MySQL is one of the positive examples of open source. However, I find that far too many people go for open source due to short term advantages (e.g., it is free), without adequate consideration of the full cost/benefit considerations and the long term implications.

To begin with, to a certain extent you get what you pay for. If the package is free, then the support and service is going to be limited as there is no funding for it. Not many people would buy a commercial package with no guaranteed support, so likewise they should carefully consider this for open source.

Secondly, although packages such as MySQL have a large development and user community, many other open source packages have a limited development community. Often the development and support community is largely driven and organised by a single individual or small group of individuals. When these people move on, even though the design community may still exist, the driving force to make it effective is lost. This is much less likely with a commercially successful package, which is protected by past financial investment and commercial commitments and future revenue streams. When looking at open source, one must ensure that the support and development community is sufficiently large and diverse that this is not an issue.

One also needs to consider the purchase price savings, in light both of the total cost of the project and in light of potential risks. A hypothetical example: an open source accounting package for free or a commercial one for $10000. The former provides a substantial day-1 saving. However, if one has 3 staff using the package full time (moderately large company) to input information then the labour costs alone are about $300 000 per year. So the $10 000 saving is trivial in the bigger picture. A commercial package which has a 5% more efficient interface will more than pay for itself in less than a year on labour cost savings alone. Also, support and documention limitations typically asssociated with open source can easily cost more than the initial savings.

As a software integrator, my experience makes me very wary of open source. There are of course noteable exceptions (such as MySQL). However, in most cases I find that commercial packages are a better long-term business decision than open source.

I'm also wary of people why say they they will try open source and if it doesn't work out they will switch to a commercial package. Although this is possible, the costs of migrating data, changing procedures and modifying interfaces to other software can be prohibitive.

Am I'm being overly negative about open source? Perhaps so. However, it is a reaction to the overly optimistic and naive approach that I often see to this subject.