November 16th, 2009
Like many other librarians with a technology focus, I’ve been following the discussions about the Sirsi/Dynix position paper ‘Integrated Library Systems on Open Source Platforms’ with interest. Many other bloggers have identified the main problems with the paper, which is superficial and uses misleading over-generalisations to argue against the adoption of integrated library systems released under a free/libre open source license. Of course, this isn’t particularly surprising given that the author works for Sirsi/Dynix, a company that has been selling proprietary integrated library system software for over 25 years. Members of the code4lib community have done an excellent job of tracking the commentary at SirsiDynix:_Integrated_Library_System_Platforms_on_Open_Source. The high quality of this feedback means that I don’t feel it’s necessary to write yet another point-by-point refutation of the position paper.
However, I do have a perspective that I haven’t seen reflected explicitly in any of the responses, and that’s the view that by talking about open source, people place themselves in a weak position. I have some sympathy with Stephen Abram’s statement that the term ‘open source’ is misunderstood and vague. Although the OSI’s Open Source Definition says that open source doesn’t just mean access to the source code, and specifies ten conditions that are necessary for software distribution to qualify as ‘open source’, it is reasonable to assume that people who hear the phrase ‘open source’ for the first time will think that it only means being able to see the source code. Most of them will not go looking for an official definition, and are likely to continue to be unaware that there is one. To me this is a problem, because using a term whose meaning is unclear leads to confusion, and diverts people into talking about side issues, which is exactly what the Sirsi/Dynix position paper has done.
Does terminology matter when we talk about software in the context of libraries? My answer to this ‘yes’, because the terminology we use says a great deal about our values and the importance we place on them. ‘Freedom’ is a concept that is often associated with libraries. For example, the American Library Association has an Office for Intellectual Freedom and published the first version of their Freedom to Read Statement in 1953; IFLA, the International Federation of Library Associations, endorses “the principles of freedom of access to information, ideas and works of imagination and freedom of expression embodied in Article 19 of the Universal Declaration of Human Rights” in its first core value. Thanks to the work of Richard Stallman and the Free Software Foundation, the concept of freedom has been extended to software, and it’s time for librarians to start recognising and promoting software freedom in addition to their more traditional freedoms. We can do this most effectively by using the term ‘free software’, rather than ‘open source’, when talking about software distributed under a license that allows users to view, change, and distribute the source code.
Why should we use the phrase ‘free software’ in order to emphasise software freedom? Because by doing so, we emphasise the benefits we gain from using software that is released under a license that preserves users’ freedom.
When people use software distributed under a free license, they are granted the four essential software freedoms:
Freedom 0, the freedom to run the program for any purpose
Freedom 1, the freedom to study the source code, and modify it to fit their needs better
Freedom 2, the freedom to redistribute copies of the original program
Freedom 3, the freedom to redistribute copies of their modifications
(from The Free Software Definition, http://www.fsf.org/licensing/essays/free-sw.html)
Taken together, these freedoms give software users the ability to be independent (freedoms 0 and 1), the ability to share (freedom 2), and the ability to participate in joint efforts to improve the software (freedom 3). It’s important to note that taking advantage of Freedoms 1 through 3 is optional—many free software users simply enjoy the benefit of Freedom 0, and never concern themselves with the other three freedoms.
However, I should also say that being independent does not prevent users from hiring outside expertise to help them install, learn, or even modify the programs on their behalf. Modern integrated library system software is complex, often consisting of hundreds of thousands of lines of source code, and it is likely that any library implementing a new (to them) system will require help from someone who is already familiar with the software. We are starting to see many examples of companies that get some or even all of their revenue from supporting free software—in the library world, this includes BibLibre, Bywater Solutions, Equinox, and PTFS, to name just a few, so a library interested in a free software ILS doesn’t need to rely completely on its own resources, or those of the community, unless it chooses to.
I can certainly see why people who are only interested in taking advantage of Freedom 0 might wonder why freedom matters. If their relationship with the software is completely passive, and they have no intention of examining the source code, making changes, or distributing it to others, which makes Freedoms 1 through 3 theoretical and irrelevant, does having free software make any difference? The answer is unquestionably ‘yes’; explaining why this is the case will also show why using the term ‘open source’ misses the point.
The best way to illustrate this is to consider phrases that mean the opposite of ‘open source’ and ’software freedom’. To me, and I suspect to most people, ‘closed source’ is the most natural phrase that is opposite to ‘open source’. If software users aren’t interested in access to their software’s source code, will they care if they are using open source or closed source? Probably not—because they will perceive no significant difference between open source software and its closed source counterpart. So using the term ‘open source’ tells people who aren’t interested in source code nothing useful, and it also leads people to evaluate software only on the basis of how well it fits their needs, not on what rights they get from the way it is licensed, and the long-term implications of those rights.
What is the opposite of software freedom? Some people might say software slavery, but I prefer to call it software imprisonment. By choosing to use non-free (or proprietary) software, users become metaphorical prisoners of their software vendor, and a common phrase for this is ‘vendor lock-in’ (which is nicely connected to the notion of imprisonment). This means that the vendor, and only the vendor, has the power to decide what features the software will have, which bugs to fix and which to call ‘features’, when to require users to upgrade to a newer version, and when to discontinue support for a software package. Some vendors may consult their user community when deciding which new features to implement in a new release, or to determine which bugs have the highest priority, but they are under no obligation to do so, and their customers who don’t like their decisions have no alternative but to accept them.
It is also possible for non-free software to have features that only the vendor knows about, and this has the potential to lead to inappropriate surveillance or loss of privacy. Since one component of an ILS is information about library users and their borrowing patterns, as a library user, I expect an iron-clad guarantee that data about me is being handled in accordance with my country’s privacy legislation. If the library can’t examine the source code, or commission someone else to do so on their behalf, how can they be sure this is the case? A report from a neutral third party, who has been able to examine the actual code the system is running, would be the best guarantee that library users’ rights are being protected.
How is free software different from proprietary software? Most importantly, it avoids vendor lock-in because its users have the four freedoms. All free software has a community of developers who are responsible for maintaining and enhancing the code–but in addition, anyone who uses free software has the option of making their own changes (Freedom 1), or commissioning someone else to do it for them (through a combination of freedoms 1 through 3), if they don’t want to rely on the developer community. The important word here is ‘option’: by deciding to use free software people have a choice of support models. In some cases, such as Koha, they also have a choice of support companies, all of whom have experience with the software. This is sometimes seen as a form of insurance, in that users of free software aren’t held hostage by a single company, and if the developer community disperses, another one can form, because the source code is available to all.
Some people argue that choosing proprietary software isn’t a problem as long as library decision makers understand that by licensing it, they are limiting their options for the future. My response is that, by failing to support software freedom, such people imply that imprisonment, in the form of vendor lock-in, is acceptable. Other people say that proprietary and free software should be considered even-handedly when evaluating new software, in order to be ‘objective’. In my view, this over-emphasises software functionality, and does not take into account the long-term implications of being locked-in to a single vendor. I find both of these positions sad and short-sighted, and would like to see more discussion of the benefits of software freedom in our professional literature.
I’m encouraged to note that at least two of the responses to the SirsiDynix position paper describe benefits of their use of ‘open source’ software that result from having software freedom, even though they don’t use the term. Mark Leggott, from the University of Prince Edward Island, describes their ability to combine software from several projects to meet their needs, by choosing the best components and integrating them locally. Cynthia Williamson describes being able to run the new (free software) and old (proprietary) systems alongside each other for an 8-month migration, which is also something that results from having software freedom for the new system and is unlikely to be possible if both systems were proprietary.
I’d like to challenge to the vendors of proprietary ILS software to release their code under a free license, because I see this as a win-win outcome. Their customers will regain their freedom, and at the same time the vendors will have access to a much wider pool of developer expertise (while still being able to control what goes into an official release). Since most libraries will not have staff with the skills to support the systems in-house, there will still be a large market for support, and much of this is still likely to be filled by the existing vendors, who, after all, have the most experience with their software. It may, of course, result in increased competition in the ILS support market, but this is likely to raise standards and result in better systems overall.
There is, however, another, perhaps more important, reason to prefer the term ‘free software’, since it will remind people that freedom is the key issue. Once people are aware of the importance of freedom, in particular the freedom to develop and enhance software on their own terms, their attention will also be drawn to the wide range of threats to this freedom. These include patents on ideas expressed in the form of an algorithm and implemented as software, extensions to copyright law, the use of proprietary file formats like Word (.doc), Excel (.xls), PowerPoint (.ppt), and Flash, and the use of DRM to restrict and monitor people’s use of digital books, articles, recordings, etc. The first clause of LIANZA’s Statement on Access to Information says that “Free circulation of information safeguards our democratic society.” To me this means that librarians have a responsiblity to acknowledge the importance of software and digital freedom, and that members of the profession need to question the limitations that proprietary formats, license agreements, and DRM place on the use of the digital information sources available in libraries much more than they currently do.
To conclude, I’d like to pose some broader questions for members of the library profession. Should libraries purchase ebooks that come with DRM, which limits what the user can do with the information? Should they purchase digital information sources that can only be used on certain types of devices? This is particularly true for digital audio ebooks, but it also applies to other types of information. Why are so few librarians visible in the Free Software Foundations’s campaign against DRM (http://www.defectivebydesign.org/amazon1984)? How are members of the profession promoting the free circulation of information if they don’t ensure that this is true not only in the print world, but also in the digital one? Finally, if members of the library profession, with its commitment to free access to information, don’t take action to preserve our own and our users’ digital freedoms, who will do it for us? What is the future of libraries if we don’t?