Edited on 10 Jul 2013: Added a note about text editors at the end.
Last week, Apple held its developer conference, WWDC13, at which they previewed the next generation of their mobile operating system for iPhone, iPad and the presumed new product. iOS7.
It has a few superficial characteristics
- It loses what's been called the skeuomorphisms. Among these, a button no longer looks or visually behaves like a button.
- It establishes a new design grammar with muted colors, new screen layout and no bounding boxes, a design grammar that is consistent and deep.
The combination of these two has been called ‘flat’ and clean and is what all the immediate, mostly negative, comments are about. This is surely a big deal. The notion of deep consistency is more important than the style itself. I like these changes — including the style — but they aren't the most important thing that has happened.
- There are some new control centers, for instance the modern multitasking and notification services. These too are welcome, and will make my life better.
But there is some deeper magic going on that really has me excited. First some background.
The Background for this Move
What is now the iPad was being considered by Steve Jobs at NeXT in the early 90's as a touchscreen device driving a next generation computer. There were some discussions about us needing such a thing; I assume that later Jobs had such a tablet in mind for the Apple resurgence planning. The experience in miniaturization gained on iPod added to the market potential in phones plus manufacturing challenges for larger devices meant iPhone came first but the iPad was already mature before iPhone.
We really can credit Jobs with the vision and balls to make this happen, and some of his guys, like Mansfield, Forstall and Kiggins. He was quick to buy key bits, like FingerWorks, and to take advantage of opportunities to widen the ecosystem. I greatly admire the grace he brought to the business.
But Jobs had limits, particularly in how he saw software. He never escaped the notion that software controls need to mimic reality. This probably was needed when the mobile devices were new and the interaction model unfamiliar. But we now have half a billion iOS devices and nearly as many from competitors using essentially the same paradigm. People know how to use these things.
That's roughly a billion devices in five years, all designed the same. People know how to use them. We don't need to be pandered to any more.
What we need next is what made the Mac great, a user interaction model for the future, say ten years. Not one that relies on buttons and panels, but one with a new physics.
As it happens, we are designing user interfaces here for a system that evolved from the earlier discussions. We're frustrated by the limits of Apple and Google devices. We've spent time with the true open source options, Sailfish, Ubuntu and Tizen. They disappoint when held against our needs. The problem is that you really do have to integrate hardware with the software controls, and you have to use low level APIs to innovate in the UI. Only Apple can do this.
What has happened at Apple couldn't have been predicted. While goofy pundits worry about whether Apple can survive after Jobs, some of us knew that Jobs and his protege was the problem now. The radical thing that has happened is that Tim Cook inherited a divided company. He did the unthinkable: booting out the Jobs guy. This is big. He did this only seven months ago and chartered something new, something not on Jobs' master plan. He's betting the future of his leading product on this. There is no question in my mind that this is the guy to lead Apple.
What's Really New
We have a completely reimagined iOS. I am supposing that some of what I mention below enables user interfaces for the unannounced wearable device. But so far as I am concerned, it makes my user interface challenges tenable for deep collaborative work.
The general problem we face is that we want to model situations and the world is set up to model facts. Our UI application at the moment is annotating movies. It is quite easy to mark and annotate objects and people. But what we want to know are things like the way shots are blocked, how the camera moves, what editing rhythms have been applied, how the space constrains and flows. We want to model situational dynamics.
The elevator pitch is that we want to model environments not objects. Sometimes I say it is architecture, not sculpture. The move from object to environment is just what the Apple team is attempting. Four things matter:
- Before, we had a system of panels that contained either buttons (and other physically inspired controls) or information in frames. Often, the information frames were objects, like the balloons in Messages, or bounded rectangles on a physically-inspired panel. Often this panel had physical texture. The new navigation paradigm is roughly the same but now the designer has to rely on the situational context in the application rather than some physically hinted context. It doesn't take much imagination to see that this will be more work for application designers but will open new possibilities for us all.
- We have lost the objects, but we now have something much richer, a spatial paradigm. We now have a system-wide notion of layers. All animations, transparencies and views now have a very clear spatial vocabulary. Now we can start planning a party instead of just arranging crackers on trays.
- Also, that system of spatial definitions has its own physics engine! The effects we have now are sparse. (Perspectives is one of them.) Those that we have give us some notion of where they are heading. All are spatially inspired and loosely based on real world physics, but with enough difference for us to know we are in a different world when we enter our device. I expect this physics vocabulary to grow slowly and to be advanced by explorers like us, but the axioms are in place to build a whole parallel world.
- Finally, we have the foundations for inter-application communication. We won't see much on this front for a while. It is a very big deal and the paradigm needs to be discovered. We will be moving from a collection of apps (think objects) to working environments. OSX and the Mac OS before has rudimentary interapplication scripting but it works poorly. They tried OpenDoc long ago, but the architecture was wrong. They'll get this right over time, I believe. I also believe that they will leverage the same reactive paradigm of spatial layers they've introduced here, and that are implicit in Quartz Composer. The design of the new MacPro tells us where the values are.
This is where we will be truly thankful for the curation and constraints applied through the App Store.
Think about how radical this is. I thought it was radical to make Unix a consumer OS. It was radical to think of pocket computers. That's nothing compared to this. The App Store has nearly a million applications, and some are very good. It is the richest programming ecosphere on the planet; nothing else comes close. There are 50 billion copies of those applications in the user base.
Apple just broke every one of those 50 billion copies. Man am I proud of these guys in whom I've placed my trust. Boy am I excited about the future.
And no, Vesper is not a model of this future.
Wait until you see what we are making!
That machine was also unveiled at WWDC, a surprise. The former Mac Pro was a machine like all big PeeCees: a box with powerful CPU, and lots of places for extra RAM, hard drives and expansion cards, including graphics cards. That machine appealed to both gamers and graphics/video professionals. This new machine is radically different. It has no internal expansion; that is to be handled by external devices connected by the amazing (to me amazing) Thunderbolt. You can't upgrade the processor, nor the solid state memory. Nor the graphics cards. No one knows the machine's cost, but many are up in arms because... well because it is not ordinary.
The relevant part of the design is that though it can come with some kickass Xeon CPUs, the real power is in two bigtime GPUs, only one of which is there to drive monitors (three 4k monitors!). As it happens, a great deal of the work that professionals do is friendly to GPU power rather than CPU. Some years ago, Apple developed OpenCL, introduced in Snow Leopard. This is a standard framework for writing programs so that the load can be dynamically shifted among CPUs and GPUs. You can get profound speed improvements if things are factored right. I'm talking several orders of magnitude. In theory, this box — for the kinds of things we are designing now — could be hundreds of times faster than the old Mac Pro.
Implications for Favored Apps
I supposed that I'll have to revisit every decision a few months after iOS7 arrives, but my iPhone broke (baby, slobber) and I had to replace it with an iPhone 5. Nearly every commonly used app has adjusted to the taller screen, except the one I am the most picky about, the text editor.
My choice was Notely, and I wrote scores of longish documents in it. I found myself mailing these rather than do the dropbox transfer because it was simply easier. So I did not notice that Notely had not updated its interface to Dropbox. It not only has not been updated for the iPhone 5, it has been removed from the store! Emails to the developer go unanswered.
Shucks. My quest has to start all over again. I think I will stick with Drafts for now because I rarely finish a doc on the phone. I expect that we'll have to count on the successful apps to modernize in he next year, so I don't want to settle too early. My requirements are simple, at least they seem so:
- a full screen view with the status bar gone. I'd rather the note title did not display.
- the ability to set the background (including the background of the keyboard) black and the text at Helvetica Neue 13 bright orange.
- a tailorable, small extra row.
- support for TextExpander, Dropbox, mail and possibly Evernote.
- vital upgrades to modern notions.
No one app has these. Unfortunately, the text editing app business is highly fragmented and probably not profitable.
In this note, I worry about drawing apps. No luck there as well in terms of satisfaction, and the investment so far has been significant. I'm using Zen Brush for now, in concert with numerous desktop postprocessing applications.