The future is open

Can you tell us a little about Drupal’s origins?

When I was a student at the University of Antwerp I created a forum, similar to an intranet, to manage my projects. It all started very informally. Basically it was a message board of sorts, used for our work, but also for coordinating our social activities. But when I graduated I wanted to keep it going so that I could continue communicating with these folks. That’s when I registered the site under the domain www.drop.org, which actually was an error. My intention was to call it ‘dorp’—the Dutch word for village, to emphasise its community feel—but I made a typo when registering the domain and created Drop instead. All very coincidental, but it turned out to be a better name I think.

The reason it became popular is probably because I always tried to experiment with new technologies as they emerged. For example, we were one of the first to implement RSS and public diaries as they were called then (blogs).  It’s that experimentation that led me to release the software to the community. People were coming up with suggestions all the time so it seemed like a good idea to let them experiment with the code themselves. That was in January 2001, when I first made the software publically available as Drupal 1.0.0. But I think even today that spirit of experimentation, of trying to be the first to introduce new technologies, is what gives us the edge. 

In the years following the release I continued working on Drupal, mainly as a hobby. Then four-five years ago I went back to university to do a PhD in computer science, which incidentally had nothing to do with the web. But when I graduated it was time to make a decision about what I’d do next.  In the years that I was back at university, Drupal had gained a lot of momentum but it still was used mainly by independent developers, small outfits.  I wanted to bring Drupal to the next level.  If you look at the really big open source projects, like Linux, Ubuntu or MySQL, you’ll note that they all have a large commercial company behind it to support the software at enterprise level. That’s why I co-founded Acquia in 2007. The reason that the White House uses Drupal is partly because of Acquia.  Large organisations need that level of support. The Economist is moving to Drupal too; to do that successfully you need a decent support organisation.

How do you see the open source movement developing?

I think people are beginning to recognise the power of open source, in particular its disruptive power.  The way it is disrupting business models; the licensing model for example.  Or the control we have over companies; at Acquia, for example, we have no IP.  And the way we innovate, in the open source community, as opposed to behind closed doors. Open source can play a huge role in the future development of technology. In a way it’s a very logical evolution. Look at the way most industries evolve. Take electricity as an example. When electricity was first discovered people started developing complex, totally customised solutions. Next came more standardised products and today electricity is totally commoditised. We don’t even think about it anymore, it has become a pure utility. But this evolution repeats itself in most industries, and now in computing too. The old computers were huge bespoke systems, and then evolved to mainframes, to PCs, and now to cloud computing. It seems to be a very natural evolution in the way industries evolve. Software is moving in the same direction, from product to utility. And I think that open source will be critical in this evolution because a utility implies much lower prices, on the verge of free. But to get there you cannot operate on a licensing model. That’s why open source will win. A licensing model simply can’t compete. And it’s not only about cost; the innovation process too is just so powerful. To illustrate, Drupal actually has a reasonably small core, but has 4000-5000 modules. The only way to build that is via open source. A proprietary vendor cannot afford to develop so many modules. In the years ahead I think that open source will be the ‘default’ software model. Obviously there will still be proprietary software, but only in specialised niche markets.  

CMS software is going through the same transition toward a utility model. First people started making websites manually, using html code. You needed technical skills to make a website. Today CMS has largely eliminated the role of the web developer, or at least, redefined it. The webmaster today is a content manager. And the process continues. Take Drupal Gardens, a service we’re setting up whereby anyone can build a social site on the fly, hosted in the cloud. CMS too is becoming a utility.

How does Drupal Gardens differ to established cloud services like Ning?

There are several differences but the core difference is the open source model. Most SaaS (software-as-a-service) offerings today still lock up their code. Some services do allow you take your data out—think of Salesforce.com. But Drupal Gardens allows you to export your data and your code. Everything is open, including the look and feel of the site that you may have added. Things are changing though. Take Twitter and Facebook. They’re not open source, but they have adopted some open source principles in the way they published open APIs for external developers. The same goes for Salesforce.com. The key shift is in the way these companies are looking at innovation. They’re opening it up.

How do you see the web developing in the years ahead?  What are your thoughts on the semantic web? 

Web semantics is basically about data interoperability. The problem with the web today is that it loses the semantics of the underlying data. There is a huge amount of structured data in the databases behind websites. CMS generates web pages on the basis of this underlying data but in the process of generating those pages it loses the underlying structure of that data. The semantic web defines that structure in the html; it’s basically about creating more metadata. Now that’s a pretty revolutionary shift because it will allow for much more interoperability between different sites and services.Tim Berners-Lee, the inventor of the web, says that the WWW will evolve into the GGG: the Giant Global Graph. It’s quite a technical story but essentially it is an evolution toward a single giant database, one giant graph, which all browsers and computers will understand. Obviously we need common standards to get there. Such standards do exist, but unfortunately mainly in the academic realm. It’s a chicken & egg problem. Someone needs to take the first step.The first step is to make data semantic; then later the applications will be developed. We’re hoping to play our small part in this by making Drupal 7 semantic enabled. I think we’re the first open source system to do so. Half a million sites are built on Drupal so it could be a significant contribution. Hopefully we can be a small footnote in the evolution of the semantic web.  

How does all this relate to the social networks, in the way that Facebook, for example, is offering social gaming developers partial access to the ‘social graph’ (ed. note: the social graph is the basic dataset of all social networks, covering users’ profile information and the connections they have with each other)?

The social graph is a subset of the GGG, it’s just one element of it. What’s important here is the portability of data, to such an extent that it doesn’t really matter what social network site you use. I expect a scenario where people have more control over their own ‘site’, their own profile, and their connections. I look forward to a much more distributed model in the future, whereby plenty of smaller sites interoperate with each other and share data. OpenID is an example of such a standard.  I guess this will be the next battle; the battle for data. Open source is well on its way of winning the battle for code. That happened quite recently, if you consider how attitudes have changed in just five years or so. But open data will be the next shift I suspect; although we still need to figure out how to manage that process.