Archive for March 25th, 2009

I respectfully disagree.

Microsoft has recently published a white paper on Microsoft and OSS, called “Participation in a world of choice”. It’s a nice 16-pages pdf, with kind words on the role of OSS in the modern IT landscape, the fact that  “It is important to acknowledge that the relationship between open source and Microsoft has at times been characterized by strong emotions and harsh words” (some of which were from Microsofties themselves, like the infamous “cancer”), and that the future will be nice and warm and fuzzy. Despite the fact that – for probably the first time – a white paper contains a significant bibliography that includes academic papers, I have to say that I am not impressed by the content, that basically tries to reposition free software and open source in a context that is not entirely appropriate, and selectively presents a view of the market that is not in my opinion accurate.

(A note of warning: this post is not written out of “hate” of Microsoft. I am old enough to remember when IBM was the death empire (and its legal team was called “the nazguls” because they were unstoppable, fearless and devoid of human emotions), when Sun claimed that linux was a “bathub of code, … and sometimes what floats in it is not pleasant”, and at the same time I recognize the great advances that both companies made  in recent years in open source. It would be my greatest pleasure to see Microsoft participate in OSS fully in the same way, but as it was for the companies I mentioned before, they have to prove themselves- and be slightly less schizophrenic in their message on OSS. End of note.)

So, I would like to point out a few things that I found in the paper:

  • page 1: “… and increasing opportunities for developers to learn and create by combining community-oriented open source with traditional commercial approaches to software development.” First of all, not all open source is community oriented; here, Microsoft implies that all open source is developed in a community-oriented way (something that is not true, for example for dual-licensed software that has a much lower external participation rate, or for vendor-dominated consortia). Then, the use of “commercial” is wrong: free software and open source can be as commercial as software from the proprietary world. I would say that the copywriter here tried to suggest that OSS is “noncommercial” (something that our study , while the reality is that FLOSS in “nonproprietary”.
  • Then we have the table in page 4:

    What the present is a very limited view both of the constituencies, and the motivations for using OSS.There are many studies that present the reason for developers to participate in OSS; among them, ethical reasons, practical reasons (what is called here “scratching an itch”), self improvement, “signaling” (demonstrating capabilities in development to increase employment opportunities), and roughly 50% of the developers are paid to work on OSS. Assuming that transient reasons are the real motivations of OSS development provides a negative view (because as those personal interests may be short-lived, the software itself may be short-lived too).

  • In the same table: the author places “IT Administrators, End Users, Entrepreneurs” in a single cathegory. Ok- this is not a research article, but even a novice would probably find differences among them. If the tags identify endogenous use, then having as a single advantage “ease of acquisition” is clearly a way to downplay the additional advantages of OSS. Among them, the reduced lock-in (that casually is never mentioned in the document), or reduced costs (as mentioned previously in my work on “OSS myths“, for example in the surveys by CIOinsight, or our work in the COSPA project). If the mention is for companies that use OSS inside their software (both for internal use or for commercialization) there are other demonstrations that OSS has substantial economic advantages. And where are the OSS companies? Those cannot be included in the last line, about “established ICT vendors”… They even mention themselves an example that does not fit in the table: “Apple shifted to what has been called an “embrace and layer” strategy for its consumer operating system by leveraging permissively licensed open source BSD code for functionality such as networking infrastructure, while focusing its commercial R&D on building a proprietary graphical user interface (GUI) “on top,” and licensing the resulting product as a whole under a traditional commercial license.” (In my classification, that would fit within the “R&D cost sharing”)
  • Page 5: ” Another industry analyst firm, the 451 Group,  identified more than 100 ICT companies who rely  on OSS to generate a significant portion of their revenue. At the same time, it found “the majority of open source vendors utilize some form of commercial licensing to distribute, or generate revenue from open source software”. Of course! The error is the same already mentioned, that is the confusion between “proprietary” and “commercial”.
  • Page 7: “OSS approaches tend to be relatively more successful when the end users of a technology are themselves developers, as opposed to nontechnical end users”. The phrase is incorrect, and arise from the identification of OSS developers as volunteers that “scratch an itch”. From a logical point of view, there are two errors: first of all, there are many technical users that are not developers (system administrators are a good example). In fact, the very high penetration of OSS in server environments is not strictly related to developer participation. Second, the assumption that OSS is inherently difficult to use (implied in the phrase) is easily dispelled by the great success of FireFox and OpenOffice (both of which require no “developer” in sight).
  • Page 7: “Windows Server product strategy continues to focus on offering a product that IT administrators will choose over alternatives (including Linux), because it is highly manageable with readily available skills, supported by a wide range of third-party applications, and offers the lowest total cost of ownership (TCO).” This is marketing, not research- first of all, the TCO debate is still not solved in favor of Microsoft (and after reviewing the TCO numbers for COSPA, I suspect that that would not be an easy win for them). Then, it implies that OSS skills are not readily available (again, something that is unproven) and it implies that OSS alternatives have a limited range of third-party applications (look at RedHat certified applications list for a good counterexample).
  • Page 8: “For developers, the entire .NET Framework is available as a reference source to enable them to debug against the source code”. And since it is not open source, this should probably not be mentioned here.
  • Page 9: “One key, supporting principle is respect for the diverse—and continually evolving—ways that individuals and companies choose to build and market what they create. No efficient, effective technical solution should be precluded or advantaged because an individual, a vendor, or a development community has chosen a particular business model—whether based on software licensing, service and support, advertising, or, increasingly, some combination thereof.” This is aimed squarely at those governments that are trying to estabilish pro-OSS policies, and ignores the fact that in many cases the inherent market situation (with a de-facto monopoly) is not a balanced market in itself. Recently it was found that “Software tenders by European public administration often may not comply with EU regulations, illegally favouring proprietary applications”; so the advantage is at the moment squarely for proprietary software vendor, and the recent guidelines from the EU are designed to provide a more balanced market. My own suggestion is to evaluate the whole cost of an IT adoption using metrics that cover the full lifetime of an application; my favourite is the German WiBe model.
  • Page 9: “A second key principle is a balance that preserves constructive competition and healthy incentives: when individuals and companies are rewarded for creative differentiation, customers benefit from a dynamic marketplace that offers more product choices. Incentives for commercial investment in new innovation should coexist and coevolve alongside practical mechanisms for sharing intellectual property (IP)—with the overarching focus on a dynamic industry that continues to bring great ideas to customers.” And this is for those that are asking for the abolition or the reduction in scope of software patents, and the invocation of open standards that are non-IPR encumbered. I already wrote in the past against software patents, and I believe to be not alone in claiming that the hypothetical advantage of software patents seems to pale in comparison to the extensive damage that it is causing to the industry. (And no- claiming that the thriving software ecosystem that we see now happens because of software patents should be rephrased in “the thriving software ecosystem that we see now happens despite of software patents”)
  • Page 10: “We have long sought to contribute to the growth of an open ecosystem, whether through publicly documenting thousands of application programming interfaces (APIs)”. It took the European Commission, and a record fine for violation of EU monopoly laws, to abtain the release of crucial APIs…
  • Page 10: “More than 500 commercial IP agreements with companies from a wide range of industries- including companies building their businesses around OSS”. Again: I would not claim that, after the substantial brouhaha in the Novell patent pact (considering that it does not covers things like OpenOffice…)
  • Page 10: ” And we have stated broad openness to noncommercial OSS development through the Patent  Pledge for Open Source Developers.” Nice- despite the fact that “noncommercial OSS” is an oxymoron, as placing additional redistribution limits make it non-OSS.

The paper ends with a very promising phrase: “We recognize that in the future, Microsoft ’s relationship with OSS may be punctuated by strong emotions and the possibility of interests that at times will be in conflict. But we are profoundly optimistic … will surface new opportunities for Microsoft and open source to “grow together” in purposeful and complementary ways.

I would be very happy to have Microsoft as a good OSS citizen, even with the recognition that their path may sometimes conflict (this is true also of other “fellows” like IBM or Sun, as well). But I would start with a more balanced introduction, or at least one that has not such a significant percentage of “hidden” ideas. Openness is before everything else in the mind; if your ideas are strong enough they will survive and thrive.