Tamale has always used open-source components in our architecture, and so our customers (especially the software analysts) often ask us about trends in open-source software. They also ask about our hybrid open/proprietary model. So, I thought I'd write a very brief analysis of the last 5 years of open-source from Tamale's perspective. Then I'll explain how our proprietary software fits into the picture.

In the 1990's, open source and Linux were considered identical, but by the time Tamale started operations in 2002, opinion was shifting. Linux was viewed as a social experiment for the first ten years of its life. But by the turn of the millennium, the questions had shifted from whether Linux would keep growing to whether Linux could replace Solaris or Windows. Linux may not have been considered enterprise-grade just yet. But the consensus was clear that not only would Linux become enterprise-grade eventually, when it did, there would be massive adoption.

As a startup planning a multi-year stealth development strategy, Tamale was looking to the long-term. So, crazy as it sounded then, Linux was an easy call. We wanted a price point that would be lower than Windows, and stability that would rival Solaris. Linux had both.

The really hard decision was for the application infrastructure. During the 1990's, software had experienced the "middle-ware revolution" (complete with dedicated media coverage). Sub-systems like application servers and messaging servers had become as well-packaged as databases. Proprietary vendors like BEA and Tibco exploded in growth, and standards like J2EE had helped consolidate the market by eliminating competing approaches. Analysts argued then that middleware vendors were "up the stack" from the operating system, and would benefit from open source because operating system budget dollars would be spent on middleware. Middleware was a new technology space, the argument went, and open source could only displace mature technologies.

Then along came JBoss. In the space of just a few years, JBoss conquered the App Server market. And it happened as that market was maturing, not after. Not only did JBoss capture an emerging market, they are now undoubtedly driving innovation. JBoss adheres to standards like EJB3.0, but they are also developing and popularizing alternative approaches like Hibernate. JBoss has become a kind of clearing house that legitimizes new open-source technologies. Projects like Hibernate, Drools, AOP frameworks, and JGroups are innovative and original infrastructure technologies. Suddenly, open-source came to mean more than Linux to the broader IT industry.

Today, there is growth in open-source technologies that are at the top of the stack. Writely, openoffice, BIRT, Nutch, GAIM, and Zimbra are all open-source applications (as opposed to application infrastructure).

There is one common characteristic for all successful open-source technologies: broad, horizontal appeal. It doesn't matter whether the software is for managing database transactions or for managing calendar appointments. If there is a large community of users, there will inevitably be a large community of developers (because users tend to pay developers' salaries). If there is a large community of developers, conditions are right for open-source solutions to win. The fundamental strength of open-source software is the effect of large developer communities, which Linus Torvalds so aptly explained (and which Eric Raymond so famously quoted): "Given enough eyeballs, all bugs are shallow." Put simply, the bigger the market, the greater the advantage of the open-source strategy.

So, if open-source software can effectively outpace innovation in an emerging market, what role is left for proprietary software? Will everything eventually be open-source? The short answer is no.

Operating systems like Linux, or search engines like Lucene, or App servers like JBoss, or collaboration servers like Zimbra serve global markets. Their customers cut across every industry, yet the problems they face have enough in common to warrant generalized treatment. In fact, all four are technologies that allow developers to focus on domain specifics, rather than on "boiler plate".

However, there will always be extremely valuable vertical markets. Tamale focuses on developing highly specialized software for the financial industry. We have the same goals as many other application vendors -- speed, scale, reliability, search -- and we use open-source components to help solve these horizontal problems.

Despite all the leverage we find in open-source, we have countless design and development goals that are specific to our target industry. Specializing search for financial research, or helping analysts discover and document the relationships between companies are specific to the investment industry. These are the type of features for which open-source would provide little benefit, and which create a market opportunity for Research Management Systems.