interface Builder. And it is resasonable—reasonable enough that I wish I had a good reason to try it out.

I’m glad people are starting to move away from the broken economies of the ad-supported web. Both companies understand that in more and more cases, successful web ventures need to charge in order to be profitable. The trick, however, is to have something people are willing to pay for.

Half a Be

or, the sad truth of the eternal present tense

Now that we’ve all either upgraded to Snowy Leopard or are patiently awaiting Windows 7, let’s extend a moment of silence for the operating systems that didn’t make it.

Inciting text for hopeful developers can still be found in the Amazon.com review of Programming the Be Operating System. While the review carefully—and, in retrospect, with good reason—avoids any claims of longevity, it makes no frivolous assertion of the powerful features of Be’s erstwhile chef-d’œuvre:

There’s not yet a big market for creators of programs that run under the Be operating system¹, but its special capabilities may prove irresistible to experiment-minded programmers.

In 2001, Be, Inc. dissolved—it ceased to Be—but Amazon’s rosy marketing prose continues to remind us of good things long gone. Indeed, certain parts of the OS still have no equivalent in today’s systems. I studied their file system in great detail because it was so far ahead of its time. But the Internet does strange things to radical ideas. Some creations are buried and long forgotten because they are too novel, only to resurface when we’ve grown the capacity to comprehend their grandeur.