|
INTEGRATION WATCH: The Next Challenge for Open Source
By Andrew Binstock
 March 15, 2004 — A widely believed fallacy is that the global population will continue to grow inexorably and that in our children’s lifetime something will have to change if the Earth is ever to feed so many mouths. The reality is that almost all social scientists who study population trends expect global population to peak around 2050, before entering into a long cycle of decline.
Signs of this trend are already evident. Italy’s population has been contracting for years; Russia’s even more so. The problem governments face in reversing this trend is that volunteers cannot be induced to undertake the long-term, high-effort assignment of rearing a child if they don’t want to.
A common misperception in software is that open-source products are likely to be widely adopted in all segments of computing. IT’s embrace of Apache and Linux and its increasing friendliness toward MySQL and JBoss suggest that open source can prevail anywhere it chooses to go. The success of Sun’s OpenOffice would seem further confirmation of this. This well-deserved success has been validated by open source’s reliability, performance and richness of features.
However, open source is at the mercy of those volunteers for the development of new products. And like the countries of the world, it’s not clear the population of open-source projects is growing in the directions that enable it to expand its franchise. The problem is that volunteers cannot be induced to undertake a long-term, high-effort assignment (developing a community to support a full-fledged product) if they don’t want to.
One look at SourceForge shows the reality of this view: Of the 75,000 hosted projects, far fewer than 1 percent are at all active. The problem fundamentally is that good software takes a lot of man-hours to write and so inherently requires a dedicated community. If a group cannot be attracted to a given product, there will simply be no open-source implementation of it.
As a result, open-source projects tend to fall into a few defined niches: programming tools (of course), operating systems and admin tools, games and music. Go to SourceForge’s top 50 projects (based on their activity levels) and you’ll find very few products outside this narrow range. There is nothing wrong with this, but it does suggest that open-source developers spend a lot of time reinventing writing tools, such as IDEs, editors and compilers, in particular, as well as music rippers and burners.
For open source to spread into other areas, developers need to be induced into writing other software, such as business apps: data warehousing, full accounting packages, business analytics, ERP. In specific industries, the lack of open-source vertical applications is striking—just as it is in the consumer space once you remove games and music. So, what needs to change in this model if open source is to move forward?
Given, as I said earlier, that you cannot induce volunteers to dedicate considerable personal time to projects that don’t interest them, the solution will need to come from elsewhere. The vendors themselves will need to open their products as open source. This might seem ridiculously impractical. Why would a vendor of business analytics open up its source code just to be considered open source?
Actually, there are good reasons. Given the hammerlock IBM, Oracle and Microsoft have on the database market, and the recent flameout of Sybase and Informix, how could a new company expect to succeed in the database market? Let’s ask MySQL. Given the hammerlock IBM, BEA, Oracle and Sun have on the Java server market, how does a company hope to succeed in this market? Let’s ask JBoss.
Increasingly, start-ups that want to play in busy markets are finding success in open-source approaches. It gives them a distinct competitive advantage; and good licensing terms mean the company can seed the market with its products. If JBoss and MySQL enjoy successful IPOs during the next few years, this open-source model might suddenly catch on.
What is not clear in this scenario is whether two companies can compete in the same market with open-source products. Successful competition of this kind has never occurred. (The Linux market comes the closest, and it is hardly a supporting example.) He who goes first tends to do best.
If the next few years prove to be good ones for established open-source vendors and, more important, for IPOs of open-source companies, then open source will indeed flourish everywhere. Otherwise, it will remain confined strictly to the niches of interest to the volunteer programmers.
Andrew Binstock is the principal analyst at Pacific Data Works LLC
|
|
Advertisement
Click here for a complete listing of Integration Watch Columns
Click here to see a complete Column Archive. |
Back to Top
|