Skip to content

It’s the Network, Not the Application: Online Desktops

Though travel is somewhere around last on a list of my favorite things about my job, there were two separate events this week that I’m disappointed I couldn’t fit into the schedule.

The first, and most obvious, is David Berlind and Doug Gold’s Mashup Camp. As I’ve written and commented on before, this has fast become my favorite conference to attend, period. It’s not just the fact that I can get away with a Sox hat and jeans there (though that doesn’t hurt). It’s not even the overabundance of creativity and skill. It is, rather, the relevance, the immediacy of the applications on display.

Over the course of my career, I’ve more or less come to terms with the fact that the overwhelming majority of technologies I cover are irrelevant – unknown, even – to the non-technical members of my family. Now that they’ve sold off the Thinkpad line, I seriously doubt that any of my non-technical friends could between them name a single IBM product. Whether or not that’s good or bad is an exercise I’ll leave to the reader (for now).

The fact that regular people don’t care about what we in the industry might consider to be regular technologies isn’t, in itself interesting. Or news. What is interesting, and may be news, is how exceptional Mashup Camp is in this regard.

Yes, some of the demos at Mashup Camp are toys or experiments or, in some cases, jokes. But a number of them are not only of serious intent but fundamentally and surprisingly useful – and potentially transformative. Think Adrian’s, Anthony’s Hype Machine, Paul’s Not billion dollar businesses, or even million dollar businesses, but useful, relevant, and – perhaps most importantly – accessible. Not because of the technologies, which are often very basic, but the unique ways they make the information on the network useful to us.

Which brings us to missed conference #2, GUADEC. While I’ll admit that part of my disappointment stems from the fact that a.) I’m a GNOME user and b.) because I know and would be happy to grab beers with a bunch of the GNOME folks, my sense of loss is especially keen because of this year’s seemingly de facto theme (from what I gather on #redmonk, anyway): online desktops. Desktops, in other words, that heavily leverage some of the same relevant, easily accessible network applications built by developers at Mashup Camp.

Those of you who are avid readers of this space (my apologies) may recall that I have a longstanding interest in the subject, if by longstanding interest you mean massive, overwhelming passion for the subject.


It’s simple, and Adam Bosworth told us this years ago: the value is not in the application, it’s in the network. Putting it in practical terms, Havoc says: “without an Internet connection I don’t even bother to turn on my laptop.” Amen, sir, amen.

My own laptop, shiny as it may be, is more or less a brick without connectivity. I could write, I suppose, but I don’t have access to my sources and I’d have to go back and insert links. Music’s an option, but I’d much rather listen to new stuff on WOXY than what I have. Even gamers are loathe to disconnect, with multi-player games the rule now rather than the exception.

It’s the network’s world, and we’re all living in it.

Or at least it should be. The operating systems most of us use today, however, haven’t quite adjusted to that reality – open source or proprietary. We still work in computing environments that the treat the network as an afterthought rather than application lifeblood.

Much has been made – and rightfully so – of the way that Microsoft had a corporate epiphany, and “got” the internet overnight. Apple, meanwhile, has practically been canonized for its hardware and software combination aimed at network enabled music services. But are the dominant desktops really leveraging the network, or just bolting it on tactically?

A year ago, as an example, I asked the simple question: why can’t Flickr be my wallpaper backend? Calvin seemed to think that was a good idea, so he went ahead and hacked it up, and someone’s apparently had the same idea over on OS X.

But go further. Could Facebook reach into and replace my local identity store? Could tags be applied to local filesystem assets? Or how about Google Apps blending seamlessly into the local client, as in this GNOME SOC project? And so on.

The difficulty here, of course, is that this advantages the network application providers at the expense of the operating system providers, which might be a good thing for users but is decidedly less so if you make money selling operating systems – as do both Apple and Microsoft. They are financially committed – chained, even – to the notion that it’s the application that enables the network, not the reverse.

Which is why I think the disruptive opportunity here is in fact for open source competitors, like the folks from GNOME. As history has demonstrated quite conclusively at this point, you cannot out-Microsoft Microsoft. As Google has demonstrated quite conclusively, however, there are tremendous disruptive opportunities still on the table. But you don’t catch up, as Bart reminds us, by going slower than they are.

Here’s how I put it more diplomatically last August:

What I’m more curious to see is what direction the desktops take in delivering a next generation desktop experience. Will they try to out-Windows Windows, or take an entirely different tack? I’m hoping, obviously, for the latter. More to the point, I think desktop Linux has an advantage in satisfying next generation users that its community has not fully perceived yet.

If one accepts that a great deal of application development and delivery will shift to the network – and there is ample evidence to support this from Google to Salesforce – it’s logical to assume that the operating system that best leverages these network applications will be well positioned for success over the longer term.

The best returns for a desktop provider, then, may be for someone who concedes that the desktop is no longer the end all and be all – the network is what’s important now. When I’m somewhere without network connectivity, I barely even think of my machine unless I’m writing. I think that the ultimate opportunity for Linux on the desktop is to forget about the desktop as it exists today. Microsoft can’t because that’s how it makes money, Apple can’t because that’s how they sell hardware. But Linux? Linux has no such shackles, no such Innovator’s Dilemma to cope with. Either way, it’ll be interesting to see how all of this plays out.

Are there questions and uncertainties about the prospect of an online desktop? You bet. Freedom. Privacy. Security. Standards. Technology.

But it would be a shame if these throttled development. Browsers were useful before the invention of SSL, and email (I’m told) is still a moderately useful tool for the bulk of the online population despite the fact that virtually no one employs either PGP or VPNs. Minimum progress to delcare victory, and all of that.

If it is true that 50% of the audience hated the idea, I hope they’ll at least keep an open mind, because in the year since I wrote the above the opportunity has done nothing but expand dramatically. Or did you think the the massive, exponential growth of Facebook and Twitter and so on was a coincidence?

It’s inevitable, I think, that someone is going to get this right. The questions to be answered, then, are who, and how soon? The timing will remain a question for some time, I think, but the who is really just a matter of will.

Categories: Network, Open Source, Software-as-a-Service.

  • Jon Collins

    Humm, harr, yes I sort of agree-ish with the online desktop thang. I agree more with the “network not the application” – its the one thing Sun Microsystems has got consistently right through the years. But aren’t these two differnt things?

    Put it this way – from a resource standpoint, its back to the magic triangle of communications, processing and storage (hat tip to Von Neumann). While bandwidth is not infinite, there will always be a requirement for offline or throttled online access, which means that online desktops can only exist for a permanently connected minority. (And while I know these are in the majority when it comes to computers, I’m thinking just as much about the evolution of mobile devices). There may be workarounds such as caching, data synchronisation etc but these are imperfect answers.

    Meanwhile, we have the “network not the application”. Again this is back to the magic triangle, but in a different way – we have standalone apps, OS etc because historically, communications were not available, and these provide the legacy environment that may be functionally rich but is clearly cumbersome.

    Just imagine what things might have been like if the network had come first, and computers followed? For duffers like me, we don’t have to imagine as this is exactly the evolution we’ve seen since the arrival of the code-enabled Web browser. Applications have been slow to come, but they are coming and in a very different way to the big-block, computer centric alternatives. The adoption of AJAX was probably the key moment, though others will be able to pinpoint it better than I can. Sure, the apps of today are clunky from a usability perspective but (to your point about Facebook and Twitter) they are clearly delivering. Meanwhile, we now have component based application frameworks and scripting based interfaces, which is exactly the prediction made for the future of applications in general.

    So, yeah, absolutely to your main premise. Do I need an online desktop to take advantage of it all? Dunno, but there’s a lot of innovation to happen yet so I’ll keep out of the jury for now.

  • Pingback: Irregular Verbiage » Blog Archive » GUADEC Day 5

  • Dan Davies Brackett

    I might not turn on my laptop much without a ‘net connection, but I also find ‘net connections *totally valueless* unless I have my laptop around to take advantage of them. I can’t listen to my favourite internet radio stations without Winamp or at least a browser; I can’t while away the hours on Metafilter or Slashdot or SomethingAwful without a browser and, preferably, an IM client.

    The local software, the local hardware, are as important to my online experience as the network is.

  • James

    Dan: Local software is part of the Online Desktop vision. As for hardware, see the Nokia n800 (which just got a Mozilla-based browser), the OpenMoko Neo1973 phone, and GMAE in general.

  • Pingback: theSalmonFarm Blog » Blog Archive » The “World Wide Web” isn’t.

  • Pingback: tecosystems » Some Work and Much Play @ UbuntuLive / OSCON

  • Pingback: Never Tell… Richard Eibrand » Blog Archive » new online station -

  • Pingback: tecosystems » Cool Gigs: EveryBlock and LiTL

  • Pingback: tecosystems » More Ozzie, Less Ballmer - Please