Bookshelf

Unplugged:

This week, speaking at a Gartner conference in Orlando, Florida, Microsoft ceo Steve Ballmer said some fascinating things about Linux and about Open Source software in general. And thanks to those remarks and the blinding realization they caused for me, I finally understand exactly why Microsoft doesn’t understand Open Source.

Ballmer asked, “Should there be a reason to believe that code that comes from a variety of people around the world would be higher-quality than from people who do it professionally? Why is its pedigree better than code done in a controlled fashion? I don’t get that. There is no road map for Linux, nobody who has his rear end on the line. We think it’s an advantage a commercial company can bring—we provide a road map, indemnify customers. They know where to send e-mail. None of that is true in the other world. So far, I think our model works pretty well,”

The model has worked well for Microsoft, that’s for sure.

At the core of Ballmer’s remarks is a fundamental misunderstanding not only of Open Source, but of software development as an art rather than as a business. Cutting to the bone of his remarks, he is saying that Microsoft developers, since they are employees, are more skilled and dedicated than Open Source developers. They are better, Ballmer suggests, because Microsoft developers have their rears (presumably their jobs) on the line. All those lines and all those rears are part of a road map, he says, and because of that road map the $30 billion plus Microsoft gets each year isn’t too much for us to pay, so the model works pretty well.

This is nonsense. It is nonsense because Steve Ballmer, like Bill Gates before him, confuses market success with technical merit. Microsoft’s product roadmap is a manifestation of a business plan, and what matters in Redmond is the plan, not the map, which is in constant flux. How many technical initiatives has Microsoft announced with fanfare and industry partners, yet never delivered? Dozens. That is no roadmap.

If Microsoft developers rampantly fail to produce good software, but the company exceeds earnings estimates anyway, how many of those rears will be actually on the line? Very few, and maybe none at all.

What Ballmer ought to have said was, “It’s true we have shipped some really bad software in the past and we are ashamed of that, but we are totally committed to improving.” But he didn’t say that. He said, “Our model works pretty well.”

I’m not sure that “pretty well” accurately describes software that is riddled with security holes exposing customers to tens of billions of dollars in lost productivity each year. And that part about indemnifying customers, indemnifying them against what? Certainly not against Microsoft software.

Against Ballmer’s glib insincerity we have Linus Torvalds, a very solid guy totally devoted to both the concept of creating powerful software and to giving it away. No wonder he is so misunderstood in Redmond.

The key to the success of Linux goes far beyond the price. Free is good, of course, but the true strength of Linux is the international movement to improve and extend it—the very “variety of people around the world” that Ballmer dismisses. What Ballmer sees as Open Source’s weakness is in fact its strength. Very few Open Source developers work full time at it. Most Open Source programmers are doing it a few hours here and there. Yet, here’s Linux, for example, a world class operating system, continually appearing in new versions and with new features. How can that be?

Linus attributes the high quality of Linux (it is very stable, certainly compared to Windows) to the very grass roots development effort that Ballmer criticizes and doesn’t understand. This would seem to contradict the idea many people have that it takes a high buck development operation to create great software. Just the opposite, says Linus, who claims that free software is nearly always better.

Huh?

“It’s very simple,” said Linus. “Because the software is free, there is no pressure to release it before it is really ready just to achieve some sales target. Every version of Linux is declared to be finished only when it is actually finished, which explains why it is so solid. The other reason why free software is better is because the personal reputation of the developer is attached to every release. If you are making something to give away to the world, something that represents to millions of users your philosophy of computing, you will always make it the very best product you can make. That’s the reason why Linux is a success.”

How can Microsoft compete with that argument? It’s hard, and the internal struggle to come up with a good response is evident in Ballmer’s remarks. They certainly won’t respond on price, since there is no way to undercut free. So we’re back to the usual campaign of fear, uncertainty and doubt.

Microsoft used to dismiss Linux as 1980s technology, which pretty much describes both Linux and Windows, it seems to me. Now they talk about “total cost of ownership” and find some way to make it look like using free software is more expensive in the long run than using software from Microsoft. Linux is certainly not free, but it is Microsoft’s tech support that has been compared to the Psychic Friends Network, not Red Hat’s or SuSE’s. Just because Microsoft has a big support operation doesn’t mean you’ll actually get a solution to your problem.

Linux scares Microsoft on several levels. There’s this business of giving the software away for free, which is totally confusing to Bill Gates—confusing and scary, since it undermines the entire basis of his fortune. But it’s the breadth of Linux and its potential on other platforms that also scares Microsoft. At a time when Microsoft is trying to be sure its software runs on all the handhelds, set-top boxes, mobile phones and any other new machine types that just might replace in our hearts the pc, versions of Linux compete on all those platforms, too.

These ideas are both obvious and old hat to people in or around the Open Source movement, but this column has a large non-technical readership that deserves to understand this, too.

When Ballmer talks about rears being on the line, what really counts at Microsoft is meeting shipping targets—meeting business goals—not quality targets. It is all about revenue. And there is nothing wrong with that if we all just say it out loud and admit the truth. But we don’t.

The lack of a roadmap for Open Source means that there can be dozens of similar projects, some being born, some dying, and others forking into new identities. Since there is no single architect for Open Source, these projects all have to compete for manpower and user interest. Most Open Source projects die, but when they do, it is a death Darwin would understand. Every death improves the software bloodline.

Against this, let’s look at how things are done at Microsoft. I’ll use as an example Universal Plug and Play, which was/is Microsoft’s initiative to make your refrigerator talk to your pc. upnp is not very successful, doesn’t work very well, has had many security problems and is in, shall we say, flux. The upnp consortium hasn’t issued a press release since 2002. But the consortium still claims 625 industry members, sucking-up a lot of time and money on technology that I would call a failure.

If upnp were an Open Source project it would be dead, which is preferable to its current state of limbo. But why is upnp so bad? Part of the answer comes down to difficulty of talking to refrigerators and stereo systems. It is tough to be clever in baby talk. But the rest of the answer, according to an ex-Microsoft developer friend of mine, is Redmond’s choice of personnel to work on the project. He said that with the notable exception of the upnp architect Yaron Golan, “the majority of the team were C-level players.”

If C-level players did an Open Source project, nobody would ever see it. But since this is Microsoft and since these were professional developers with their rears on the line, upnp limps on. And since Microsoft says whatever it does is the best it can be and deserves the notice and support of the entire world, we end up with a kludge that helps nobody, wastes resources, distracts customers, competitors and industry partners, and didn’t cost a single rear in the process.

I don’t think the model works well at all.

We aim to please.

Posted by cds at October 23, 2003 06:52 PM