Aping the worst of proprietary software
Open source looks eerily like proprietary software in building products, not solutions, for end customers. We need to open up more of our software to allow for integration between projects.
Dave Dargo has written a thoughtful piece on one problem with proprietary software today: it spends too much time isolating itself as a product, rather than opening up itself and combining to create solutions. As it turns out, open-source software is following in these same destructive footsteps, as we notes:
In the open-source world we still mimic much of the marketing and product management methods of these old-school software business models. One of the greatest promises to me of open-source is the ability to integrate products and build completely new solutions. One can not take Oracle, or a sub-set of Oracle, and combine it with technology from Microsoft Windows to build something that didn't previously exist. Licensing and intellectual property protection prevents that type of innovation.
With open source, however, one can take source code and products from different projects and combine it into unique, new solutions designed to solve actual problems faced by large IT organizations. It's one thing to save an IT department money at the acquisition point, it's quite another to solve the real-world problems they face every day.
Dave is right: we're not much better at this than the proprietary world, possibly because our training comes from that proprietary world. Having said that, when I look at my company's product, I see integration of the Spring Framework, Jakarta, Hibernate, and a range of other open source products. (In fact, if you look closely, you'll see that many proprietary products incorporate open source technology, including things like MSN Messenger from Microsoft.
I, personally, am not looking for a third-party foundation or a dedicated software integration to do this. I think the best way to realize Dave's vision is for product companies/community projects to look beyond themselves to find interesting ways to extend functionality to create these solutions that Dave envisions.
Of course, this requires true open source licensing. It's tough (actually, impossible without a contract/license) to integrate mixed source code, because the process of code integration is bogged down by the bureaucracy of corporate integration.
I would therefore suspect that the most widely integrated third-party technologies will be those that are the most open. The question will be, however, whether ubiquity is a good trade-off for conversion of all use into cash. I think it is, but that's a serious discussion to be had....