Social network virtual machine

Social networks have become topic de jour over the last year, and with good reason. According to Alexa, Facebook now has more pageviews than Google. Pageviews of course translate into profits at some stage, and so the market for social networks like Facebook has become extremely hot over the last while, with Microsoft winning a bidding war against Google and other players to invest $240 mil in a minority stake valuing Facebook at $15 billion. That’s a big number, and has some people worried about another dotcom overvaluation meltdown.

Financials aside though, Facebook taking over the global social network has worried me for other reasons. Firstly, there is nothing that Mr Zuckerberg et al have done that is really revolutionary – as this Economist article points out, Facebook merely took the idea of social graphing and implemented it more effectively than the likes of Plaxo and other online address books. There is nothing here that cannot be copied and improved upon though. Perhaps their breakthough was in opening up their data source to external developers via an API, creating the new class of "Facebook apps" – certainly many companies have already spent countless dollars trying to infiltrate the social network.

As a developer though, this brings me to my second concern – Facebook is by no means unique. MySpace, Orkut, Salesforce, LinkedIn, Ning, Hi5 and Plaxo to name but a few all add to the global social graph, connecting people with proprietary datasources. Are we to have to choose the "flavour of the month" tool, or port our social network apps to every platform? A problem indeed, and where there is a problem, there is money to be made.

What is needed is a solution like that which Java brought to programming languages – given a variety of platforms, create an abstract layer on top, a "virtual machine", which will allow a standard language to be used regardless of the actual machine below. And, as Sun created Java, so enter Google.

Google’s OpenSocial is a set of common APIs to retrieve most of the kind of information you see on Facebook – profile information, the social graph, news feeds and other similar items. It’s not released yet, but fear of Facebook seems to have convinced just about every other social network to have jumped aboard – and a Facebook adaptor is not so hard to write anyway.

The power of this idea is quite monstrous – you will now be able to easily write an application that searches across almost ALL the social networks in the world, collates the information and returns a global view of an individual. Applications written on this platform will be able to infiltrate all supported networks. And I think Google might just have the clout to pull it off.

Where is this going? Personalised search – identify yourself on the OpenSocial platform and Google will be able to customise your search results,

and ads

, based not only on your profile (or part of it), but the profiles of the friends in your social network. It’s different to GMail in that by definition a social network is partly public domain and therefore fair game to data mining. Also, I don’t think it’s entirely coincidence that OpenSocial sounds so much like OpenID, and the two are likely to work together, so your social profile becomes linked to your online activity.

As always, be careful what you share online.

Comment on “Social network virtual machine”

Subscribe to our Newsletter