In the past few years, free and open source software (FOSS) adoption has enjoyed tremendous growth in the IT industry. This trend is echoed in the geospatial technology industry. In the geospatial arena, growth is evidenced by the popularity of many tools and applications, and the emergence of a vibrant community of FOSS developers and users (search OSGeo for examples). In spite of these impressive gains and widespread adoption, the general understanding of FOSS remains cloudy.
One reason for the confusion that surrounds FOSS is that FOSS caters to multiple stakeholders with widely differing interests. These different points of view are often merged into single threads of discussion, obscuring the issues. In order to shed light on this situation, I suggest using “personas” to aid in the analysis of various perspectives related to the acquisition and usage of FOSS. Personas are a technique used in marketing and in software development to identify user types. This definition comes from Wikipedia: “A user persona is a representation of the goals and behavior of a real group of users. In most cases, personas are synthesized from data collected from interviews with users.”
This article does not intend to engage in a thorough research of FOSS as viewed from the lens of various personas. Even properly defining the right personas requires a good deal of research and analysis, as well as interviews with real practitioners. This could, indeed, be a nice topic for a graduate thesis. A very interesting aspect of using personas is that the researcher can define a persona that represents a specific community. In this manner, the analysis can be specifically tailored to determine how FOSS is viewed by the geospatial industry.
In order to explore the topic of selecting FOSS software, I will define a specific persona. This persona is “Jack E. Parcel,” a GIS analyst for a medium-sized city. Jack has a degree in geography and is proficient in using GIS desktop software. Jack has become familiar with Web GIS by attending presentations given by his friendly GIS vendor. Jack understands application requirements well; however, Jack relies on the external consultants for system implementations. He has been assigned to work with “Kate M. Server,” his IT manager, to select the GIS platform for a customer facing Web application. Jack has heard good things about FOSS from Kate, so he decides to do some research on the Internet. Very soon Jack gets confused.
Now that we have our persona, Jack, let’s examine FOSS selection from his vantage point. First, Jack is not a programmer by any stretch. Thus he has no interest at all in the source code. He is interested in the functionality: does the software do the job? Many articles Jack has found on the Internet talk about the development process, committers, steering committee, the trunk, etc. All of these subjects are very interesting, but irrelevant to Jack since he will not manipulate the code. From Jack’s perspective it looks like there is not much difference between selecting proprietary software or FOSS. Let’s look at this a bit closer.
When selecting software, you first consider your requirements. Does the software do the job? Here you will find detractors who will argue that FOSS is not as strong as some proprietary software. This may be so in many cases but the same holds true for proprietary software: there are proprietary packages that are not as strong as others (an obvious observation). In a recent Directions Magazine webcast a listener asked whether open source scales. The context of the question was about map servers. Again, the answer is in the same vein: some FOSS map servers scale very well, others not as well, just as with proprietary software. Thus, this question boils down to choosing the right tool for the right job and price, versus performance tradeoffs. Perhaps one aspect where FOSS enjoys some advantage is in the no-strings-attached opportunity to try and test a full copy of the software. Proprietary software evaluation copies are often restrictive.
Next, you generally look at total cost of ownership (TCO). This includes initial license cost, annual maintenance, implementation cost, training cost, customization, etc. For FOSS, the factors to consider and the calculations are identical; the only difference is that license costs for FOSS are zero. FOSS detractors will argue that more often than not there is no support, training or other services available for FOSS. The reality is that today most successful FOSS geospatial products have well-developed support, training, consulting and implementation ecosystems. The existence of such should be a primary selection criterion. Granted, in many cases such services are not available, but the same is true for proprietary software. Just as for FOSS, there is proprietary software with poor customer support, no training or other services.
Having explored Jack’s perspective, let’s turn our attention briefly to a very different persona: “Joe C. Sharp.” Joe is a programmer who can figure his way around complex systems and complicated code and has great fun doing it. Joe is a GIS consultant implementing websites for small businesses leveraging on third-party GIS. Joe’s view of open source is totally different. For Joe the availability of source code and the ability to make changes quickly to customize functionality or make fixes are primary concerns. Thus, existence of a robust and active network of developers for peer support is a very important consideration.
It is important to analyze adoption and selection of open source from a specific point of view - a persona. From the perspective of an end-user persona exemplified by Jack E. Parcel, there is really no difference between choosing FOSS versus proprietary software. The same selection criteria should apply in both cases. Regardless, risks, cost and other challenges have to be considered carefully.