Benjamin Button operating system

Must have been very disappointing. ‘The Curious Case of Benjamin Button’ got a record number of nominations for the Oscars, but only won three minor awards. It won’t provide any solace, but I certainly would have granted the movie my personal award for ‘Best Metaphor for IT-related Subjects’. When I watch Brad Pritt mature from a baby in a old-mans body to a man in the strength of his life, then finally end as an old man in a baby’s body, I can’t help but relating it to – well - operating systems.

A sad example of an occupational disability indeed.

But think about it: most operating systems by now are complete ready for some life cycle reversal. And I am not just talking about Microsoft Windows. I recently bought the 2009 editions of Apple’s iLife and iWork software suites. What a bulky bunch of fat client applications that really is. Spontaneously, it starts to feel like an anachronism. As I wait for the huge software stack to install itself from DVD, I realise that I don’t even know what new features are contained in this edition. What can I say? I am one of these unquestioning Apple junkies. Eventually, I find an interesting functionality in iPhoto: self-calibrating face recognition software scans your photo library, identifies people and tags them in handy categories. With the size of my library, the initial scan takes more than 5 hours. It fully absorbs my poor Dual Core Apple Mini and I think about how an optimised, Googlified server farm probably could have done the job in seconds. No doubt it also would do much better in storing the photo’s. And in processing the movies I am editing. And in managing the music that I buy from iTunes. So what will be the point of the 2010 editions – and beyond – of this type of applications and the operating systems that run them?

If infrastructure, processing and storage moves to the cloud, we need lighter client operating systems. They need to be optimised to communicate with whatever is inside the cloud and deal effectively with user interaction. That’s about it. Quite a different mission statement. Operating systems will grow smaller and as they mature, they will start to look younger.

The signs are everywhere. The new Snow Leopard release of OS X will be lighter and faster, rather than rich on additional features. Linux suppliers are slimming down their distributions, also to make them more suitable to the new generation of cloud-enabled Netbooks. Also, there are some new players knocking on the door. Palm’s much anticipated WebOS is really a denial of an operating system, purely designed to connect to the Internet and then let the exciting things happen elsewhere (which actually is a move back to the no-nonsense roots of the company). And we can only speculate what Google’s Gdrive might supply through its expected support for booting from the network: could it be a small, lean and mean operating system (‘G’ would be a nice name) that only contains the features needed to run the Chrome browser?

Realising this, it seems like a missed opportunity for Microsoft to position Windows 7 as just a new version of Vista with some additional functionality. Especially if we know that it will feature a powerful virtual machine that can run older versions of the operating system. With some reckless courage, Microsoft could have created a small, cloud client OS – just imagine, ostentatiously developed in open source - that would include a virtual window to run any legacy Windows application (in the cloud, obviously). It has been done before: it proved to be the only way for Steve Jobs to migrate from legacy OS 9 to the dramatically different, Unix-based OS X.

Oh well, just another scenario that did not make it into the nominations.

In any case, the client operating systems of the near future will look lighter, younger and in general very different from how they look now. And finally, on one day, they will just seem to have gone, leaving us wondering what we were bothered with in the first place. A curious reverse life cycle, that I will gladly watch.

Just sit back and let the movie play.

First published on Capgemini's CTO blog