The Platform and the Web – what can Facebook and Talis tell us?
There has been much talk of Platforms in the context of the Web recently, although usage has proved somewhat troublesome to pin down. In a recent podcast conversation with me, Tom Heath probed some fledgling ideas about the way in which resources might be considered as on or in the Web. I’m not certain, but it appeared to me that this useful distinction was taking form in Tom’s mind as we spoke (although he’s subsequently found a similar expression from way back in 2002). Thinking about ‘on’ and ‘in’ more over the past few days led me to wonder if the distinction might also be usefully applied to some of the different flavours of Web Platform that we see springing up around us at the moment. I also began to wonder if ‘of‘ (yes, one ‘f’) might be more compelling than ‘in’, other than the likelihood for everyone to read the dichotomy as ‘on’ or ‘off’ the Web, which would be exactly wrong.
I’d like to share some of these ideas here, and see whether or not they have substance. Examples for my hypothesising will be drawn from Facebook and the Talis Platform, but I’m sure that other platforms might also be examined in a similar vein.
The rise of the Web Platform
A couple of years or so ago, when Talis developers began working on our Platform in earnest, ‘platform’ and ‘web’ didn’t really go together in most people’s minds. As Wikipedia notes, computer scientists are liberal in their use of the ‘p’ word, and it cropped up in almost every part of our industry except in direct association to the Web itself;
“In computing, a platform describes some sort of framework, either in hardware or software, which allows software to run. Typical platforms include a computer’s architecture, operating system, or programming languages and their runtime libraries.”
We were therefore reduced to analogy in describing the things that our Platform would make possible; we talked about containerisation; we talked about Microsoft Windows and its removal of the need for developers to worry about painting pixels, spinning disks and understanding the movement of mice; we talked about the electricity grid; we talked about getting on and off steam trains without getting your dress muddy. We talked about providing robust, extensible, open and web scale foundations upon which we and others could build data-rich applications capable of externalising the silos in which so much potentially valuable data lurks.
Then Salesforce began to be described in platform language, and earlier this year Facebook launched ‘the Platform’ to usher in a period of explosive growth that has seen 100-150,000 new registrations each day, over 24 million active users (50% of whom apparently visit the site at least once per day), and phenomenal take-up of apps built on this Facebook platform by third parties. In another recent podcast, Jamie Taylor of Metaweb also described their activities in terms of ‘a platform’ that their first application, Freebase, builds upon, and we spent a little time considering just how many platforms the world might need…
All hail Facebook
In a recent piece by Kevin Allison in the Financial Times, Andreas Stavropoulous of Draper Fisher Jurvetson is quoted as saying;
“There is a giant sucking sound you are hearing in Web 2.0 start-ups, and it’s all going towards Facebook.”
Facebook has opened up to developers, and the site has become remarkably sticky as a result. The openness is not really two-way though, and as a result I’m forced to wonder just how much of a platform this is. Does it offer capabilities upon which third parties can build to deliver rich internet applications? Or does it just make the Facebook site a little better, drawing ever more eyeballs to that one place on the Web? An early answer in Facebook’s own Platform FAQ [PDF] is telling in this regard;
“[The] Facebook Platform is a development system that enables companies and developers to build applications for the Facebook website, where all of Facebook’s 24 million active users can interact with them. Facebook Platform offers deep integration into the Facebook website, distribution through the social graph and an opportunity to build a business.”
(my emphasis)
Sure, there’s value in something like that, and 24,000,000 human beings can’t all be wrong (all the time), but the potential is surely so much greater if you’re brave and powerful enough to fundamentally shift the model. Why does Facebook ‘make’ me visit my page on their site? Why does it send me an email to tell me that someone has written something for me, and then make me go to facebook.com to see what? Why not just tell me who said what, and annoy me a lot less? In these days of RSS feeds, widgets everywhere, more one-stop-shops than any of us could possibly ever need, why can’t Facebook (or MySpace, which I have never used, or LinkedIn [I'm here], or Bebo, or whoever) operate on my terms, allowing me to place the functionality that I use in contexts of my own choosing; whether on or off their site. Openness is great. But it’s a two-way thing, and Facebook don’t seem to get that yet. Presumably their revenue model would require a pretty significant shake-up before they would consider the idea, but they’d better be careful as one of their competitors might take the bold step before they get around to it.
Which brings us back to on and of the Web…
Facebook and Talis might very well be offering ‘Platforms’, but they’re quite different in intention. Facebook’s platform seems to be all about making the Facebook site as rich, compelling and sticky as possible; everything is sucked to one point. The Talis Platform, on the other hand, is about providing developers – wherever they are – with the tools and capabilities to easily link and manipulate data across and through the web. The former sits heavily ‘on’ the web, and feeds upon it to suck ever more into its maw. The latter is truly ‘of’ the web, giving a distributed community of developers and users powerful new capabilities to enmesh their applications, and to deliver capabilities at the point of need.
Both appear, when viewed from one perspective, to be the same.
But they’re so very different when examined properly.
Platforms that are ‘on’ the web are a richer and more capable return to the days of portalisation. Platforms that are ‘of’ the web work with what they find, and make everything that much easier, that much better. There’s a place for both, but we need to be clear as to which type we’re referring if we are to avoid confusion.
This, from Nova Spivack [another of my podcast guests], contributes to the picture too, as do two posts from the Harvard Business School’s Andrew McAfee. There’s clearly some further digging to do here…
What do you think? Does the story make sense? Is the distinction useful?
Today’s Creative Commons licensed picture of Platform 9¾ at London’s Kings Cross station is from Flickr and by Siriol Griffiths.
Technorati Tags: API, Facebook, Freebase, Metaweb, Platforms, Facebook, Radar Networks, Semantic Web, Talis, Talis Platform, Tom Heath, Web 2.0, Web 3.0, web services
July 5th, 2007 at 5:42 pm
Paul,
This all comes down to an interesting Dichotomy relative to the Web:
1. Platforms for Exposing Data & Services On The Web
2. Platforms for Injecting Data & Services Into The Web
If the end result (output) of platform usage / incorporation doesn’t result in Web Resolvable Identifiers for Conceptual Entities (URIs or IRIs) in the domain covered by the Platform, then at best it is emitting Domain Specific Data “On To The Web” i.e. the Data is “Mash-able” but not “Mesh-able”.
As I’ve postulated for eons (internet time), we are dealing with exposure of individual perspectives via data containers best described as “Data Spaces” (ultimately). This is a important but often misunderstood concept (btw – you got it right in your post on the subject). In due course I expect the concept to “self crystallize” since “Objects” and associated “Values” that connect us are becoming clearer by the second (again, Internet time)
The Web is more about “Presence” than “Web Sites” or “Web Services” or “Web Data”. The points of “Web Presence” are evolving rapidly.
btw – Semantic Data Web middleware can take Facebook Data “on the Web” and Inject “into the Web”. In a nutshell, I am saying that FQL & Facebooks REST services will be emitting Triples very soon via the Virtuoso RDFizer known as the “Sponger”
So we will have dereferencable URI/IRS (i.e. Entity Identifiers that act as data pointers) for Facebook data alongside Googlebase, Amazon, eBay and other Web Services Providers.
July 5th, 2007 at 9:49 pm
Flippin’ ‘eck
Andrew McAfee, associate professor at Harvard Business School and the man who coined the Enterprise 2.0 term, has been advocating a ‘flip test’ for assessing the likely impact of new technologies. He got the idea from Professor Stanley Crouch. Presumably
July 6th, 2007 at 12:45 pm
A great post, Paul!
I think the distinction between the two models is a very valid one!
Applications of the near-future are either going to HIEST information from other silos, or the silos will COOPERATE with the web.
I don’t think there’s any way around this fact.
I hate to put bad mo-jo on an exciting company, but really, I don’t think FaceBook is sustainable as a “platform” for much longer than a year or so. If they make their API become more like a limited RDF access-point or something, then maybe they can survive as a big owner-of-users for a while. Open data is coming to the Web! It must, because Users will take matters into their own hands and share data either way.
Thanks for the post!
July 6th, 2007 at 1:03 pm
Kingsley – thanks for the comments, which Danny Ayers has some thoughts about over on his blog.
To anthropomorphise these things a little (probably too much!), there’s almost a selfishness about the services that simply expose data on the web; they’re taking, without giving much/anything back to the wider Web in return. That’s not to suggest that their users don’t receive some value from the site, but the wider commons certainly misses out.
The difficulty, I guess, is in persuading regular users – and app creators – that there is a bigger market for their ideas out on the open web. 24million is plenty big enough for most!
July 6th, 2007 at 1:09 pm
Andrew – thanks…
Without some compelling alternatives, the ‘on the web’ Platforms like Facebook probably have a far longer life than the year you give them.
Many of those 24million Facebook users are probably perfectly happy with what they have. It’ll take more than an ‘open data’ argument to make them shift allegiance.
So here’s the thing; how to make ‘linked and open data’ the next cool and fashionable must-have for the typical Facebook user… and answers can’t include ‘triple’, ‘RDF’, api’, ‘data’, ‘uri’, ‘REST’, ‘triple’, or any of the other arcana to which we turn in explaining what we mean to one another.
July 6th, 2007 at 1:12 pm
Hmm. Two ‘triples’ in the previous comment. I wonder what that makes…?
July 7th, 2007 at 6:40 pm
I appreciate your effort to try and link the Talis opportunity to Facebook however I think most end users are looking for more of the simplicity of Intel’s Mashmaker.
http://www.siliconvalleysleuth.com/2007/06/intel-envisions.html
I think platform’s like Talis should focus on data that could benefit non social network applications and mashups. Rather than Facebook, which appeals to egos, Talis and other data neutral efforts should focus on the dollars and sense of government. Their is a real benefit if we could be more closely monitoring what government is spending money on and ways to decrease government expenditures.