 Thank you, and to all of those of you in the audience for whom English is not a first language, I apologize, because Spanish is not only not my first language, it's really not a language for me at all. I want to thank our hosts for inviting me to talk again this morning. It's been sort of a tradition the last few years for me to say something about the relationship between HP and the Debian Project. We have a relationship that goes back for many years. I think you've just heard some words and some thoughts about the importance that the Debian Project plays to at least one government organization. What I want to try and do today is a little bit different than what I've done in the past. So those of you who've heard me talk about HP and Debian hopefully in the past won't be completely bored today. I want to talk a little bit about the relationship between HP and Debian from the perspective of what it's like for a large corporation to try and participate in the free software process and how HP tries to do that with Debian. But first, a little bit about who I am. Out of curiosity, how many of you in the room have not heard me speak before somewhere? A few hands in the back. Okay. I won't spend a lot of time on this then. The probably the most interesting things are that I tell people I made my first personal contribution of source code to this thing that we now call free software in about 1979. And if you know your history, this is before the GPL existed. It was before we had great terminology for talking about this whole phenomenon. It was even before I was employed by HP. In fact, I was still a high school student at the time. And if you'd like to know the story of what that contribution was and what it was about, I can tell you about that some other time. I've had the pleasure of being employed by Hewlett Packard where it's been off Agilent technologies since 1986. And so, not as long as my involvement in free software, but also, you know, very long and very rewarding for me experience. I became a contributor to the Debian project very early in the project's history. In fact, today, there are very few people still active in Debian who have been around as long as I have and I don't know sometimes if that means I'm just too dumb to find something more fresh and exciting to go work on or whether it's a really great demonstration of just how important I think the Debian project is in the world. I've had lots of different roles in Debian over the years. I continue to maintain a non-trivial number of packages. Some would say perhaps not as well as I should, but that's life. I also, at the moment, serve as president or chairman of the Debian Technical Committee. But perhaps more interesting, I've been for the last couple of years the president of Software and the Public Interest, which is the nonprofit corporation that exists in the United States to hold property and trademarks and finances and so forth on behalf of not only the Debian project, but a number of other very significant free software projects these days. I invite those of you who will be here for the rest of DEBCONF to come to the SPI BOF that we have scheduled sometime later in the week. I don't remember which day it's on, but there are actually several members of the board of directors of SPI who are here now or will be here by then and some past members of the board and so forth, and it's a great opportunity to learn a little bit more about that organization and its relationship to Debian, how the finances of the project as seen from the U.S. perspective are going and so forth. I also have the privilege these days of serving on the board of directors of the Linux Foundation and the Consumer Electronics Linux Forum, both representing HP. And for those of you who pay attention to news in the United States, you may have seen that there's a new coalition that was formed within the last week called Open Source for America, and I'm pleased to be one of the founding members of this. It's an organization that will attempt to encourage greater use of free software within the U.S. federal government system. But then those of you who know me well at all will understand that, you know, what I could spend all day talking about are the things I do with free software associated with my hobbies, most notably playing with high powered model rockets and building satellites. Since the last time we had a DEBCONF last November, I had the opportunity to fly this rocket, which is so far my biggest toy. That's, I guess, if I do the math in my head about 26 centimeters in diameter and, you know, it's as tall as I am. And this is what it looked like taking off on its way to about 6,000 feet or not quite 2,000 meters elevation. I happen to think that's a really pretty picture. That's Pike's Peak in the background, which is a very significant landmark in the front range of Colorado where I live. And not only do I spend time playing with building and flying the rockets, but this is a completely open hardware and open source solution for the avionics for collecting data on these rockets and controlling their operation that I've been working on jointly with Keith Packard, who many of you will recognize as another Debian developer and long-term contributor to the X-WINDOW system. Keith and I have given some talks about this at other conferences, including places like Linux Conference Australia. And in fact, just last night we submitted some talk proposals for LCA and Wellington, New Zealand next January with an update on the new version of this system that we're currently working on. If you're interested in this, there's lots more information on our website at altosmetrum.org. But, you know, it isn't all just about playing with rockets. This has been a really interesting year for me. Some crazy things have happened. I don't know how many people have had the experience of having Linus Torvalds as a barber. I'm not sure I recommend it, particularly when he comes at you with shears like this. If you haven't heard the story, at the Linux Conference in Hobart, Tasmania in January this year, there was a fundraising activity to try and generate some money to help with research to try and save the Tasmanian devil. These cute furry little ferocious beasts are at risk of becoming extinct because they suffer from a facial cancer that is transmissible between different members of the species. And the population is fairly small and constrained to the island of Tasmania. And so they're very concerned about the fact that this facial cancer could end up causing them to become extinct. This story is sort of long and complicated. Go search on facial dignity if you'd like to find my blog post about all of the details. But by the end of the week, we had raised more than $40,000 Australian to go to this charity in exchange for which I agreed to let Linus shave me. And the results, frankly, were pretty hideous. Some folks said, oh, BDL, you look so much younger. My wife said, please don't wait until you come home to start growing your beard back. So as you can see, fortunately, she congratulated me for raising so much money for charity using something that cost me nothing to acquire and would cost me nothing to replace. And I'm pleased that it did actually grow back. I also, last month, had another one of those sort of mind-blowing, life-altering kinds of experiences. This was a picture that was taken at the Forum International Software Livret in Brazil last month. That's an amazing event by any measure. More than 8,200 people attended that Congress this year. It's clearly one of, if not the largest free software events that happens in the world. This was the 10th anniversary of the event. And to celebrate, the Brazilian president, Lula, showed up at the conference. And not only did he show up and shake hands, but I was very impressed to see him give about a 30 or 45 minute talk without notes, without teleprompter, or anything, speaking very clearly and very passionately about how important free software is to the economy of Brazil and their success at growing to be what is now the world's 12th largest economy. It was the first time I've had the chance to shake the hand of a sitting head of state of a significant country. And I have no idea if that will ever happen again. It certainly never happened to me in my own country. So anyway, enough about me. It's been a fun year. Lots of interesting things have happened. This is a quote that I love to share with people from Evan Moglin, who many of you will recognize as one of the people who helped to create the GNU General Public License, particularly who helped to run the process by which GPLv3 was drafted. His law professor in New York is now the head of the Software Freedom Law Center and various other related things. He said a few years ago in a speech given at Harvard Law School that we now have a body of software accessible to everybody on earth so robust and so profound and its possibilities that we are a few men months away from doing whatever it is that anyone wants to do with computers all the time. He was, of course, speaking about the body of available free software. And this is something that caused me to pause and stop and think, because I believe it's absolutely true. And it helps to focus us on what I think is a fairly profound thought, which is that the question really is no longer whether free software makes sense in some fundamental way, but rather when and how best we should use free software, particularly in the corporate world, to try and pursue our business objectives. And part of the reason that this ends up being such an interesting question is that the way free software works is so fundamentally different in how it's produced and maintained and consumed than other software models. A key attribute of Linux and many free software applications is that they are developed and supported quote unquote by the community. What does that really mean? Well, a couple of the key elements of it are that no single company is in charge. There are in fact a range of contributors with varying interests and abilities and motivations who come together to create something that represents a much more substantial aggregate contribution of human intellect and willpower than any single company or any single government could ever afford to bring together in any other sort of model. And as a consequence of this community process and the rules that the community has chosen for engaging with each other, often expressed in the terms of the licenses that we choose to use for our software, users end up having flexibility in how they want to acquire support for open source technologies. And this is because we've in effect lowered the barrier of participation by licensing our software in a way that causes everyone to have, everyone who uses the binary programs to have access to the source code, we in effect have said that any user can become a developer or in many cases, more importantly, they can pay someone else to be a developer on their behalf because frankly as much as some of us would like to believe otherwise, most of the users of our free software don't really want to be developers but the freedom that they get as a consequence of the licensing structure and the relationship between producer and consumer of software that we've chosen opens up a whole range of possible behaviors that they wouldn't have access to otherwise. And of course there's this ultimate notion that if whoever the upstream creator or maintainer of a piece of software is behaves badly, we always have the option to make a copy of that source code, create a new focal point for activity in the process doing something we call forking and to carry on developing the software in a new and better way. And we've actually seen this used a few times in the past. I think the transition of the X window system from the old X386 structure to the way it's managed now through X.org and free desktop.org is an excellent example of this. Development had stagnated, there was a small group of people who were gating the ability of contributors to make progress with the software and the community eventually said no, no, no, this is not healthy, it's not good, we wanna do this better. And they were able to do that because of this open collaborative process that's enabled by the kinds of licenses that we've chosen. So what's the impact of this whole sort of community development process for companies? Well, they're sort of good and bad things. There's exciting stuff and there are challenges. One of the exciting things is that the emergence of these sort of internet enabled communities of developers and users that frankly would have had a hard time existing before the internet because one of the keys is that in order to create a critical mass, you have to have enough people with common interests who can come together to form a community. The existence of the internet allows this to happen over broad geographic distances and among people from different cultural and economic and social backgrounds. But the combination of the emergence of these communities and the behavioral models that these communities have chosen enabled by open access to standards really has accelerated the development and adoption of free software in the PC industry. And I think it's interesting, one of the big major industry analyst firms, Gartner, has put out these predictions where they they changed the date and the number a little bit from time to time but they're basically saying that we're a year or two away from something like 80% of all software in the world having at least some open source content in it. And I have no trouble believing this because sitting as I do on HP's internal open source review board, I can tell you that today I believe HP is in the software that we build and ship to customers and we are now one of the 10 largest software companies in the world though that's not usually I think how people think about us. I suspect that we're at or beyond that point already. And the other thing that's really interesting to me that's very powerful is that the free software movement allows innovation to come from really surprising places. And this is because of what we call the long tail of contribution. As an example, one of the recent releases of the Linux kernel had more than 2400 distinct people who had made at least one patch submission in that kernel release cycle. If you go and look at a histogram of how many patches and how many lines of code and all that, you will see that the number of significant contributors who do a lot of work on the kernel is much smaller than that. It measures probably in the dozens for the Linux kernel for other software projects it might be even smaller than that. But the exciting thing is how many people there are out in that long decaying tail of contribution who made one patch submission or maybe two. And what's exciting about that, these are people who are using this software for purposes other than to be a developer of the Linux kernel or some other project who have run into some personal issue or concern that got them motivated enough to go understand a problem, figure out how to fix that and make a contribution. And what we have seen time and time again is that this ability of people who are not fundamentally setting about to work on a particular piece of software but who in the process of trying to do something else run into something that motivates them to put in a chunk of time to help discover and triage a particular bug to come up with a particular fix just opens the world in a fundamental way to a level of contribution and innovation and a rapidity of development and improvement of software that could not exist under any other software development or related economic model. So if we wanna take advantage of this as companies, if we want to have access to this incredible stream of development and technological enthusiasm, we have to learn to productively participate as citizens in these community development processes if we want to be able to benefit from the open innovation that can come as a result. So I said I was gonna talk a little bit about sort of what it's like for a big company to try and work with a free software project. I think one way I can do that is to compare and contrast a little bit what I think free software developers like me really want and then talk a little bit about what I think large companies really want and then we can see what some of the challenges might be. It's really difficult to characterize what free software developers want. Why? Well, we're all very different people. We come from different backgrounds. We have different social and economic pressures. We have different experience bases. We speak different languages. We look different. We all look funny, right? But I do think that we share some very common elements. One of them is that we're usually motivated by doing what I call scratching itches. This is the desire to work on something that matters to us personally. Sometimes we work on a piece of free software because someone is paying us to. But the most productive contributors to our communities of development are the people who are there because they want to be. They are personally motivated and they're working on something that really matters to them. Another thing that I think we share in common is that we really don't like working on schedules. It's not because we're a bunch of crazy anarchists, though sometimes maybe we are. It's really because we are working on things that we are personally passionate about. We often are working on things we're not being directly paid to work on. And therefore, we're working on this, even if we're doing it with the approval of the manager of our company, saying, yes, it's okay for you to work on that for a while. It's still sort of a spare time activity for many of us. I laughingly explained to somebody once that I work on free software a whole lot more than I would if I were really being paid to do it all the time because as a passionate thing that's sort of an avocation for me, I work on it every time I have free time that isn't being soaked up playing with rockets. And that's actually a lot of time. But this notion of trying to make deliveries of particular pieces of development on a schedule is something that's very difficult for us. We can sometimes get motivated to meet a common deadline to sort of agree that we'll try to accomplish something by a certain point in time. But that's very different than sitting down in the way that many corporations think about developing software of how many engineers do we need to put on this for how many work days in order to deliver a particular result. It's also, I think, a common element among all of us that we really value the collaboration. We like sharing with each other. We like it when other people look at the work that we've done and say, that's cool, I wanna use that. It's particularly interesting to me when people say, oh, that's really cool and I've thought of a use for that that you didn't think of and I'm gonna go do something that you never anticipated that's really cool. And then I look at it and I go, oh, wow, that's really great. I'd like to take that back and do something else with it. We get these sort of feedback loops going where we're collaborating and working to drive forward something that maybe none of us anticipated in advance. And fundamentally at the end of the day, the thing that drives a lot of our behavior is personal reputation. I've said in the past that one of the reasons I think Debian has such a history of delivering high quality software is that we all understand that every time we upload a package, every time we make a release, the thing it's going to affect is our personal reputations and we care a lot about our personal reputations. Every once in a while you look at some contributor behaving in a way that you think is a little strange and you wonder what they could possibly be thinking because you know that what they've just done is going to have a negative hit on their personal reputation and you can't understand why they would want to tolerate that. And I think at the end of the day it's important to understand this and to think about it because there is a big distinction between the concept of personal reputation and what a company actually cares about. So what is it that a company like HP really needs? Well, I hate to be this way but the number one item is revenue. We need money. That's how companies operate. It's what makes the whole thing go. You know, I could say profit. I could say return on investment. I could use all sorts of things. But fundamentally what we have observed through the history of business is that if you take more money in on the top line the rest of the things all sort of take care of themselves if you're not being an idiot. The other thing that companies really need is growth. If you have studied economics you just take this as a given. If you haven't you might go well why do you have to grow? Why can't you just stay the way you are and keep making the same amount of revenue? I would encourage you if that's something you don't understand to go read an introductory text on macroeconomics. The way the economy of the world works standing still in a business is an excellent recipe for failure. It just doesn't work that way. And in order to get revenue and growth what we've come to realize is that companies really need to be able to differentiate themselves. This is particularly true for companies that participate in commodity markets. How many of you make decisions on which x86 family servers you're gonna buy based on the logo that's on the front of the box? I bet you that's not often the first thing that you look for. Unless of course you're enlightened and always buy HP gear. But my suspicion is you care a lot about what it costs. You want to look at the features and say does it have sort of the minimum set of features I expect? And then after that you're largely driven by questions like cost. If you are enlightened, if you've been doing this for a while you may care about the quality, you may care about the ability to acquire support, you may care about many other things that we think are very important in the products that we deliver to the market. But at the end of the day we have to find some way to differentiate ourselves whether it's by building better hardware, providing better services, better support options. In some way we have to figure out how to differentiate ourselves so that when there are three different similar things available to you you'll pick the one that we make. And part of all of this is building a sense of corporate reputation. Corporate reputation, if you talk to marketing folks in a big company they talk about the brand. They talk about the importance of building the brand, of building trust in the brand, of people admiring and appreciating the brand, the trademark. And what is that? That is building a reputation for the corporation. The reason this is an important distinction is that companies actually have a really hard time being enthusiastic about individuals who work for the company building and developing their personal reputations. It's not because they think it's bad, it's just because that's not something that immediately obviously leads to more revenue, more growth and differentiation for the company. And I think it's very important because the way the free software world works, if you go and ask Linus Torvalds how he makes decisions about which patches to accept in the Linux kernel, the element of reputation that matters is the reputation that that individual submitter of a patch has established over time as being someone who is clueful and delivers good content to be incorporated. He doesn't actually generally care who they work for. It's not important to him. He doesn't look at the part on the right-hand side of the email address except to recognize what person it is. And this is something that I think companies have to be conscious of and have to think about. It's certainly something we've had to recognize at HP that if we want to be able to successfully build corporate reputation in the free software world, an important component of that is to enable our employees to be able to build their personal reputations in the community. So let me talk just a little bit more about the whole revenue and growth thing. I tell people this and sometimes they seem to have a hard time understanding it but participation in the free software process for a company like HP is a business, not a hobby. What I mean by that is that publicly traded companies must legally, must morally, must pursue profit. It's the thing that we're expected to do. It's all about producing a return on the shareholders' investment and helping to grow the company and be more successful. So I think it's important as free software developers that we always ask ourselves what is it the company expects to get in return for their involvement and support in our project? Because over time, the expenses increase, the revenue has to grow to match, the companies are very interested in creating opportunities for their employees. So at the end of the day, I think it's important for us to remember that the motivations that a company has for participating in our project may be a little bit different from those that we have as individuals. And when it comes to this concept of differentiation, the problem is that technology companies really can't afford to compete only at delivering commodities. That drives the profit margins way down. If the only thing we can say is we make something just like those guys, but it's a little cheaper, that's a losing business to be in. You don't generate enough profit to be able to afford to invest and continuing to develop and continuing to innovate and to do exciting stuff in the world. So what companies really find themselves often striving to do is to figure out how to add value in ways that customers will be willing to pay us for. When it comes to things like servers, it's pretty obvious. If we provide the world's best remote management interfaces, that may differentiate our servers from somebody else's servers and make them more attractive, even if in some cases they're a little bit more expensive. I think, given our market share in the server world, that's a strategy that's worked pretty well for HP. But we are constantly having to look for ways to create something that is an industry standard that works just as well and plugs in just as compatible as everyone else's and yet has some differentiating feature that will cause people to be willing to provide more money to us than they would otherwise. And traditionally, the goal for doing this was to create control points. And often control points meant patents. If you have a patent on something and that guarantees that only you can successfully participate with that feature in the marketplace or you can force other people to pay you a little bit every time they do the same thing, that's clearly a model that can work for a company that wants to produce differentiation. But it doesn't work well in the free software world. So one of the challenges that we face in collaborating between large corporations and the free software universe is how best to create affinity for our products in an open source collaborative context where we don't have these sorts of hard control points. So I think there is something that we can ultimately agree on though. And that is that what we're really trying to do is deliver positive user experiences. As developers, we want to do this because we want others to be able to benefit from and use our work. At the sort of most fundamental level, lots of us write software because it solves some immediate problem that we have. And so maybe we're satisfied on some level if it works for us. And maybe we don't always care if it works well for other people. But I suspect that no matter what they say, almost everyone who actually participates in the free software process gets a positive feeling and appreciates the feedback that they get when other people look at what they've done, appreciate it and want to use it. As companies, I think it's all about how we motivate customers and want us to give us our money. Look at some of the most popular and attractive consumer devices over the last few years. The iPhone is one really good example. Why is it that it's so successful? It's certainly not because it's open. It's certainly not because it uses free software. It's because it delivers a compelling user experience. And so I think to some extent, what we can agree on is that at the end of the day, we want to build software that allows us to deliver really compelling, very positive, strong customer experiences. So what has this led us to think at HP and how do we try to engage with the community given this sort of context and the set of constraints? If you've heard me talk in the past, you've probably seen this slide. This hasn't changed in years because it works. It's just a very fundamental part of how we think about the free software and open source world and how we participate. We decided very early in our history of engagement that we wanted to participate directly as open source community members, that we would support existing community values and behaviors while developing robust enterprise capabilities and other things that are important to our customers. And one of the most important ways that we've done that is that we chose to use existing open source licenses. In the history of involvement of large corporations in the free software world, I think HP remains unique among sort of the major players as the only company that has not yet felt the need to create our own open source license. IBM has certainly flirted with different licenses. Sun has introduced a string of interesting and frustrating licenses. Our friends at Apple have introduced licenses that they thought were open. Even Microsoft has sort of said, well, hey, this open source thing's kind of cool, but only if we can have our own licenses. You know, great, fine. I'm pleased to see more free software in the world under good open source licenses regardless of what that license is. But I think it's actually a more genuine, more direct and healthier relationship with the community when we can figure out how to participate using the existing licenses and therefore the existing expressions of values and behaviors that communities want. And then of course we work with our commercial distributions and other ISV partners to take this innovative work that we've participated in the community and deliver that as solutions to customers. And as I've said a couple times, and this is, I promise, the only sort of real marketing slide I'll throw at you, this has been a very successful strategy for us. These are some numbers that the IDC analyst firm puts out on a regular basis. And what it shows is that, you know, if we combine the HP numbers with the compact numbers from before the acquisition a few years ago, HP's now had over a decade of clear leadership in sales of Linux servers. And the numbers, the absolute numbers are actually fairly interesting. They claim that we've sold more than $13.9 billion worth of servers to people who want to run Linux. That represents more than 3.2 million servers. And in fact right now I'm told that HP builds and ships a new server once over 11 seconds, 24 hours a day, seven days a week. And if we look at the percentage of those that we and the analyst firms believe end up running Linux, that means that we're shipping a new Linux server much faster than once a minute, 24 hours a day, seven days a week. Those are pretty significant numbers. This is why this is such a significant part of our overall business and why it's important to the company to figure out how to be successful here. Okay, so maybe it makes sense for us to participate in the whole collaborative development process. Maybe it makes sense to do Linux, but why do we work with Debian? Well frankly, it's largely about the inclusive collaborative nature of this project. In some sense, I think Debian is perhaps the broadest, truest expression of the sort of fundamental expectations of what it means to be a free software project. And to embody the four fundamental freedoms that Richard and the Free Software Foundation folks have talked about for years. From a very practical standpoint, HP Working with Debian has enabled us to enter new markets that were not well served by commercial distributions. You heard some references to Telefonica and the telco industry in the first talk this morning. When we at HP chose to enter the market of carrier grade telecommunication servers, we went and talked to our commercial distribution partners, the Red Hats and Novels and other relevant companies at that time of the world and said, we wanna go do some hardware here, are you interested in participating with us? And they all said, that sounds like a really great market, but it's pretty far removed from what we're focused on right now. And if you HP want us to do this, feel free to give us lots of money and we'll be happy to go there with you. And we scratched our heads a little bit and said, well most of the work that we need to do is actually the platform enablement part, the part that we would have to be engaged in anyway. So at the end of the day, we ended up delivering something called a product that's called Debian plus HP Telco extensions. And in fact, we have shipped a surprisingly large number of itanium based servers running Debian and some Telco software that are used to do things like route mobile phone calls. In fact, I was told a couple of years ago that we were very close to the point where one third of all mobile phone calls in the world would at some point in the progress of the call touch a machine that was built by HP running Debian. That was a fundamental part of maintaining the call progress infrastructure for those mobile phone calls. I haven't kept up with this market. I don't know whether that number is still true, whether it's gone up or gone down. But I do know that at many times the number of itanium servers we have shipped with that particular software stack has exceeded the number of itanium servers that folks like the register believe have been shipped in the entire world. It's been a very big, very lucrative and reasonable business for us to participate in enabled by our relationship with Debian. Debian's also been a really high quality source of bits for us to build many embedded products that use Linux. There are right now well over 200 different products from HP that ship with a Linux kernel and related open source technology inside in such a way that it's completely invisible to the end user. They're buying a cool little network attached something. They're buying a blade server chassis and the fact that the processor that operates the chassis and does the resource allocation for all of the blades within that chassis happens to be running Linux is and should be irrelevant to the end user. They care about the feature set that that product delivers. And we used to laugh and say that the only way they would know what operating system that processor was running was if it was broken. And at the end of the day, it's interesting when I sit on our internal open source review board and I see proposals come by, even when they don't explicitly mention Debian, when they show us the tree of source code that they're gonna use and provide it to us to run through our analysis tools, I'm impressed with how often the version strings I see in the various sources and executables make it clear that they've grabbed a Debian source package to build that particular product from. Which isn't surprising because we discovered a long time ago that Debian was a great platform for internal developers to use. And it's not surprising that when people are running a particular distribution as their development platform that they might choose software from that when they're building products. Something that they're enabled to do in the free software world that's more challenging than other models. And frankly, I would be remiss if I didn't mention customer demand. The reason we ended up providing official support for Debian on our servers wasn't just because folks like me thought it was a good idea. It's because some of our customers said we really want you to do this. And some of those customers were important enough to us to have us listen and pay attention. So let me give you just a few highlights of the history of HP's relationship with Debian. Much briefer than I've done at some talks in the past. First of all, I don't know how many of you know or remember, but the first ever dedicated Debian server called master.debian.org is something I put together from a cast-off 48633 megahertz tower and probably be amused. Two, five and a quarter inch full height SCSI drives, a 330 megabyte disk for the operating system and a 660 megabyte disk which held the entire archive with lots of room for growth. I think our FTP masters or dams someone are gonna give a presentation later this week and talk about the current resource requirements and the way the mirroring network works and so forth. And Ganna fastened me the other day for some details about this system to paint a contrast. So you may actually hear a little bit more about that later in the week. At a certain point in history, it was very important to HP to have a Linux distribution for the PA risk processor architecture and then for our itanium two based servers. And so HP directly invested in ports of this. Some of you in the room today are people that I recognize were working for companies at that time that we contracted with to do some of that work. And so in that period, we were actually paying people to work on Debian and to help with the port of Debian to those architectures. Some of you may remember big flurries of in amused to fix porting problems and yes, those were being done by people that were being paid by HP to go make those architectures happen. And at the end of the day, the results from that I think were really quite positive. Another thing that Sam Hartman reminded me of a couple of days ago is that HP was the company that funded the legal opinion that finally allowed us to include crypto software in the US main archive. I don't know how many of you have been around the project long enough to remember, but this was a real problem for us for a while. We were contemplating the possibility of almost having to build two entirely different distributions. The one that we could put on machines in the US and not have to worry about it being exported from the US and the one that everyone would actually want to run in the rest of the world. And you know, there was a certain cry that said, well, gee, what's the problem here? Let's just host all the Debian servers outside of the United States and dodge this whole problem. The challenge at that time was that most of the bandwidth and most of the hosting resources that were being offered to the project were inside the US. And so to walk away from that would have been a big problem. The other thing was that we didn't really want to disadvantage all of the users of Debian in the US to having to use second class software. You know, short crypto bit strings and things like this. And so I was very pleased that in the end, it turned out to not be very hard to get competent legal help to analyze the situation and help us understand the process by which we could legally include crypto software in our US main archive. In fact, we gamed the system just a little bit by providing them substantially more documentation than we think they really wanted. And I was highly amused some amount of time later when they told us that, oh, yes, we were doing it perfectly and we were the example they held up to show other people how to do it. So some of these things are funny. Over the years, HP has donated a lot of hardware to the Debian project. To be honest, I've lost track of how much. But if you go look at the machine database part of db.debian.org, you'll see that there are an awful lot of servers that have come from HP over the years. And frankly, you know, we've been the largest single financial sponsor of Debian Conf going all the way back to Debian Conf 3 in Oslo. I personally have had the privilege of being a participant in every Debian Conf, since Debian Conf 2. I was really sorry to have to miss zero and one, both of which were in Bordeaux, which is a beautiful part of the world that I really like. But they conflicted with my family's annual beach trip and I had to keep my priorities straight back in those days. I don't know that my priorities have really changed much, but now I just bring parts of my family along to Debian Conf and it all kind of works out in the end. And finally, and perhaps most importantly to some people, HP was the first and I think still the only major server vendor in the world that has chosen to provide worldwide support for Debian on our servers. And I announced a year ago that there was this website that you could go look at hp.com.go.debian. That URL still works, but it actually points you to a new place now that I'll talk about in just a minute. So what matters to HP going forward in our relationship with Debian? Well, I've already talked about some of this. It's important for us to be able to increase shareholder value to focus on end user experiences and to be able to balance sort of the things that the company chooses to do directly and what we expect others to do on our behalf under contract and so forth. And I articulated last year in Argentina in my presentation about HP and Debian's relationship at that time that there are some challenges for HP or any company like us in working with Debian. And frankly, while we've made progress in some areas, these challenges all still exist. They maybe always will exist. One of them is that just, you know, Debian isn't a company. And so the way we do marketing and support of Debian has to be different than the way we do it with a commercial distribution partner. You know, Debian doesn't really care about non-free software vendors running on our distribution or not. It's not a principal concern that many of us have. And yet a lot of customers in the enterprise world who want to be able to run Linux as their operating system platform really do care about whether they can run Oracle, whether they can run SAP. And you know, frankly, until and unless something really fundamental in the universe changes, I don't know that Debian as a whole will ever care enough about that to cause the world to change in a way that makes those things happen. So in the meantime, there are some places where no matter how much we might appreciate Debian and want to support Debian, we do end up recommending other distributions to our customers because there's a support relationship or a contract situation that just works better for a customer. Another challenge is predictability. Frankly, we've gotten better about this. Last couple of stable releases we've done have been much closer to the times that we advertised we were trying to hit than we used to be way back in the past. But as I commented in an interview that I gave to somebody in the press a while back, one of the things that is important is that when teams are trying to provide official support for stable Debian, they have to be able to anticipate when they need to commit the resources to requalify, recertify, and do the related integration work around new releases. Sometimes we finesse this a little bit by having people that work for us be permitted and encouraged to be key members of various teams. Things like Dan Frazier's involvement in the kernel team is one good example of that. But at the end of the day, it's the project as a whole and the contributions that all of us make to trying to hit the release schedules that our release managers announced some period of time in advance that ends up constraining whether the project seems predictable to our commercial partners or not. And I won't get wound around the axel on the kernel driver and firmware issues. I think everybody is working to try and do the right things at the right times with this, but it is challenging to us right now that the economically and technically most viable alternatives for things like gigabit ethernet interfaces in our servers happen to be from a company that likes to deliver a binary downloadable firmware blob for their card. And so having a good solution for how we enable customers to be able to install on those sorts of systems is pretty fundamental to their ability to use Debian on our hardware. So at the end of the day, the point we've gotten to is this realization that by directly participating in and collaborating in the free software community development process, in effect, doing as much of the work as we can as far upstream in the chain of development as possible, that we end up enabling ourselves to be very broad in what we can support. And in fact, one of the things that's changed in the last year is we don't just provide a certain level of support for Debian in addition to our principal commercial distribution partners, Red Hat and Novell Suza, but we have realized in the last year that the things we did in order to successfully support Debian also allow us to successfully support many other community distributions of Linux. And so the hp.com.go.debian, that link still works, but it actually points you now to hp.com.go.communitylinux. And this is actually, I think, an improvement, not a retraction or reduction in the quality of support and solutions that are available for Debian. It's just a realization that a lot of the things that we had to do in order to make it possible to support Debian have also made it possible to support Ubuntu and CentOS and Fedora and so forth. And so today, I firmly believe that hp provides a broader and deeper level of support for Linux on our server systems and a growing amount of solutions for Linux on things like notebooks and netbooks than any other company in the world, and that's something I'm really quite proud of. So I'll wrap up with another one of my favorite quotes. In our relationship with Debian, one of the things that we were constantly reminded of is this quote by Margaret Mead, that one should never doubt that a small group of thoughtful, committed citizens can change the world for Indeed. It's the only thing that ever has. Thanks to all of you for being thoughtful, committed contributors to Debian. I look forward and my company looks forward to continuing to work with you very productively for the future. Thanks for your time and attention.