Bob Young prefaces his answers to your questions by saying, "You may notice I've ducked some of the answers below - there is a reason for this. My role at Red Hat these days is as Chairman of the board. Matthew Szulik is Red Hat's CEO and will be a better person to answer some of the specific issues that these questions raise...."I have been spending more of my time recently working with Laurie Racine and her team at the Center for the Public Domain, www.centerpd.org. The Center's goal is to help improve the quality of the debate on Intellectual Property issues in the public arena. Our support of ibiblio.org (formerly Sunsite and Metalab) is just one example of the kinds of things we are doing to improve the ecosystem that makes up open source in specific and the public domain of knowledge generally.
"My answers will not be "official" Red Hat policy (that would be Matthew's job), or even the Center's official positions, but rather my personal take on the answers to your questions."
Do you think a recession will help RedHat/Linux?
We had an Ask Slashdot a few days ago wondering whether a recession will help Linux or not. Since you're the CEO of RedHat, you probably have a better idea as to what effect a recession will have on RedHat and Linux. So, do you think you will gain more market share during a recession than you would otherwise?
As above, I haven't been CEO of Red Hat for quite a while now (which is why Red Hat Inc is doing so well - but that's another story ;-) but here's my take on that question:
Back in the early days of Red Hat we used to sit around debating how we were going to get MIS directors to take us seriously, when one of our directors, Frank Batten Jr (now involved with open source database company Great Bridge) would insist that we shouldn't be trying to sell the MIS directors at all. He insisted we should focus our sales efforts on the CFOs (Chief Financial Officers).
His logic was simply that the open source price/performance ratio was simply so much better than the proprietary software vendors that the CFO's would eventually override the MIS dept's preference for the safe, tried and true, binary-only proprietary software model they were used to, and require that they consider open source alternatives.
In a strong economy the CFOs were willing to fund the MIS directors budgets. In a slowing economy CFO's start to scrutinize every expense much more carefully. The existence of the lower cost open source alternatives are going to be very attractive to many companies who are currently paying millions of dollars of royalties to software upgrades they really would rather not have to purchase.
Who decides what goes in and how?
I'd like some insight on how the decision is made to include something in Red Hat Linux, how quickly to roll in new releases of software, etc.
For example, I've seen pre-releases of KDE get included and updated in rawhide (and I believe in actual Red Hat releases) rather often, but even the individual GNOME components are almost never updated until well after a full stable release is announced. There are other examples, but that's the main one that comes to mind.
There also still seems to be a lot of 0.x version software in Red Hat to this day. So, I'm just curious how you make these technical decisions.
There is no one method. It varies depending on which component of Red Hat Linux is involved, and what the development community around that component's advice to us is. Ask Slashdot to arrange an interview with one of our engineering leaders and you'll get better answers than I can give.
But... "There also still seems to be a lot of 0.x version software in Red Hat to this day" .. is a -much- smaller issue than it used to be. It was only a few years ago that we used to ship beta code as part of our official releases simply because there were so few alternatives.
Today the primary reason we ship any beta code is so that the community of users who rely on Red Hat get more insight into where our technology is going and more opportunity to help influence what it does, how it does it, and how reliable future releases will be. Keep in mind that a big percentage of Red Hat users and developers still, to this day, do not have reliable high speed access and rely on CD-Roms as the source of the sources.
Would you really recommend it for desktop use?
I'm a Linux enthusiast and contributor but I still don't see where it's "ready for the desktop" as I would understand that phrase.
Bob, if you had a non-technical friend or relative who currently uses Windows, Quicken, Office, IE and AOL, could you in good conscience tell him it would be in his best interest to use Linux instead? What exactly would be in it for him?
Er, -I- am that non-technical friend, and I use a Red Hat Linux-based desktop exclusively.
The real answer to your question goes like this: No one (other than maybe some Slashdot reader) buys operating systems. People buy applications and then chose the operating system that best runs those applications. If you need to do desktop publishing you may go into a CompUSA store, find a copy of Aldus Pagemaker and then read on the side of the box that it runs on Windows or MacOS so you buy a computer with one of those OSes to run your app.
If you work as I do: on the net (Netscape) including all the net-based apps that I can run from my browser, reading email (exmh), and printing the occasional Word file (Applixware), you can do all these things on a Linux box every bit as easily and a great deal more reliably than on any of the 1980s legacy OSes. On the other hand if you need some application that only runs on Windows you may have problems with a Linux-only computer.
Fortunately, the future of the desktop is not in the 1980s applications that required you to load the application yourself, run it, back it up, and otherwise play sys-admin to your own computer. The future of the desktop will be using Internet appliances where the applications will sit out on the net (or your corporate Intranet) and you just download the small pieces that allow you to use those applications without having to take responsibility for them. Your sys-admin will not have to walk down the hall to have hands-on access to your machine. He or she may be in Australia and will support your machine remotely. This model requires a "real" operating system and will spell the end to a lot of the OS lock-in that all the Windows 95 and 98-based apps on the shelves of CompUSA represent.
Why invest in RedHat?
RedHat has the biggest name recognition of all the Linux distributions. To many non-tech types Linux == RedHat. And you are now breaking even, yet despite that RedHat's stock went from $80 a share to less than $5 and there doesn't seem to be a sign of that turning around.
What do you say to people who ask why they should invest in RedHat? Also, as a high-tech company I'm sure employees got stock options, how are they dealing with the crash in share prices and how do you convince them their options are still worth something?
As I've been preaching to anyone who would listen since long before Red Hat went public, Red Hat is a good investment if you believe that the proprietary binary-only software model as practiced by most of the software industry today is broken.
It is broken in that it does not conform to what customers expect from suppliers in free-market democracies, namely that the customer is normally in control of the customer-vendor relationship. It is only in the software industry, as this industry has evolved over the last thirty years, where the vendor is in control of his customers in an almost feudal way.
In the middle ages the feudal system was based on the ruling classes keeping the population under absolute control by not giving them any insight into the laws they were governed under. In effect you could be thrown into jail for breaking a law, and the policing authority did not have to tell you what law you broke or why it was in place.
Software users today are prohibited from making any changes to the software that they are building their organizations around, whether to add features their users desperately need, or stop their servers from crashing unexpectedly, or to patch a security hole, completely arbitrarily. In fact they can be thrown in jail for improving the systems they are using. If you don't believe me just read any shrink-wrapped proprietary software license.
Changing this industry is not going to happen overnight. So buying a portfolio of open source companies stocks, including Red Hat, may not pay off in the short term, or it might -- I'm not an expert on the stock market. But open source is solving a major structural problem in the software industry on behalf of the consumers of the products of that industry. The companies who enable this change to occur on behalf of their customers are going to be good investments in the long term.
Your impressions on the recent MS Interview
[Earlier] on Slashdot we got responses from an interview of MS exec Doug Miller and he touched upon some areas of Linux that caused a lot of debate and discussion in the forum. My question to you is, would you skim through Doug's reponses and provide us with your counterarguments or comments?
Doug Miller and I go way back. All I can say is how disappointed I am that Doug has joined the "dark side". He is a really decent human being.
He used to provide the exact opposite answers to the ones he provided in that interview and he did so with much more conviction.
Also see 9 below.
Linux-Mandrake started off simply as basically a copy of Red Hat Linux with KDE installed (which was the most advanced desktop environment at that time). However since 'growing' from Red Hat, it has become a distibution of it's own, with a incredibly simple install, more features/applications (ReiserFS, more intergrated Gnome/KDE menu...), Pentium optimised applications, and generally is more 'bleeding edge'.
How do you feel about the fact that Red Hat Linux was 'the womb' of what would now have to be considered one of the strongest Linux distros for the desktop, and a major competitior to Red Hat on the desktop with it's claimed "99% Red Hat compatibility"?
Mandrake is a great example of why open source is so valuable for the user of software.
The software using marketplace is not a simple single market. It is a vast collection of markets. Software developers who need good C and Java compilers are as different from dentists who need good dental office billing systems as two markets can get.
If Mandrake do their job properly they will serve some market(s) better than Red Hat does.
So for all those potential Linux-users who might not use Red Hat Linux because it does not include some application that they need, or support in a language Red Hat does not offer, we can point them to Mandrake.
The result is more choice in the marketplace. The customer wins, which is the whole point.
Recently we've seen several worms attacking vulnerabilities in the default install of Red Hat Linux. What is being done to make the default installation more newbie-friendly from a security point of view? The average desktop user probably doesn't want or need BIND, do they?
I have to duck this one. Security deserves precise and detailed answers and I'm not qualified to give them.
Other than to say the price of security is eternal vigilance.
Also, you should check out Red Hat Network. (www.redhat.com)
Competing against MacOSX
With MacOSX arriving as a desktop Unix (more or less) backed by a known, (sometimes) respected name, do you consider Apple to be a serious competitor, the same as Microsoft? Would Red Hat ever consider a PPC release to try and steer people away from MacOSX? Or, instead, do you think Apple will remain largely a niche player, but one that adds weight to the all-purpose viability of Unix?
"Apple will remain largely a niche player, but one that adds weight to the all-purpose viability of Unix."
This is -exactly- what I think. Go to the head of the class.
Doug Miller, a Microsoft executive, was recently interviewed for Slashdot. Many of the questions posed were regarding the competitiveness of Linux with Windows in the medium-term. To paraphrase, Doug said that there was no viable business model based on Linux, that the lack of standardization (ie. KDE v. Gnome) would be enough of an economic disincentive to commercial application developers to prevent them from venturing into the market.
On the face of it, he seems to have a point. What do you think? Does Linux need to be herded down the path towards a super-majority recognized 'standard' to be successful, or can the type of open-source movement to date provide enough tools and applications to drive Linux to dominance?
Saying that there is no business model to open source is like saying there is no business model to democracy.
Miller's arguments are red herrings thrown to distract the debate from the real issues. The real issue is not standards or technical compatibility with specific pieces of binary-only software. In an open source world compatibility is not dependent on whether a binary-only supplier adopts the standard as written or not. Whoever is trying to achieve compatibility simply has to check the sources. So the games that the binary-only proprietary software vendors play with standards go away. The real issue is where are the innovations of the next generation going to come from. The Millers of this world think that their employers should have preferential rights on offering those innovations.
This control over future innovation is being assisted by the trend in our legislatures worldwide to expand Intellectual Property (patents and copyrights, otherwise known as government granted monopolies) rights. These enable corporations to maximize their profits at the expense of the citizens in our society and are not just bad for our democracy, it is bad for business.
Let me be clear: Intellectual Property rights (IP) such as patents and copyrights are good things. But like anything in life too much of a good thing no longer is good. Too little vitamin D and you get bone diseases. Too much vitamin D will kill you. IP can be useful to independent inventors to protect their invention from being copied and marketed by larger distributors without compensation to the inventor. But today IP is mostly used by the armies of lawyers employed by the largest technology and publishing companies to squash potential competitors who don't have access to equivalent legal resources.
Just one example: For the first hundred years of copyright, copyright terms lasted less than 20 years. In the last 40 years copyright has been extended to 70 years plus the lifetime of the author. How government granted monopolies, justified as a means for providing incentive to authors to create additional works, achieves that goal *70 years after the author's death* mystifies me.
Needless to say this change in the structure of the rules that govern our society was not promoted by authors. It was promoted by the people who truly benefit from the extended IP rules, namely the publishers. Why we as a society are so keen to reward global publishing companies at the expense of the authors, musicians, researchers, artists, software developers, inventors, innovators, and entrepreneurs is another mystery.
All of these groups, in fact our society in general relies on the public access to knowledge. For every Metallica who worries that they will not be well enough paid for their next album there are literally hundreds of thousands of musicians whose ability to create and perform is being gradually eroded by the additional IP rules being imposed on us by our elected officials in the name of supporting the technology and publishing industries.
(For more detail on this check out www.centerpd.org.)
Giving the major global publishing and technology companies ever-greater government granted monopolies on vast definitions of technology (think "one-click" patents on a website) is inconsistent with how free-market democracies are supposed to work. Then arguing, as Miller does, that any alternative model cannot be legitimate because it does not generate the monopoly profits his employer collects would be a joke - if were not for so many of our legislators buying this line of reasoning.
I don't want to discourage you from writing to your representative, or your member of parliament, they need to hear from you. But here's a more positive thought on this topic: The cool thing about free-market democracies is that the citizens are the consumers. So you can sometimes solve societal problems in the marketplace. This is where Red Hat's and the rest of the open source suppliers opportunity lies.
Linux and open source is succeeding simply because it works in the customers interest better than the proprietary binary-only model does. It gets around the innovation deadening impact of software patents. It avoids the creativity sapping effect of locking up knowledge and expression behind 70 year copyrights.
While open source may not be a business model any more than democracy is a business model, it is possible to use open source to serve your customers better than the competition. Matthew Szulik and his team have driven Red Hat from revenues of less than $15 million when we went public 21 months ago, to over $100 million today with gross margins in excess of 55%.
On one thing Doug Miller and I agree: serving your customers is how you build great companies. I just don't see how locking your customers into inflexible binary-only proprietary technologies over which they have no control can be defined as serving them. I guess we still have some work to do getting the word out that there is a more robust, reliable and economic way to use technology, it's called open source.
Red Hat Acquisitions
I noticed that while Red Hat was valued highly, Red Hat used its funding to purchase companies like Cygnus and C2Net. Escpecially with the purchase of Cygnus, you appear to be consildating the infrastructure that makes linux viable commercially. One could conjecture that you are trying to provide developer tools and resources, both as a product and as a way to build into Linux (as in the motto "it is the developers/ISVs stupid!"). Red Hat is currently valued much lower than it was at the top of the hype, but one could argue that these (and other) strategic acquisitions give Red Hat an edge over the competition or the chance at surviving the tech stock maelstrom.
Q: How do you see these acquisitions as helping Red Hat and its position in the market?
In general the answer to the above is: yes. As in, yes we see the need to offer comprehensive and high quality development tools as extremely important to the future of the platform we are promoting.
The only point to be clear on is who we see as competition. Red Hat's success to date has been due to our focus on the real competition. When we started the whole Linux/open source market was not big enough to pay our credit card bills, much less the rent.
So we new we had to take customers from the established industry leaders. We knew we could do this by delivering benefits that those billion dollar competitors were not willing to offer their customers.
This benefit was the control over the technology we were asking them to invest in, that open source enabled us to deliver.
Which is why we don't see Mandrake as the competition. We have products and services we can sell to a Mandrake user. The same is not true of the big binary-only proprietary software suppliers.
Unfortunately, most hardware vendors support Microsoft because MS has the largest share of the market and they know it will pay to support MS with drivers.
Linux is not in that state, save for (perhaps) networking devices. Has RedHat considered helping to fund driver development for other forms of hardware? I'm thinking mostly of 3D accelerated video cards (by helping to fund the DRI group), but other items (scanners, USB- IDE interfaces, etc.) would be nice too.
Yup, this has historically been a problem. Red Hat (and many of the other leading Linux distributors) has contributed drivers, and has contributed development resources to other device driver developers.
But more and more manufacturers recognize that they will sell more of their boards, add-ons, peripherals, and systems if they make the small effort of ensuring that it will run the latest Linux kernel and libraries. For example HP is now actively writing open source Linux printer drivers for the popular HP printers.
- 30 -