2007 was a year where Open Source saw big gains in its acceptance. Open Source as applied to enterprise software, mobile applications, and social/collaborative software have all been popular and have seen a lot of growth.
But one problem that continues to haunt Open Source is interoperability of all the pieces. And there often are a lot of pieces. On the one hand, Open Source licensing encourages innovation because there’s no need to reinvent the wheel. It’s very rare to find an Open Source solution that can’t be decomposed into at least a half dozen or more Open Source components.
This is clearly one battle where Microsoft’s monolithic approach to architecture shines. Dig into a Java project, and you’re likely to find a list of components that include things like Lucene, Hibernate, MySQL, JSF, Richfaces, Tomcat, Spring, Xerces, and many more. Yikes! How can you keep track?
The trick is to get a stack of integrated Open Source products that all work in harmony. Popular frameworks and stacks that combine many components together can ease the problem, but headaches still abound.
Enterprises customers are wary of these kinds of problems and are delaying adoption because it it. These customers need assurance that these stacks of software can be updated without breaking their applications.
“There is generally a lack of certifiable quality when it comes to solutions that build on multiple open-source components, where a change in one underlying component can break the entire solution,” Kim Polese of SpikeSource said. “As long as this remains a concern for enterprises, open-source vendors will not realize the rates of adoption they are hoping for, and adoption beyond the enterprise into SMB [small and midsize business] customers will also be limited.”
A recent report of the Open Solutions Alliance also cited Open Source interoperability as a major problem. Companies like SpikeSource are addressing interoperability, but there is much more that still needs to be done.
