"If you asked a developer if they'd like to work with open-source or commercial software, eight times out of 10 they'll say open source," Lyman contends. And some developers may charge less than developers who work with commercial products.
Hamadeh says that with SugarCRM, it's even possible to "have a local student come in and program something in a couple of hours," Hamadeh says, or a tech-savvy business person can create custom modules. But, he cautions, while there are some SugarCRM consultants who will do a great job, they can be expensive, so having internal IT talent can help you avoid added costs.
Brisbin points out that the success of open source at NPC is due largely to the fact that its developers have a breadth of knowledge and are willing to work outside of narrowly defined silos.
"We have small development teams, and we cross areas of responsibility," he says, noting that he routinely moves among RPG, Java, Web front-end development, PostgreSQL and the underlying application system. "There is a critical mass of information you need to have as a developer to do open source effectively," Brisbin adds.
And then there's one of the more hard-to-quantify costs: cultural change. Mentkow says Roanoke's move to OpenOffice involved changing the culture as much as it did changing the desktops. "Cultural change does not happen in moments," he says. "As we move to different platforms and different standards, what we have to see is an acceptance of those changes."
Sims adds that it's easier to achieve cultural change at organizations that value resourcefulness and courage, since moving to open source represents a break from the approach that involves seeking traditional answers to difficult problems. "People still say you can't get fired for buying Microsoft or Oracle -- how about, you should get fired for not coming up with the best scenario that meets your company's unique criteria, regardless of conventional wisdom," he says.
As open source matures, companies will begin to get past the misconceptions, understand the implications and balance the benefits with the downsides. "Most of the time when there's a problem, it's because there's an assumption of 'It works, and when it doesn't, we'll fix it ourselves or find the answer on the Internet,' " Driver says. "Or there's an assumption that the cost of acquisition can be extrapolated to total cost of ownership. But there's a care and feeding cost to everything."
Brandel is a Computerworld contributing writer. Contact her at firstname.lastname@example.org.