[LWN Logo]

Linux Weekly News
Daily updates
Linux Stocks Page
Penguin Gallery
Book Reviews

Contact us

December 15, 1999

Bob Young's Bazaar Keynote

Intro by Mad Dog Hall

Bob Young has set the course in Red Hat by their philosophy, the fact that they GPL their own internally-developed software, and by providing good sites to support free download of their software.

[The crowd is smaller than for yesterday's keynote by Richard Stallman, but everyone seems pretty friendly.]

Bob started the talk by mentioning that it was his usual high tech multimedia talk, meaning no media at all. He stopped by Stetson's on 5th avenue to look at hats. "We're in the hat business, not the software business, since we give away our software."

He tries on a couple of hats, beret and soft felt. People seem to like the beret better, so he put that on.

When asked if he was going to talk about the IPO, he led off with:

Going public gives you the resources to grow and look after your customers. One of our key contributions is that we give away all of the software we right under open source licenses. That's our biggest contribution to this revolution and it continues to get larger as we hire more engineers to develop more software.

Speaking to that, let me tell you a story from my early days in New York. For my business, Vernon Computer Rentals, I started going to Unix users groups and started writing a newsletter for them. I Asked them what type of article they wanted to hear that they didn't already get from Infoworld, etc. The answer was all about free software. "From engineers according to their skill to engineers according to their need." I was skeptical of the model, the "berlin wall falling". I could see that it gave a unique benefit to the user, but I was missing the economic model to make it work for business.

When I dug into it, I found three communities of developers of free software. The engineers that built it and two bigger pieces: the engineers that build tools that they need for their companies and the commercial companies that were already contributing. The X Windows system is 110 MB and it comes from the X consortium, Sun, DEC, HP, etc., who need a common interface from their systems, couldn't trust each other, so they published under an open source license. A barter system based on open source.

Launched a catalog called PC Unix/Linux catalog. Bob, if the software is free, how are you going to make money at this? Trust me, our customers love it and we'll work this out. Early on, we found we were not in the software business. Hooked up with Mark Ewing where we take parts from a worldwide set of engineers and give anything we develop back to them. Totally different from the model of binary control of the software. What business are we in? We looked around, are we in the legal business? If a lawyer wins a case in court, another lawyer can use his words as soon as they are out of his mouth.

We're making a commodity. How do you make a name in the commodities in the market? You build a brand name. Here is the Tomato Ketchup story: Why does Heinz have the largest share of the market? Some people say it tastes better. However, test it in the third world and you'll find out, no one really likes ketchup and they like all ketchups differently. We've defined the concept of ketchup so well according to the parameters of what Heinz is, to the point that ketchup that doesn't come out of the bottle is better than ketchup that does.

The open source model would give us the best technology, unique benefits to users. If we built a strong brand around the product, we had a saleable commodity. The bulk of the work only starts when you install, moves on from there. IBM is famous for their products, but they effectively give away their products and make their money on services and support, making those products useful for their customers.

We give away the software in a shrink-wrapped box that contains free software and some services.

Turn it over to Q&A

  • Are you doing a Red Hat Linux for Netwinder?

    That deal ended a while ago, based on them selling a large number of units, which never happened. We don't have a commercial agreement with Corel.

  • Concerning patents and free software. Looking at a tracking system on free software and patents. Is this true?

    Just to be very clear on my response, on a personal basis. Red Hat is not ideological, they exist for their customers and their shareholders, Individuals within a company have ideology, are there services that we can provide to our customers based on these ideologies. I'm fascinated by the problems created by the US patent office by granting patents for everything including the kitchen sink. Patents on code, in my personal belief, are a bad thing. If you can't patent a Shakespear sonnet, how can you patent code? We need a whole conference to talk about this.

  • Don't think the patent issue is ideological, it is also commercial.

    Agreed, but the creation of a database to help us defend against free software infringement on patents is not commercial. We haven't done anything in that area, I must confess.

  • You used to say "free software" instead of "open source", now including more commercial software. How do you support the free software community.

    I take pride in being ambidextrous, I can say both free and open source and tend to use them interchangeably, understanding that there is an ideological difference between the two.

    Our whole success, our opportunity to build a successful business, hinges on the value proposition of giving our customers the control over the technology they are using. That is what gives us power over our competitors. We aren't going to dictate to the market. We will make our customers more successful. We will assist Oracle so that our customers can run Oracle on Linux rather than Oracle on a proprietary platform. We see it as a necessary step, to do that for their business and believe that the model is so powerful, that someday when people draw up a list of features they need, one of them will be, "is it free software". First, we have to prove to the customers that this is a good thing.

    Red Hat does not ship proprietary software that we develop as part of our products.

  • I have trouble believing that the model for a software company that is out there today is really going to vanish, that people should focus on being marketers and organizers instead of developing software.

    I gave a talk in Netherlands last year at a Unix conference, 50 year old Unix programmers with IQs two or three times mine. My talk was going to be, "How Linux solves the fatal flaw of Linux". That was a little arrogant. I was writing my resume, I've been working in this industry for 26 years. I've been here for 50 percent of the life of the industry. What we are doing is not that radical. Looking at what is being done as the way things should be done when it only started in 1965. Before that, the software was shipped for free with a product so that their customers could help them support their code. They changed the model when they got large enough to make more money by making their customers more dependent on them. It is only 30 years old and it has a fatal flaw.

    I'm called at Red Hat the "Red Hat spokesmodel". It is, in fact, my full time job to take credit for the work that everyone else does. In this particular case, when you look at the proprietary software model, intellectuals come up and ask us how the lessons learned in the open source industry to other arenas? In my mind, the lessons are already applied. If you build a bridge, everyone can see it. The engineering plans become part of our common heritage and we build bridges based on the ones that stood up, not the ones that fell down. In the CS industry, we're not allowed to build on what has been done before. It deprives customers of the ability to control their technology they have purchased. They sign away their rights on software that their very lives may depend on.

    This really looks like an inevitability, we are bringing the software industry into the 21st century, the model under which most systems work.

    Five years ago, I couldn't equate the hype about the Internet with reality. Today, I don't write letters on paper, I do my shopping on the Internet, etc.

    It isn't defined by any of us, but a new model is being developed and I'm confident the days are numbered for the old model.

  • Questions about what should people do?

    I know I'm going to get a thrashing about this no matter what I say. Taking my Red Hat off. With my Red Hat on, if you work for Compaq, get to me afterward and I'll go buy a beer and tell you the real answer.

    With my Red Hat off, do what's best for your customers. VALinux' approach is different from Dell or Compaq. I know i'm not clever enough to figure out which approach is right but I know the market will figure it out.

    I'm going to duck the question because I know the answers that are in the best interests of Red Hat, but my friends at VALInux and elsewhere are in the audience, so I'm not going to go there.

  • Comparisons between Linux and the highend servers like Solaris. How long before Linux runs on the high-end servers?

    The story goes back to Unix Expo 1995, here in the JJ center, when the Unix engineers had heard of free software and were interested in it, but the MIS directors ran the other way. They didn't understand from it and ran from it like the plague.

    We had launched our 2.0 product and we were hoping to sell a lot of products at the show. I'm a sales guy, the doors open, a guy with pinstripes slows down to approach our booth, planning on passing it by. Grabbed him, talking to him. He's from Merrill Lynch, "being a federally chartered bank, we have adopted a policy of knowing where every line of code has come." That's one of the best excuses I've heard for not adopting a better technology. Over the course of conference, two or three people from Merrill Lynch buy copies, they have dozens of systems running Linux. I tell them my first conversation, they say, "Oh damn, you were talking to an MIS director. If you tell him what we said, we'll come back and kill you."

    IDC numbers, 15% of new operating systems last year were Linux based in 1998. This wasn't so new, but the users became brave enough to admit that they are using it.

    The biggest challenge is that MIS directors are highly conservative because their mission is not to screw it up. They are running million dollar systems for billion dollar companies. Saving a few million is very little compared to the damage they could do if they guess wrong. Prefer choices that have big names behind them.

    The big opening today with Red Hat and VA today is deploying solutions on those scales today is to get them to deploy platforms today.

  • What about scalability of Linux systems?

    It is still true that, while we're dramatically bigger, we still inherit our software from the larger Linux community. I'm thrilled with the progress we're seeing because teams from SAP and Intel are starting to contribute. Network management tools are becoming available, better than not having anything.

  • Marketing strategy that Red Hat is doing now. Offering the platinum support package, $45,000 for an entire network, more expensive than Sun's comparable price. Increasing cost of Red Hat boxes.

    We sell services. The basic system is the $30 system, the more expensive version is the professional version and it comes with more services.

    Supplying high quality support is expensive. Sun can discount their software support by offsetting it with the cost of their proprietary hardware. Combine VA's hardware with Red Hat's support, it will be a less expensive solution compared to a proprietary solution.

  • One of the things, is Red Hat ready to commit to sending a person by plane to get something up and running. If I'm running the stock market on Linux.

    We intend to be able to do that directly and through our partners. Could we sign that contract today to support you in New York? I'd have to defer that, but we wouldn't sign the contract unless we have the resources to do it. If you don't get a yes today, you'll get it tomorrow. We understand that is critical here in New York.

  • The purpose of a company is to create value for their stockholders. Some distributions add proprietary software to add value?

    I think it is a self-defeating approach. The better technology seldom wins. The technology that is marketed better wins. What open source does for us is give us a unique value proposition. The minute we hook them on proprietary tools, we lose that unique value proposition and become just another little software company. Our company's value didn't come out from being first. There are others that were there first and were much better financed. Our success has come *from* our commitment to the philosophy. The competitors that take the semi-proprietary approach lose that value. I see it as a self-correcting system. A great deal of education of the media is necessary.

    PC Week is a great example, took a lot of quotes of mine out of a talk I did at Comdex to say that we're taking on Microsoft. Then they wrote a model to say we were arrogant. We give away everything we do, yet we have to change our spots to keep from becoming another MS. Help us educate the journalists, that this is not just another operating system. It isn't just better or cheaper, It has a value proposition that customers can't find anywhere else, from the control that customers get over their technology.

Eklektix, Inc. Linux powered! Copyright 1999 Eklektix, Inc. all rights reserved.
Linux ® is a registered trademark of Linus Torvalds