Of VARs, Linux And Guinness

If you want to get an interview with Mr. Linux himself--Linus Torvalds--you need just two things: steak tartare and Guinness beer. Torvalds, the 29-year-old Finnish developer of the Linux operating system (OS), wolfed down the raw, ground beef delicacy and washed it down with a pint of fine Irish stout as he fielded questions from VARBusiness investigative editor Lawrence Aragon and software editor Matthew A. De Bellis. The laid-back "programmer" opined on a number of topics:

VARBusiness: Are you surprised by the momentum behind Linux right now?

TORVALDS: To some degree, yes, but at the same time, it wasn't entirely unexpected. More than two years ago, really, there were a lot of companies that were looking at Linux. But one of the largest things holding people back was the fact that nobody else was using Linux. A lot of companies that were kind of on the brink decided to do it, and now it's snowballing. It's kind of like critical mass, some kind of eruption.

VB: What's your market goal for Linux?

TORVALDS: I don't care if Linux holds 90 percent of the market. What I want is when some random person walks into a computer store, he at least has to spend some time making a decision. Does he want to buy Windows; does he want to buy Linux; does he want to buy Mac? Right now, nobody does that. There's very little real choice in the market.

VB: Why is that important to you?

TORVALDS: Partly because one of the reasons Linux came about is that I knew I didn't want to have Windows. This was eight years ago, and Windows was even worse. I knew it wouldn't cut it. It still doesn't cut it. But I didn't have many alternatives. I had the commercial alternative [Unix], which was way too expensive. It was really out of reach for a normal human being, and not only out of reach in a monetary sense, but because years ago commercial Unix vendors weren't interested in selling to individuals. They were interested in selling to large corporations and banks. So for a normal person, there was no choice. You could go for a Mac, but for a programmer it wasn't an option.

VB: How do you see VARs fitting into the Linux market?

TORVALDS: I haven't actually talked to very many of them, and I haven't talked to any lately, but I remember that it wasn't that long ago, maybe two years, that I spoke to somebody who said he didn't choose Linux because he just hated having to explain [what it was] to his customers. That may not be true anymore. Today, you probably don't have to explain it ... I think what changed is that if a VAR now says, "We have this new product, and we can go with the Linux-based version instead," the people he's talking to--the MIS manager or CFO--have at least heard about Linux. So [a VAR customer] doesn't go, "What?" He says, "Oh, Linux," and tries to look intelligent. I think just the fact that people have heard about Linux makes a big difference to a systems rep.

VB: A number of VARs are doing quite a bit with Linux, but there are others who aren't exploring it, particularly NT VARs. What's your advice to them?

TORVALDS: Well, to a lot of VARs, the problem may be that they don't have any Unix knowledge at all in-house. It's the Unix VARs that have been really excited about Linux because they have the Unix knowledge and hated the licensing terms and support. And they are probably the ones that are actually doing Linux today. For somebody who is coming from a Windows environment moving to Linux, even for a technical person, you have to learn. Some people just aren't willing to do that. And I can say as much as I can [about the benefits of Linux] until I turn blue in the face but they won't listen.

VB: Why should VARs resell Linux?

TORVALDS: If you're a VAR and you're at least open to considering the idea, one reason to go for Linux is that even though the profit margins aren't that big, not paying a licensing fee is an issue. The other thing is, it's good for a lot of boxes that sit in corners. The most important thing is that it's stable and, like it or not, even Windows NT is not stable [in every configuration] ... Another issue is that [Microsoft] doesn't want to sell you the operating system. It wants you to lease it, and there are a lot of customers who do not want to lease what they get. For VARs, that can be really, really painful because they may have customers who are ready to pay a large amount of money up front, but [they balk because] a lot of customers don't like the notion of knowing that they'll be paying forever. Look into Linux so you don't get caught with your pants down.

VB: What's another selling point for VARs?

TORVALDS: A lot of people hate updating their entire systems because they just need a small part of it that they need to get fixed. With Linux, you can upgrade the kernel without having to operate all the shared libraries, all the DLLs. Usually, it's not actually a kernel problem. You can go the other way: You can upgrade one or two libraries without having to upgrade the kernel or the application.

VB: Does that give Linux an edge over Windows?

TORVALDS: As a VAR you want your customers to upgrade. You want them coming back to you. So the fact that they upgrade--if they want to upgrade--that's good for you, I would expect. But you don't want them to have to upgrade internally and you don't want to have to force them to upgrade. You can send them a chain letter every six months and say, "Ok, we have a new version out," and they're so used to upgrading that they'll just do it.

VB: There's been talk that Microsoft Office will port to Linux. Do you believe it?

TORVALDS: That's pure rumor. Hey, for all I know, maybe it's true. I suspect that if it's happening, then it's someone inside Microsoft doing it just to see how easy it is to do. But if it actually ever happens I'd be really happy because that means I won.

VB: When will we see a critical mass of desktop applications?

TORVALDS: I'd say realistically within three years, but I think that in '99 you will start seeing some of it happening. The desktop market is the hardest by far, for lots of reasons. One major reason is the desktop is also the place where you have the most variation in applications, so that's where the advantage of Windows is the greatest, because it can handle a lot of applications. At the same time, I just think Microsoft has been doing a great job with applications, and it's certainly doing a great job on marketing. But I think it's done a bad enough job on some basic technical issues [to give Linux an opening].

VB: What are some of those technical issues?

TORVALDS: [Microsoft] started from the wrong kind of platform, so it accrued a lot of crap early on. It probably could get rid of a lot of that in [Windows 2000], but if you've actually followed what happened to Windows NT, it's actually moving more into the same old crap because it notices that it needs to run the same games [and other applications] on Windows 98.

VB: You recently said that Linux does not compete directly with Microsoft or Windows. Is that the case?

TORVALDS: What I meant to say is that it kind of competes in a different market space. It competes for the same people. That's obvious. Especially Windows NT and Linux head on. And I'm trying to make it head on for Windows 98, too. But it competes in a different kind of way--for the same people that is. Microsoft has been very successful in marketing out the competition. Either you buy them out or you just flood them out. The way Linux has developed and been released is just so different from a normal kind of commercial market space that it competes on its own terms instead of competing on Microsoft's turf. And that's a big advantage. You don't want to give Microsoft the choice.

VB: What are some market segments you think are ripe for VARs to target?

TORVALDS: Right now, from what I've seen, you probably want to target ISPs, where networking is a large issue, you have a fairly technical kind of support structure and price is important. But what I think is kind of interesting, and I haven't seen very much of, is point of sale (POS) systems. You might install 20 POS systems and have them connected all the time to a server.

VB: With regard to POS, do you think Linux can kick-start the network computer market?

TORVALDS: I'm not actually a big believer in the thin client.

VB: Why not?

TORVALDS: I think most of the time it makes a lot of sense to be centralized, but you want to have at least some intelligence in the client because if the network goes down, you still want the client to be functional.

VB: So you don't believe Sun Microsystems CEO Scott McNealy's statement that the network is the computer?

TORVALDS: No. I believe in networks but I believe in networks as a way of decentralization while Scott McNealy seems to think of networks as a way of centralization. And I don't believe in centralization. I believe that you want to have central management because otherwise it's just too hard, but you don't want to have central processing. ... You don't want to limit what you're doing in some arbitrary way. You do it on a per-case basis, and that argues against thin client.

VB: Some VARs are worried that Red Hat Software will be the Microsoft of Linux. Are you worried or pleased that Red Hat is getting so much support from all those big shots?

TORVALDS: I'm not worried about it. It's good to have a focal point, because it concentrates people and makes it easier to talk about Linux. ... I'm not really worried because I can certainly see them being big, but I don't think it's ever going to be much of a worry that you'll have the Linux market just completely owned by one company, because of the whole copyright issue. ... I think that even if Red Hat would be able to obtain Microsoft status, it wouldn't hold onto it very long because people would have been aware of how it got there.

VB: What are the major challenges that Linux still needs to overcome?

TORVALDS: I'd like to see a company that has nice success selling Linux applications. Or a VAR that just makes a killing selling 10,000 machines to the U.S. Navy or whatever. Those kinds of success stories have been missing. There have been a lot of fairly successful sales already, but even just six months ago when people bought into Linux, large companies said you cannot use this as marketing information.

VB: Why was that?

TORVALDS: It was a perception issue. It was horrible. They would say, "It's working really well for us but don't tell anybody, because what would a stockholder say when they learned that we're using a free OS?" That means that we still haven't broken through that kind of barrier where we can have a few really good success stories that just makes people say, "Hey, this is actually a really nice way of making money."