OpenSource vs. SaaS - The Final Word

Let me be the last to post about "Open Source vs. SaaS". Two excellent posts have been put up recently (O.K. not so recently.) Anshu Sharma's and Dave Rosenberg's. Both are very well written, and I agree with Anshu's arguments. That said, they both are essentially missing one essential point; all applications will be Web Applications (I think I'm going to say this in every post from here on out.) It doesn't matter how they are developed, people won't use them unless they can access them on the Web (my three kids don't even know what a disk drive is.) The question is why aren't more Web Applications being developed specifically as Open Source projects.

Let me say first, that the entire argument "Open Source vs. SaaS" is facetious. Open Source is a development model, SaaS is a delivery and usage model. Open Source applications can be delivered as SaaS and SaaS applications can be developed using Open Source methods. The argument arises because so few true Open Source apps are actually delivered as Web Applications (I use SaaS and Web Applications interchangeably.) Instead they are developed as single instance applications that a user installs.

Some companies then take this Open Source base and add Web Application functionality such as multi-tenancy and scalability as well as business functionality and flow to it. (We did as much with Dave's tremendous MuleSource product when we created the OpSource Services Bus.) But to say these apps are Open Source is the equivalent of saying SalesForce.com is Oracle since they built an app on top of an Oracle Database.

So why aren't there more native Open Source applications that are run as true Web Applications. Most are single-instance enterprise software that someone installs to use. The most compelling apps of the last 15 years, from eBay's bidding app, to Yahoo's Portal, to Google's Search and SFDC's CRM are all proprietary apps. Some say that SugarCRM is a Web App, but I think of them as a hybrid company selling both installed and SaaS versions of a singe app (and we know what I think of hybrids.) Ruminating with John Rowell, the only one we could come up with was Wikipedia.

Why is Wikipedia the only OpenSource/Web Application? Because running a Web Application costs money. You have to pay for servers and power and network and security and backup and so many different items, and that takes the Benjamins. Usually only commercial enterprises have the Benjamins to make that work, and Open Source communities don't want to develop for commercial enterprises. They'll do it for Wikimedia (the organization behind Wikipedia), because it's a charitable organization, but who wants to develop an app for Google or SFDC?

So back to my earlier thoughts. If the OpenSource development model is a good one, but all apps will be Web Apps (memorize this people) we need a platform where all of the expensive stuff is taken care of for the high-minded developers to start making apps. Then we can find a whole new non-topic to blog about.