"Hey dude, is it time for me to switch to open-source software?" That's a characteristic opening line of an e-mail message from Jim; he doesn't beat around the bush. Over the years I've had a running correspondence (kind of a low-profile consultancy) with him, especially since he turned entrepreneur and started a retail company. Jim's a friendly but unstoppable go-getter with success following in his wake, so I wasn't surprised that he managed to just about double his business every year. That doesn't mean it was easy, and I know from the correspondence that computer systems contributed to slower growth than Jim hoped for.

After all the hoopla the past couple of years, some people might wonder how Jim is just getting around to thinking about open-source software. He isn't. Jim trained on UNIX systems in college. He's no nerd, but he is IT-literate. However, as a businessman he's been preoccupied with running the company. Most technical issues were resolved by taking the shortest distance between two points, that is, how quickly could he get something working rather than answering the question, "Is this the best way?"

Until recently, the default position for most of his computer decisions favored, shall we say, branded approaches. Now he seems to have come to a point with his business where an overview is possible. Jim's starting to ask the bigger questions, e.g., can we do better? Of course, better is a slippery word, but I think in Jim's straightforward way he means some combination of more functional, more reliable, more flexible, and less expensive. Fair enough.

That's what we're talkin' about

Although he didn't define it, I'm assuming Jim meant open-source software such as the Linux operating system, Apache Web server, PHP or Perl scripting languages, and MySQL database management system. There are many others, but these are the best known. I don't think Jim is directly concerned about the open-source philosophy. As a small-business owner, he's less interested in how programs are written than in their potential for ROI.

Three reasons are often cited for aligning with the Linux, Apache, PHP, and MySQL axis: These programs work, they're not from Microsoft, and they can be much less expensive. Jim doesn't have much time for hating Microsoft, but he does have concerns about reliability and cost. In his e-mail he posed four questions along these lines; I'll use them for a framework.

1. What does it take to do it?

A lot of companies have been using open-source software for years, though most have been big companies. The movement to adopt open-source software by mid-size and small businesses is relatively new and it's different, something like playing baseball with a two-man team. Most smaller companies don't have IT departments, which is another way of saying they may not have the people--people with the right knowledge--to implement open-source products.

Converting a server to Linux, setting up a Web site with an Apache server, or attaching an application to a MySQL database requires people who are knowledgeable. Does it take more knowledge than similar products, like a Microsoft operating system or an Oracle database manager? Not necessarily, although on average open-source software is technically demanding, but the real difficulty has been finding the right people.

Fortunately, that's changing. If a company is located in a relatively large community (say a city of around 100,000), there will be people with open-source knowledge available for consultancy or employment. For another thing, some very large companies--notably IBM--now support open-source software (at least some of it).

Having, or getting, the people with the right knowledge is a general requirement. Beyond that, smaller businesses present individual situations. A new business doesn't have to fight legacy software or retrain an office full of people. Open-source products don't support some applications; a company may have critical software that runs only on Windows. One of the important caveats for all software, and open source in particular, is to be certain the appropriate software for a company is available.

2. How far can we go?

One possibility is to use only open-source software. It can be done. Although best known for server-based computing, there are a growing number of programs for the desktop and office use. Sun's Star Office and Lindows (the Linux shell that works like Windows) are covered often by the press, but there are many other products. With few changes in hardware, a company could replace not only server software but also all the desktop software.

This sweeping approach should cause a puckering in the gut, because such a tsunami of change is terrifying. There are so many people who are accustomed to and depend on particular software, that radical change risks all hell breaking loose. This applies to the Windows desktop hegemony, but it's also true of dependencies related to Oracle databases, IBM systems, and Sun servers (among others). Sometimes--no, often--it will be necessary to mix open-source with proprietary software. A company might run only Linux servers, but it will need to have proprietary database software. This is what companies like IBM and Oracle now promote.

Whatever the situation, most companies (not just smaller ones) will need to hire expertise. This costs money, which works against the cost advantages of open-source software. So how far can you go with open source? In most cases about as far as the people you have and the people you can afford to hire will let you go.

3. How do we measure success?

Jim has essentially two ways to look at success: It costs him less than what he's doing now or the software provides something he doesn't already have, which could be better performance, new features, or more reliability.

A lot of open-source software is free ... or at least sort of free. The software itself may not have a price, but implementing it certainly does. All the major titles can be acquired from multiple sources, many of which are companies that provide a range of fee-based items such as improved documentation, technical support, and consulting. However, even including such costs, it amounts to a fraction of what similar software costs are from a brand like Microsoft, Oracle, or IBM.

Savings are most likely to be seen if a company is currently paying yearly licensing fees or is locked into an expensive upgrade pattern. In Jim's case, he will be able to switch out of very expensive data management system (which was oversold in the first place) and bank nearly $80,000 a year in savings. This is one metric that should be relatively easy to apply.

I've noticed over the years that open-source software is very successful doing the routine things. Of course, the software can be pushed in many directions; after all, open source means code can be modified and customized. But most companies don't need to modify the software; they need reliability and acceptable performance. Reliability can be measured by hours of up (or down) time. Performance can be measured in transaction processing capacity. In general, open source software will be a good fit for small and mid-size companies.

4. Will the open-source idea last?

Not long ago Jim had a visit from a Microsoft representative, and he was left with concerns. A few years ago there were a lot of doubters, people who said that the open-source route to commercial computing wasn't going to prosper. Well, it's done all right--or very well, depending on your perspective. Though there are still doubters, even they tend to agree that for the immediate future open source is viable. Advocates are generally convinced that open source is the way of the future.

A relative consensus should be enough for people like Jim who are thinking about adopting open source products. In the fast-paced technology business, very little endures more than a decade. Some technology will last longer; just don't expect it. So if you install Linux, Apache, and MySQL systems, what matters is that they will be supportable for at least a decade.

I'm glad to say Jim isn't leaping into open-source software. He decided to find out what specific open-source software is appropriate for his company. He is locating people in and out of the company who can work with the software. Open-source software is a good bet for him, but he's using knowledge to make it less of a gamble.