Does open source matter?

February 27, 2006— I can't understand why folks want to vilify Oracle for buying up open source companies such as Sleepycat and Innobase. By the same token, I can't really fault those companies for selling out, either. For small software companies, getting bought has always been a viable exit strategy.

It seems that, although some misgivings may remain, the open source community is coming to terms with the realities of the software business. Whereas at one time an open source software company might have been booed out of the room for charging license fees for commercial deployments, these days that practice is becoming the norm, as Matt Asay observed in a recent blog post.

But what of that? If a business model that says "develop for free but pay to deploy" is acceptable, then what does enterprise IT need open source for, anyway?

While we're talking about Oracle, let's use databases as an example. When you choose to deploy an open source database, you aren't basing your decision on technical superiority. MySQL can't compete with Oracle 10g on features. Despite the technical excellence it brings to the table, neither can PostgreSQL. No open source database can. Oracle's R&D head start was decades long.

What you do get from open source, however, is the freedom to experiment. You can set up your server and begin building your application. Your developers can get up to speed with its capabilities and its unique dialect of SQL. Along the way, you can begin to answer that all-important question: Will this work? And, most important of all, you can do all this without shelling out a dime.

In the Bad Old Days of proprietary software, this was impossible. In some cases, developers couldn't even so much as crack a reference manual before a contract was signed. But today this is changing. Databases are a prime example. Each of the top three relational database vendors—IBM, Microsoft, and Oracle—now offers a version of its flagship product for download, free of charge.

True, the free versions don't offer the full feature sets of the enterprise versions, but to call them "crippled" is overstating the case. IBM's free DB2 Express-C, for example, supports up to four x86 processor cores and 4 GB of RAM on either Linux or Windows, with no limits on the amount of data stored or the number of users. You can even deploy it commercially if you're okay with those constraints. Not so long ago, a lot of production databases were rolled out on less.

What caused this change of heart on the part of some of the world's biggest commercial software companies? Make no mistake, open source had everything to do with it. If it weren't for the proliferation of viable free alternatives to their products, these companies would have had no reason to change their tune. But the reality is that the software market itself has changed—changed forever—and software vendors have no choice but to adapt.

Community-based development is the next frontier. All companies respond to the needs of their customers, but traditional proprietary software development processes don't foster the same kind of relationship between developers and end-users that open source projects do.

Still, that too is changing. Sun's Java Community Process, for example, isn't open source, but it's close enough for the majority of serious Java developers. And Oracle's recent open source acquisitions are a positive sign that Oracle is thinking seriously about the ways it interacts with its developer and customer communities.

So how about it? Traditional, proprietary software practices are no longer serving the needs of enterprise IT, but so far open source hasn't provided all the answers, either. If commercial software companies can combine their technical know-how with the community interaction and easy access to technology available with open source projects, wouldn't that be the best of all possible worlds?

This story, "Does open source matter?" was originally published by InfoWorld.