recently took some time to do an email interview with LinuxQuestions.org (Thanks Michael!). As you can probably tell from some of the questions, this interview is a touch old. If you have a question that you'd like answered, post it in this thread. I'll send a few of the best questions, as followups, to Michael.
LQ) Tell us a little about yourself. Where are you from, where did you go to school and the other basics.
MT) When most people ask this question, they mean "where did you get your degree?" I got my BS CSE from the Moore School at the University of Pennsylvania. That's the final resting place for several chunks of the first general purpose all-digital computer, the ENIAC. But I started learning about computers at home, about 1974, when my father bought and assembled an IMSAI 8080, then later a Cromemco Z2-D with three or four 64KB banks of RAM and a 10MB winchester hard disk. As I recall, the Z2-D computer cost as much as our station wagon. And that's when I started to learn BASIC, PL/I, Pascal, C, FORTH, LISP, and many other programming languages. It was a passion of mine since I was 12 to write a compiler, and after writing a few toy compilers in CS class, I got my chance in 1987 to transform the GNU C Compiler into the GNU C++ compiler, and later, to merge it as part of the GNU Compiler Collection.
Believe it or not, the Z2-D from 1976 was my PC in college (1982-1986). With my summer job at Cromemco, I'd upgraded it with parts from the scrap heap: a 68020 processor, 1.5 MB of RAM (3 512KB modules), a 48KB two-port graphics card. I also bought a shiny new 50MB harddisk which consumed my entire summer earnings.
LQ) What's the hostname of your favorite linux box and why is it named that? Also, if you couldn't use Red Hat or Fedora, which distribution would you use?
MT) I haven't paid attention to hostnames in forever, but if I were not using Red Hat or Fedora, I'd probably use Mandrake. Mandrake seems to have a very large number of RPMs available for it.
LQ) What was your first introduction to Linux? What was the reason behind you using Linux and was anyone in particular responsible for turning you on to Linux?
MT) My first introduction was via Adam Richter, creator of the Yggdrasil distribution. He called me up and took me to lunch one day, mainly to try to understand whether what I'd learned at Cygnus (the world's first company to commercialize free software) could be applied to the business he was thinking about starting. I didn't think so: we were selling support contracts for $35,000 to more than $1M per year, and he wanted to sell CDs for $99 (or perhaps even less). The two models could not have been more different.
I forgot about Linux until I got a call from Larry McVoy, telling me that there was this software company in North Carolina (software company in North Carolina!?) that had about 15 people and was growing by leaps and bounds. It was committed to free software, and Cygnus should look at acquiring it. While I was not that excited about Yggdrasil, I did become excited about Red Hat. We held a board meeting to discuss spending 10% of our equity in 1995 to acquire Red Hat but I could not convince the two other co-founders to make an offer. Four years later, Red Hat acquired Cygnus with 10% of their equity. Sigh.
LQ) I remember reading an interview with you in late 2000 in which you answered the question "Which distribution do you feel is your main competitor?" with "Right now our main competitors are Sun Solaris and Microsoft." Fast forward to today, do you think that same answer still applies?
MT) Moreso than ever.
LQ) Now that the dust from the initial Fedora announcement has settled and FC has a couple releases under its belt, would you say the project is as successful as Red Hat had hoped? In what areas would you say it really shines and what do you think are its biggest shortcomings?
MT) The project has been incredibly successful, and we have a lot of people outside of Red Hat to thank for that. What Red Hat must now do is to finish the job of making Fedora a true community project by publishing, and getting accepted, a governance model.
As a Linux distribution, Fedora shines, pure and simple. But it's hard not to: the work that the GNOME team, the Open Office team, the Evolution team, the Firefox team, the kernel team, I could go on..., are doing is simply amazing. And the Fedora community is blessed with a large number of people seriously committed to making it great. I think that Fedora does a fantastic job of showing people "the best of what works today in 100% free and open source software". I think that Fedora isthe perfect balance of leading edge technologies, yet robust enough that I can give my dad 4 CDs, he can install them, and then tell me "everything worked, as expected."
Although it is still playing out, I believe that SE Linux will become a mainstream technology, and I believe that Fedora was instrumental in making this possible. Perhaps some other approach could have yielded a positive outcome, but from what I saw on the inside, Fedora made it happen sooner, better, and with less risk than any other than we evaluated.
What do you see as the long term roadmap for Fedora? How closely tied to Red Hat will it remain and how to you see it fitting into the overall Red Hat product offering?
One of the best articles I have read in the past 5 years is Eric Von Hippel's article in the Harvard Business Review titled "Customers as Innovators: A New Way to Create Value" (April 2002). In the framework of that document, Fedora is Red Hat's user-innovation toolkit. Let me explain the difference between Fedora and Enterprise Linux, and then it will be obvious how they fit together. As I said, Fedora is the best of what works (or almost works) right now. In three to four months, there's a new best--some packages will be dropped, others added. Enterprise Linux has different properties: a release cycle of 18 months instead of 3-4, and a 7 year support lifecycle, which means that any technology we select into Enterprise Linux must be stable (unchanging) and supportable for a long time. There are lots of great technologies in Fedora that work today, but which we cannot yet commit to support for 7 years. Those technologies will remain in Fedora, maturing until we can make that commitment or find an alternative that can be supported in an enterprise context.
Last year was a difficult year: people were screaming at Red Hat for creating Fedora, for creating a version of Linux that could really stand up in enterprise environments (not just working right the first time--that's easy--but working over many years, which is hard), and for demonstrating that we were the first to demonstrate a profitable business model that scaled. Many people bought into our Enterprise Linux model, then immediately wanted to us to make all sorts of roadmap decisions that were inconsistent with the long-term support model. Becaue of the mess around our Fedora communications last year, people didn't understand that the place to innovate--for customers and the community alike--is Fedora. They tried to push changes into Enterprise Linux, and we pushed back. Now, people understand what Fedora is, how it works, how it relates to our future Linux products, and this has accelerated what we've been able to do with Enterprise Linux. If you look at our latest public beta, it's quite impressive. And when we ship it, people can have confidence that like earlier versions, it will have long-term support without breaking our company.
LQ) Do you think that RHEL clones such as CentOS and Whitebox have any impact on the bottom line? Do you foresee a gratis version of RHEL (or even RHL) ever returning to Red Hat?
MT) If the RHEL clones are having any impact, it's not observable to me. Our quarterly sales went from 98,000 to 144,000 subscriptions from fiscal Q1 to fiscal Q2 (the latest public data). As for a gratis version of RHEL--I don't see it. For people who want something that's supportable, they need the technology subscription, pure and simple. For people who want the best of what's available today, we already have that and its already gratis--Fedora. A gratis version of RHEL would have all the downside of technology and engineering compromises with none of the benefit of long-term stability and supportability.
LQ) A desktop Linux offering is currently absent from the Red Hat product offering. One has to guess that it's only a matter of time before one returns. Are there any particular milestones that are being waited on? Can you give us a hint as to when we may expect something?
MT) Well, if you check out http://www.redhat.com/software/rhel/desktop/
you will see that we /do/ have an enterprise desktop solution that is highly competitive. While we have already announced some substantial wins with this product, I think it gets to be even more exciting when more of the Stateless Linux project is incorporated. Stateless Linux, along with SE Linux, are perhaps the two "killer" technologies which, together, will redefine "secure" and "manageable" as it pertains to both desktops and servers.
LQ) On a similar note, Red Hat doesn't currently have any product aimed at the Home user. Is this something currently being considered, or is the market just not there yet in your opinion?
MT) We don't want to own the whole pie. We've got a business model that works, and we want to encourage others to grow the pie. I think that Fedora makes a great starting place for anybody who wants to address the home market. I give Fedora to my friends and family who want to run Linux at home--and they seem to like it! Still, it's far easier to underestimate the power of the open source model than it is to overestimate it. I am optimistic that we'll continue to see the open source community grow, and with it, increased support for the home user.
LQ) With an idea that started in 1987 and was incorporated in 1989, you are clearly one of the pioneers when it comes to turning Open Source into a viable and successful business model. What has been the most important lesson you have learned in this regard in the last 15 years? What do you think the industry needs to go to gain mass adoption and really hit critical mass (or have we)?
MT) The most important lesson I've learned is that no matter how you price, package, or place your products, at the end of the day the customer makes their own determination of value. If you can understand value from the customer's perspective, you can win more often than not.
As for critical mass, I do believe that free software and open source have hit critical mass, at least as the physicists define critical mass. Put another way, if any single entity were removed from the equation: IBM, Red Hat, the FSF, etc., the revolution would continue. OTOH, if Sun disappeared, Solaris would disappear with it.
LQ) Going way back, you have been a hacker who clearly enjoys coding. Do you still get a chance to really sit down and hack away at code these days? If so, which projects do you enjoying working on and what is the programming language you currently prefer?
MT) I have started to become really interested in GIS and statistics. I am learning GRASS and R. I need to learn more about the field before I can do meaningful hacking, but I imagine that in a few years, I might be writing software again (instead of just using it).
LQ) Red Hat just announced the purchase of the Netscape Directory Server and Certificate Management System from AOL, which seems to be a slight departure from the usual business plan. Can you give us some insight into what lead up to this acquisition, where you think these products fit into the Red Hat product lineup and what long term plans you have for the acquired products?
MT) I don't think this is a departure at all. Our customers want and need a scalable directory service, and the proprietary alternatives (Microsoft and Novell) were inconsitent with our vision for an Open Source Architecture. Purchasing the Directory Server and Certificate Management System help complete our solution without compromising our principles.
LQ) Do you think that situations such as the recent Mambo one are something that will hinder the adoption of Open Source? As a community, what do you think we should be doing to protect ourselves from these type of issues?
MT) Should everybody stop sending any emails because a spammer gets 9 years in jail? Of course not. People need to put events into perspective, educate themselves about reality (not what is reported by a sensationalistic or hyped-up press) and be responsible. In my view it is infinitely expensive to try to protect oneself from all possible risks, but if one is responsible, common sense can protect one from 99% or more of all risks, and for the remaining 1%, there is strength in numbers. People who buy from Red Hat know what they are getting, and for a great many people, it's pretty good.