In a recent article from Harvard Business School’s Working Knowledge, Sean Silverthorne, does some Q&A with Ramon Casadesus-Masanell and Pankaj Ghemawat about their research on the competition between Microsoft and Free and open source software (FOSS). It’s detailed and raises issues on FOSS distribution versus proprietary in relation to user adoption.

The article notes that “By lowering the price of Windows, the demand for Linux shrinks to the point where Linux is not a threat to the survival of Windows.” If I understood correctly, I don’t think that their study was intended to look at issues outside the scope of their economic model. Thus, the following is not criticism but rather some extra thought on the matter. I think there could be other issues that might make Linux a threat to the wide-spread survival of Windows.

For example, perhaps this is unlikely but if Windows exploits, viruses, etc. increased to the point where nobody could realistically use the operating system safely I would imagine totally different sorts of reasons compelling people to adopt Linux, namely privacy or safety concerns. Some time ago, I presented an article on how I thought a lean OS delivery strategy could really impact user adoption. I don’t see Microsoft able to do this, I think it is something that only a FOSS OS could accomplish because of the nature of the FOSS development/community/business models.

And what does the study discussed in the article illuminate?

The article discusses FOSS “demand-side learning” in which the development cycle is shorter because users have the opportunity to improve the software by modifying the code or contributing ideas. This may give the impression that FOSS, by virtue of increasing demand-side learning, would displace the position of proprietary software. However, the study’s authors note that their economic model does not show that to be the case.

They point out that “…the value of an operating system depends critically on the number of users, traditional software has an advantage…” that is, its usage is already spread far and wide. This first-mover advantage seems critical, according to their model, in what would prevent Linux from overtaking Windows.

In spite of demand-side learning, technically better software, and cost advantages, they note that without strategic buyers (such as governments or large organizations that choose to do wide-scale Linux rollouts). It’s not likely that Linux will overtake Windows. Because of some those factors, they also note that Microsoft ultimately gains from people copying and distributing the Windows OS, even when MS doesn’t receive payment for the copies.

I also thought their points on societal welfare were interesting, in that they find “…a monopoly of Linux is always preferable…to a Windows monopoly…” but that it’s “ambiguous whether a duopoly Linux-Windows is better than a Windows monopoly.” I don’t think this takes into account issues like the importance of freedom within the context of modern technical societies. I would like to know more about what they considered in societal welfare.

Finally, among their recommendations for Microsoft, if it wishes to remain competitive (and I guess “remain” is the correct word since we do see sizeable Linux increases in demand-side learning as well as the key strategic buyers they identified, taking action) is one that MS increase its demand-side learning. The thing is, how could Microsoft do that? The study recommends a number of methods. However, I don’t see a way for proprietary vendors such as Microsoft to seriously increase their demand-side learning enough to be competitive with FOSS communities, unless they themselves go FOSS.

Even though some sites have reported on this study from the perspective that Linux cannot overtake Windows, rather I think that so long as MS stays proprietary, the study points toward the direction of Linux.

——

Update: Dana Blankenthorn at ZDNet commented on this study as well. Dana brings up some other ideas that weren’t addressed in the study, like “…the idea that open source isn’t filled with clever entrepreneurs…” which I suppose heads toward something I was trying to show, that there are outside factors, which may be unexpected and could contribute to the competition in wholly different ways.

However, Dana seems to argue that the Harvard article is filled with conjecture and overreaching conclusions. I didn’t interpret it that way. I had the impression that the Harvard conclusions were drawn for only a very specific set of parameters as defined in their economic model. I don’t believe the conclusion was that “Microsoft will always beat open source.” in fact, while I saw that quote in Dana’s commentary I can’t find it in the original article.

Rather the original article seems to provide conditions that could lead to either side gaining or losing ground. The article says “Ultimately, the authors believe, neither side is likely to be forced from the battlefield” which is a rather different conclusion. The authors also hypothesized the following if MS set the price of Windows to zero “…Thus, we conjecture that even in this case, there would be people developing and using Linux…” and I don’t believe that conjecture helps support Bankenthorn’s interpretation of the article.

Why do I bring this up? Simply because I don’t think the original article was prophesying the future so much as examining what might happen under different conditions.

The Canadian Association for Open Source (Clue) published a thought provoking letter to an ITBusiness.ca article today. The Clue letter says that “…What is needed is for the government to separate the pricing and procurement of the source product from the various value-add services…” which is an interesting reflection for current musings about Public Works and Government Services Canada’s proposed potential changes to government procurement processes. It’s certainly feasible to evaluate these separate areas in a sophisticated way that still allows for a comprehensive decision.

The point here is that with Free and open source software, frequently one is at a loss trying to get a development community to respond to particular business issues when the issues are peripheral to the development of the software. The reason is not because the development community is basking under the dream of an intoxicating four-leaf clover high grown in artificial pleasure pods, but because it’s outside the scope of what they’re pursuing and toiling at, namely developing the software. Such business functions are often taken up by other service organizations (which are generally a part of the development community too) that do focus on implementing, supporting, customizing, etc. the software.

Why address the letter to an ITBusiness.ca article? I’m not entirely clear on that, however I do see the sense in linking the issues. Various other ITBusiness.ca articles report on the change in procurement processes as involving a decrease in the number of qualified suppliers, perceived increase of barriers to SMB providers, and introduction of methods such as electronic reverse auctions, which some people seem to be claiming would emphasize low initial costs at the expense better long-term purchase strategies. If I understand correctly, I think that what the Clue article proposes fits with what would work well for the channel partners of large vendors (whether they’re open source or not). It recognizes that channel partners provide valuable services, which risk being slashed from the procurement process (if the ITBusiness.ca articles’ various representative quotes are accepted). These services actually may be valuable toward saving taxpayers’ money in ways that could not be accounted for if the government focuses on purchase costs with only a few vendors directly. So these channel partners are essentially the equivalent to the general open source type of business, which is about providing value-added services around a typically, zero-cost product. They naturally share a goal here.

On the one hand the proposed procurement process change sounds like it may favour FOSS solutions because of upfront cost factors. But if these solutions cannot even be considered because the actual providers can’t get the opportunity to be part of the process, it’s moot.