An interview with Linus Torvalds

By way of an apology for the paucity of posts today, here’s an interview I did with Linus Torvalds earlier this year. In it, Linus talks about the future of proprietary software models, Microsoft, and, of course, open source software.


Open source programs have clearly made some great advances in the past few years. We’ve seen it with Linux and Mozilla. What sort of future do you foresee for open source? What do you foresee happening to software vendors tied to proprietary software models?

The power of open source really lies in various groups improving and building on an increasingly bigger existing base and slowly turning that base into commodity. One of the keywords here is "slowly" – it’s by its very nature pretty evolutionary, i.e. it takes time. It’s, in my opinion, also pretty unstoppable, but the process definitely makes it possible for proprietary vendors to generally take advantage of the open source commodity base, and continue to be proprietary "on top" of that base.

So I don’t think the proprietary vendors go away per se, I think they just end up moving higher up in the food chain. For most of them, that isn’t even necessarily very painful: Regardless of how commoditized the base is, commercial users will pretty clearly want to have support for it, so even with a totally open source technical product, the "higher up the food chain" thing will likely always exist in the form of services and support.

Also, even ignoring services and support, there’s always specialization: By its very nature, open source generally tends to fill the "basic niche", and I don’t see it as being in any way unlikely that proprietary vendors will fill specialized niches.

As open source software becomes more pervasive, what will happen to companies that have built their infrastructure around proprietary software?

Obviously, with open source becoming more pervasive, one of the things that happens is that the open source projects themselves tend to try to make that migration easier. Interoperability tends to be fairly high up on the requirements of any open source project, and you can see it in how open source projects have fairly actively attacked the issues of file format and network protocol compatibility.

But clearly a company that expects to be able to move over to open source infrastructure needs to have some flexibility of its own. But quite often it would tend to be sufficient to try to avoid the most obvious vendor lock-in: Any time there is a choice between a single-vendor solution and a more widely deployed and documented one, try to avoid the vendor solution.

Of course, that strategy tends to be a good idea even regardless of any open source issues.

Microsoft has spent so much time protecting the desktop that a myriad of other opportunities passed it by — search, for example. Open source software is cutting into Microsoft’s virtual monopoly of desktop software as governments and corporations become attracted to the cost savings they offer. Meanwhile, innovators like Apple and Google are becoming more prominent in the technology universe. What do you see happening to Microsoft in the years ahead?

This is exactly the kind of question I have a very hard time answering.

Microsoft really is a fairly interesting vendor in this space. Unlike most proprietary vendors, it’s one of the very few ones whose bread and butter comes directly from its commodity market, and even its
specialized offerings often sell because of its near dominance of a market that certainly looks to be commoditized over the next decade or so.

So it’s no wonder that Microsoft is one of the very few players who really don’t seem to like open source. Most other vendors seem to see open source as a platform that they can ride on, while to MS it’s a threat to how they do business.

That said, I don’t see the MS market going away very fast, and I don’t see why MS couldn’t continue to function as a software company even if they don’t control the commodity market any more. In many ways I think MS is in the same situation that IBM was in two decades ago, losing control of the basic market — and thus the dominance of the market — but not necessarily going away or even necessarily shrinking.

I think the really interesting question is what happens to their profit margins. It’s almost all profit for them right now. I don’t think that’s sustainable in any market, and yes, I believe that open source is one of the things that will "correct" the software market.

What will happen to the PC world if Microsoft’s market share should decline? What will the technology landscape look like?

If Microsoft loses its dominance, that’s likely a good thing for the market in general. Again, see the IBM connection from a few decades back: More open competition tends to make the market not only more lively, but also tends to grow it.

So the notion that many people seem to have that the PC market lives and dies by MS dominance would seem to have no basis in reality. If anything, near-monopolies tend to stifle things.

These days companies aren’t quite as dependent on Microsoft as they once were. Certainly Microsoft remains an important strategic partner for many companies. But the company is seeing its influence erode. What companies and technologies will overtake Microsoft should it continue to decline?

I do not believe that anything can "replace" Microsoft in the market that MS is right now. Instead, what I think happens is that markets mature, and as they mature and become commoditized, the kind of dominant player like MS just doesn’t happen any more. You don’t have another dominant player coming in and taking its place — to find a new dominant player you actually have to start looking at a totally different market altogether.

What that "next market" is, I have no clue. It doesn’t even have to be computers per se — to the next wave, computers are not necessarily what is being sold, they are just part of the required infrastructure for creating the next big thing in the first place.

In other words, I don’t think the PC "goes away." It’s an indispensable general tool for communication and computation. I just think it becomes some commodity thing you take for granted, like your groceries. You need them, every day, but you don’t think of them as controlling your life.

Do you think we might someday see Microsoft becoming something more along the lines of a utility, a venture that provides basic services largely taken for granted by those who purchase them?

I think that is how they have always seen themselves; it’s just that they much prefer the "no competition" landscape where they can have high profit margins. And yes, I think the big difference 10 years from now is not that MS is gone or even necessarily does anything very different, but that they have profit margins in line with the rest of the industry.

Of course, it’s likely that they’ll have to operate very differently then, but that’s more of an internal operations thing than anything external in the market.

What will happen if Microsoft’s market share doesn’t decline? What will the tech business landscape look like in a world in which Microsoft not only retains its current position, but extends it to other realms — the living room, financial services, etc.? What does the "Microsoft forever" world look like to you?

I just don’t believe in dynasties. Things erode over time. Successes start to take themselves for granted, and the successful companies aren’t nimble and hungry enough any more.

In the tech market in particular, companies just don’t tend to stay on top forever — they become irrelevant either because of their own missteps or because their market just isn’t the "happening thing" any more. You can only skate the cutting edge for so long.

So the question is how the decline happens, and in what timeframe. Will open source be a factor? Almost certainly. Will it be the factor? I don’t know.


Share this Post