GNU/Linux vs Microsoft on Research & Development: What's the Truth?

LXer Feature: 13-Jan-06 Linux News raises questions about Microsoft's alleged $6 billion per year investment in R&D. Is it money well spent, or are they just throwing it away? Digg Story

A recent Microsoft Certified Professional Magazine article, Winning the Linux Wars, suggested that Microsoft invests more than $6 billion (USD) in research and development. Rich Freeman, the article's author, reveals some of the arguments Microsoft partners make in trying to keep customers from jumping the Microsoft ship. One of the alleged selling points is the investment in R&D that Microsoft makes. That raises a couple of questions. Just how does Microsoft's R&D investment stack up against the libre software community's investment in R&D? Is Microsoft's investment well spent, or simply a waste?

Freeman quotes John Hollinger, an executive for a Microsoft Gold Certified Partner on Microsoft's investment in R&D. "Microsoft invests north of $6 billion a year on R&D. There is nobody in the Linux world" that does that, he says. "Whose economic best interest is it in to invest the extra $10 million or $50 million over some period of time to make sure that's a secure computing platform?" This statement has at least two major problems. The first lies in the notion that no one in the GNU/Linux world invests $6 billion a year in R&D. The second lies in the fact that, for all its investment in R&D, Microsoft's computing platform is no more secure than a wet paper bag.

Some in the technology community laugh when hearing about Microsoft's investment in R&D. As one person sarcastically noted, "They spend it on research and development for sure... Researching the market, and developing their marketing ploys, researching their competitors patent portfolio, and developing their own.... The term 'Research and Development' can mean anything you want it to (a bit like statistics, really)." I also question whether customers really are gullible enough to accept "we invest lots of money in R&D" as a selling point for software. I would hope that most IT decision makers are intelligent enough to see through that. However, I think it's worthwhile examining the argument presented, since the article claims this is a selling point. Personally, I doubt it. Still, they fired the volley.

First, comparing the R&D investment for the GNU/Linux community to that of Microsoft is absurd and likely an impossible task to accomplish. Microsoft is a single company with a range of software in which it must invest. That is, they not only invest in Windows, they also invest in Office, SQL Server, IIS, and games. They obviously don't invest too heavily in Internet Explorer, or in security for that matter. The GNU/Linux community, on the other hand, has a variety of different companies supporting the development of the Linux kernel. The R&D for the rest of the platform is handled by the individual project teams and distribution vendors. Obviously, while IBM, Novell, Red Hat and others contribute to the development of the Linux kernel, they also contribute a great deal to their own distributions and projects.

What makes the comparison so absurd is that, with Red Hat, Novell, Mandriva, Mepis, Linspire, Xandros, and so many variations of the same core system, how do we quantify the overall investment in R&D? There certainly is no need for me to spend $6 billion if I want to start my own new GNU system. I wonder how much money Mepis invested in R&D? Yet, they have a solid, stable distribution that is growing in popularity. In other words, no single vendor actually needs to invest $6 billion in the kernel and core utilities. Now throw in the respective investments of time and money into projects like Apache, Firefox, GNOME, KDE, and OpenOffice.org.

The problem lies in understanding the difference in the two economies. Microsoft can pull up the pay records and equipment costs for its R&D teams. So, too, can IBM, Novell, Sun, and others. However, does the KDE project track it's investment in R&D on the same terms? What about OpenOffice.org? Just because a libre software developer doesn't account for their investment in the R&D column of the accounting spreadsheet doesn't mean that no money was invested in R&D. Libre software projects count their investments in R&D - and the return on their investments - in rather different terms. Translated into dollars and cents, one might well conclude that the value of R&D in the libre software community far exceeds even Microsoft's wildest dreams.

What of the investments of libre software developers in circuit boards for devices they create to run libre software? Calculate the cost of a libre software developer's time in experimenting with variations in the user interface, or a given library. Calculate the cost of time spend reviewing the users' feedback on your experiments, and then making adjustments where necessary. Just because libre software developers don't always have a balance sheet to show where they spent their time and money doesn't mean they didn't invest heavily in R&D. Again, bring these investments of time and money from thousands of projects into a single platform, and one can begin to see the value of these investments.

The members of The Freely Project have been working on a software solution stack that churches and other ministries can deploy. The project is defining the software that churches need, and then will establish a repository for the appropriate applications. Ubuntu is the primary OS. The Freely Project's investment in R&D is minimal. The application stack consists of existing applications, so the only real investment in R&D needed (by Freely) is in defining the needs of a variety of churches. Once the repository is made available, churches will be able to download and install Ubuntu Linux, then add the repository to the list, and download the applications they need and want. Thus, by capitalizing on the investments in R&D from the numerous existing projects, The Freely Project can empower churches around the globe to migrate to a platform that enhances their stability and security and reduces their overall computing costs.

The Freely Project is taking advantage of a collective R&D investment that probably amounts to millions of dollars, even though its own investment is likely a few hundred dollars. The expected ROI is the ability of churches to enhance their technological capabilities while reducing costs, thus allowing them to re-allocate funds to the church's main mission - ministering to people. It's true that The Freely Project is not a corporate business entity. Still, the same lesson applies to businesses. By capitalizing on the current investments in R&D, no single business needs to make a heavy investment of their own. They can, and may choose to do so anyway, in order to reap the returns from a new technology. This is what the major players in the libre software community are doing. The whole economy benefits in the process.

Just as the investments in R&D are measured differently, so too, are the expected returns. A developer who contributes a security patch for Apache or an e-mail program calculates the return on investment, not in terms of dollars and cents, but in terms of the reduced impact of viruses on the global internet. Likewise, what would be the return on investment in the $100 laptop? That cannot be measured in mere dollars and cents. I doubt $6 billion was invested in the R&D, yet the overall impact will be far greater than that. There is also the emotional return - simply knowing that users use and appreciate your software. Calculate that.

Let's turn our attention now to that $10-$50 million Hollinger said Microsoft invested to ensure a secure computing platform. There is no doubt that it is in Microsoft's economic best interest to ensure their computing platform is secure. In light of the recent WMF fiasco, one might wonder just where that $10-$50 million was invested - the security of the code, or the new alarm system for the OS development team's offices? Perhaps it's going to the marketing literature about their security improvements. It may be true that Microsoft is improving the security of Windows. Yet the nature and number of vulnerabilities suggest that they should consider shifting their R&D budget.

Microsoft's approach to organizing and securing the filesystem doesn't really show much for their R&D investment. Windows 95/98 were pathetic when compared to GNU/Linux. And Windows NT/2000/XP isn't much better. GNU/Linux still does a much better job of separating users and their settings. Also, Windows is world famous for its unchecked buffers and over-reliance on remote procedure calls. The combination of ActiveX and Internet Explorer is viewed by many as an open invitation to attackers. Some might even ask whether that money was invested in researching how to open up Windows to attacks.

Microsoft once kept the graphical desktop, Windows 3.11, separate from DOS, the underlying operating system. With Windows 95 Microsoft integrated the graphical desktop into the underlying OS, going against the common wisdom of keeping the graphical desktop separate from the underlying OS. They did so in an effort to increase the system performance. The problem is that Windows 95 was more crash prone than Windows 3.11. It is interesting to note that the GNU/Linux world never made this sacrifice, and every GNU/Linux OS I've used loads quite a bit faster than Windows 98 or XP on my boxes.

If you're an IT executive or manager listening to the Microsoft Platinum/Gold/Silver/Bronze or whatever Certified partners talking about Microsoft's investment in R&D, you really ought to ask some questions. Ask them why, with a supposed investment of $6 billion per year, they can't catch more security vulnerabilities. Will that money not pay for enough security experts? Ask them why, with that kind of money invested in R&D, they have to integrate a web browser into the core OS. What Research proved that to be a wise technical decision? (In reality, it was a marketing decision.) In other words, they don't even seem to be throwing money at the problem. They're just throwing it away.

If you don't ask questions about Microsoft's R&D, and just blindly accept that spending $6 billion per year in R&D is money well spent, then just prepare to accept the consequences when your network is hacked and your data is exploited all because someone thought that integrating a browser into an OS was a technically sound decision. If you don't think that the libre software community can match Microsoft's expenditures in R&D, it's because you don't understand how R&D is measured in the community. And if you don't understand that Microsoft's marketing and political machine is better developed than their software, then you'll keep paying thousands of dollars for a wet paper bag when you could have a real computing platform to show for your investment.