Continuum computing: I want the right thing, everywhere

And it seems everyone wants to deliver that.

I'm going to be demanding. Software - or rather tools and experiences - have to work wherever I am, on whatever device I'm using, with the appropriate interface and tools and context for whatever device and input I'm using. It just makes sense. If I put some photos onto our network drive from my PC (it has the SD card slot), I'll tag and rotate them in Windows Live Photo Gallery and maybe do some auto fixes. But when it comes to uploading them to Flickr, I'd quite like to do that from the Media Center in the lounge, or from my phone or maybe from my digital photo frame. If I'm looking for a cheap flight for CES, I should be able to set up the query on my phone and see the different options on the bigger screen on my PC. My Flowerz achievements should show up on my phone, my PC, my Xbox and my Android tablet. I shouldn't have to resort to Web apps to get a cross-device experience, and I shouldn't have to deal with fiddly buttons from a remote control or have to hunt out different apps for each device - I should be able to pick an app and use it several places.

I can consume my music and videos and photos on pretty much any device and you can see my update from one social network on another and again, on most devices. But if I edit a photo on photoshop.com, can I undo that edit in the desktop version of Photoshop?

The idea of things you do over a period of time, on different devices, at your own convenience is something we've been thinking about since the Web got useful in the 90s. Ubiquitous and context-sensitive computing has been a research area for decades; check out the Xerox PARC ideas about pads, tabs and walls for a sense that the technology is finally catching up with some of the vision - but having the hardware isn't enough, until we have the processes and tools and apps and services to deliver it as well.

We recently spent a month at different conferences and promisingly, that was one of the themes of just about all of them. Nvidia's GTC conference was about horsepower for making things happen in the cloud or (eventually) on the phone and tablet. CTIA Enterprise was a great place to see the latest devices and decide it's going to be 6-9 months before anything except the iPad is really ready for prime time (something that the Nvidia Tegra team confirmed - they expect tablets with Tegra 3 and equivalent graphics processors from the summer on). BlackBerry Devcon put context and working cross multiple devices front and centre, with the PlayBook - a tablet that can be a bigger BlackBerry or a smaller 'PC' depending on whether you run a BlackBerry app or an AIR app or an OpenGL game (which should have the context services RIM talked about the year before, that know where you are and when you're going to arrive - and if you're likely to be late). And apart from the hardware news, Intel's developer conference and the AppUp Elements developer (un)conference both had a couple of big themes: contextual experiences like a travel guide that suggests the kind of outings it knows you like, and the continuum of experiences that go across devices.

If you've ever wondered why Intel is interested in software, there are three good reasons. If you want people to get the most out of your hardware, it's a good idea to write compilers that take full advantage of them so the companies whose day job is writing developer tools know what's possible. If you want to sell advanced hardware capabilities, it's easier to make money by selling the software that needs them (which is why Intel bought McAfee). And then there's the big one; the hardware manufacturers pick Intel chips because people want devices that run the software that runs on them - PCs for Windows and Windows apps, Macs for Mac OS X and the Mac experience (which I think far outweighs any individual Mac app) and, if Intel gets it right, one day Atom phones and TVs and tablets. But I don't want a different set of software on every device I touch and developers don't want to keep rewriting apps again and again for new devices. But then again, I don't want least-common-denominator apps that don't take advantage of the features of my device; whether that's touch, or voice, or GPS, or the power of Windows and a Core i7. I don't want powerful keyboard shortcuts or power-hungry animated transitions on a phone, I don't need big, touch-friendly buttons on a desktop with no touch screen and I don't want the basic features of a phone app on a tablet that has a screen big enough to let me cope with doing more.

Intel's Renee James brought up the idea of a continuum - apps, tools, experiences, content that's available across devices. And the developers said if there were open standards and common APIs across devices that made it easy to build apps that automatically scaled themselves to suit the device, they'd switch away from their existing developer tools - although there's a clear division between what I think of as the write once, run anywhere believers and the write once, port anywhere realists. (You might be able to tell where my own view lies!)

Microsoft's Brian Goldfarb used a similar phrase when we were talking about Silverlight at the MIX conference this spring; write once, optimise anywhere. That's the continuum approach Microsoft is taking with Windows Phone 7 and Symbian and Xbox and the PC and the Web (covering Mac and, through Mono, Linux too); write to Silverlight, write to XNA and you can make the apps work on each platform far more quickly, but you can't expect the interface that's right for a game controller to suite a phone screen.

I don't think phone OS's stretch up to tablets; Windows Mobile on the 5" HTC Advantage was frustrating, I think RIM was absolutely right to build the PlayBook on something like QNX and if I was carrying something the size of the iPad or the Galaxy Tab I'd want it to have full Windows apps, or at least a full desktop browser (and handwriting recognition - the Windows technology I've been relying on since about 2001). And iPad apps aren't just the same iPhone apps anyway. Windows Phone 7 won't go onto tablets because Microsoft knows that it would be a frustrating experience; because you can see more at once on a larger screen you can cope with more complicated concepts and you want to do more complex things. The Windows Phone team will stick to their knitting and leave the Windows team to work on scaling the Windows 8 technology and interface down to the 7" screen (maybe they could take a bit more advantage of the Media Center interface that's nicely finger sized?)

Put it all together and my suspicion is that neither multiple platforms nor multiple ways of developing apps to run on different devices are going away. But there's an increasing awareness that no device is an island, and instead of just data passing between them there needs to be this continuum of what you can do and where. So next time you're wondering what Intel AppUp is for, apart from waving at Microsoft as a threat and being prepared for when Microsoft does its own app store, remember that leading developers across the continuum is one of the answers. M