 And he's going to talk about Hewlett-Packard's relationship with Debian, so I'll make a next stage. Yeah, thank you very much. So Debian, it's really wonderful to meet you back at the end of the site, and I hope that sometime while we're here, we're going to have a site that does a lot of good effect. I'm curious, how many of you have heard me talk about HP and Debian's relationship in the past? A few, yeah, even a few HP folks, surprise, surprise. How many of you have never seen me in person before anywhere? Oh, hey, that's actually really gratifying. I love coming to the Debian conference every year in part because it's a place that I feel very comfortable. There are many people that I have known for many years and had the opportunity to work together with on interesting things, but part of what I have really appreciated the last few years is that in particular, this Debian day as a first day to kick off the official part of the conference has provided a neat opportunity for us to sort of step back and talk about some things relating to the Debian project and how it fits into the larger ecosystem of free and open source software, and that's, I think, gives us an opportunity to sort of help explain the project's place in the world a little bit, and I'm going to try and talk about that in a slightly different way than some of the other presenters today, both because I personally have a long history of association with this project and with the whole idea of free software in general, made my first personal contribution of source code back in about 1979, and I've been involved with HP or its spinoff agilent technology since about 1986, so I've sort of had the opportunity to be one of the people who lived through the evolution and emergence of the Debian project, and at the same time, I've been part of HP one way or the other for a very long time, and as it says here, I've been working on Debian since about 1995. Actually, I first installed it sometime in 1994, but my real involvement in the project began right at the beginning of 1995. I did serve for years the Debian project leader back in the 2002 sort of time in history, and I continue to serve as chairman of the technical committee as well as being president of software in the public interest, which we'll have an SPI session on Monday, I guess, we have a board meeting that we're going to try and actually have the folks that are here in person be meeting in person, and we'll be happy to take questions about SPI and answer questions about where we're trying to take that organization. SPI provides a legal and financial umbrella for the Debian project in the United States, which is important for being able to hold trademarks and things like that, and I also build pieces of amateur radio satellites and play around with high-powered rocketry and all sorts of other fun things, which I occasionally allow to sneak into my presentations. But you know, people ask me, so what exactly is it that you do as chief technologist for open source and Linux at HP? And I've spent a lot of time trying to think about the right way to explain that, and I think the easy way to do it is actually to show you three quotations from wildly different places. The first one's from somebody that I only quote rarely and usually with great mirth, but this time it actually is sort of interestingly appropriate, and that's this quote from George W. Bush about three years ago when he said that my job is to like think beyond the immediate. I read that quote, and I just had to laugh because this is in some sense part of what I think my responsibility at HP is. There are other people who worry about the budgets for the engineering activities that we're engaged in this very minute to try and deliver products in the next quarter. My job is largely to be looking at longer-term trends and behaviors and to help the company figure out what it should be doing to be in the right place at the right time to successfully participate in the larger free and open source software movement and to find interesting things to do. I also really love this quote from Evan Moglen from a talk that he gave at the Harvard Law School a few years ago, where in speaking about the free software world, he said that we now have a body of software accessible to everyone on earth, so robust and so profound in its possibilities that we are a few man months away from doing whatever it is that anybody wants to do with computers all the time, and that's really very neat. That describes free software as, in effect, an infinite range of possibilities, and then Dan Shearer, who I actually saw this morning for the first time and quite some time, threw out a quip at a conference in Australia a few years ago where he said there's so many clever things that could be done. And if you take this combination of things, trying to look a little bit beyond the immediate, this incredible breadth of things that can be done with free software, and then trying to pick the right set of clever things for the company to be involved in and trying to pursue as we navigate the free and open source future, that's really sort of what I do. But you know, enough about me, let's talk about the larger bits of HP. If there's only one thing that you remember from this presentation today, this is what I'd like it to be, hp.com slash go slash Debian. This is the sort of portal launch page, home page, whatever your favorite word is for finding out information about the products that HP has specific Debian support kinds of things associated with it. I'm now going to sort of jump away from this and talk a little bit about history, how we got where we are, what we actually do in terms of using technology from Debian and various products and some of where we'd like to go in the future. And then I'll wrap up by reminding you of this URL once again. So if you go back to the beginning of my personal involvement in the Debian project around the end of 1994, the main Debian server was a machine called ftp.debian.org. And this was a PC under a student's desk at a university. He happened to be a student who was working in the university systems administration community. And as a result, was no big deal really for him to have an open source project machine sitting under his desk. But there were problems associated with this, not the least of which was that every time Debian had anything like a release, all of a sudden that university's network connections would be totally inundated with file transfer requests from people all over the world wanting to download newer software. And the university networking folks would get concerned. And every time they got concerned and dug into it and found out that it was this PC under this guy's desk, questions would get raised. And so Bruce parents who was very involved in the leadership of the Debian project at that time was really concerned about the stability of the situation and asked me, since he knew I was managing a team inside the old test and measurement part of HP that maintained lots of systems and had lots of networking resources available to us. If I'd be willing to go grab a mirror copy of that system and just keep it sort of up to date and available in case something bad happened at the university and that machine got shut down or the connection got severed or something like that. And so I grabbed an old HP Vectra 486 Tower PC that was on a pallet about to go out for scrap and stuck a SCSI controller in it, an ESA SCSI controller for those of you who ever had the dubious pleasure of dealing with those things. And a couple of hard disks. And I love pointing out that at that time the entire archive comfortably fit on a 660 megabyte SCSI drive with lots of room available for growth and put a very early version of Debian on it. And that actually ran on an HP network connection from an HP data center in Colorado Springs for a number of years until we changed internal accounting rules for how we paid ourselves for the use of network bandwidth. And all of a sudden it became a high profile visible thing that maybe should live somewhere else. But I love the fact that today my management at HP likes having me talk about that part of our early history. Since of course at the time I just happened to be a manager running a group who sort of had permission to do things like that. It's not like it was some huge corporate strategic decision involved. And then if you're on the clock forward a little bit in that same period of time it had gotten slightly embarrassing that HP's PA risk processor architecture was sort of the last 32 bit CPU family that didn't have a real Linux port associated with it. And there had been lots of internal discussions in HP about ways to do this. But frankly a lot of individual contributors working in the company were worried that if they started contributing to something like a Linux kernel port that they might get in trouble inside the company particularly if they had previously worked on the HPX kernel or things like this and maybe had access to information that the company wouldn't want them sharing publicly. Then this little group of consulting folks in I think we're all in Ottawa, Canada called the Puffin Group emerged. And as a way of getting attention for themselves and sort of making a name in the world they announced that they were going to begin a PA risk Linux port. It helped that some of them were real aficionados of that architecture and knew a fair amount about the systems. But it was also sort of a brave and audacious thing to do because they were in effect diving in to try and port Linux to a processor architecture that wasn't completely well externally documented outside of HP. But through a happy series of coincidences and because of the amount of attention that this drew by early the next year and early 1999 HP had agreed to participate in the port and the moment that announcement went out there was this amazing effect on people working inside the company because now that HP had agreed to support the port of Linux to PA risk hardware individual contributors in HP felt free to go participate in the project and help make this happen. And there were some explicit things happened regarding permissions to make certain kinds of hardware information available and so forth that really aided that. And this was one of this was a major sort of stepping stone in the history of HP's direct involvement in the Linux flavor of the free and open source software community. We've been very involved in other free software things previously. We've been supporters of the Internet software consortium and their bind and DHCP demons and so forth for many, many years. But this is a really big deal in the history of HP's involvement with the Linux community. Not long after that Linux quit care acquired the Puffin group. And at that point, it wasn't just a fun project anymore. If work was going to continue, you know, somebody needed to get paid for it and HP actually agreed to put some money in so that tool chain and kernel porting activities could continue. But you know, you get to a certain point and it's like, okay, we have a working Linux kernel. We have a compiler and tool chain that work pretty well. But if anybody's ever going to actually use that, you need to have a Linux distribution around it, not just a kit of parts. And the reason that that distribution for PA risk ended up being Debian was that while on one hand, this port would be sort of irrelevant without a full distribution built from it, HP had already publicly announced the itanium processor architecture as its future direction for many of its systems and that that was going to be replacing PA risk. And as a consequence, it was it was really hard to sell a commercial Linux distribution on the notion that they should make a big investment in porting their distribution to PA risk because they looked to the future and said, well, you know, this architecture is not going to be around forever. You've already announced what your plans are for transitioning to something else. If we're not going to sell millions and millions of units, how would we ever make back the money that we would put into causing a port to happen? And so PA risk, the PA risk report was largely perceived as a community driven activity and not something for which there was a really strongly articulated business need. And at the same time, there were some key engineers in this group in HP who were Debian developers or for other reasons had gotten to know the project and liked it. This was actually before I went back to work for HP full time to do Linux things, but I knew a lot of the people who were involved and I would get invited as many other people would to participate in these hackfests that used to be organized around the PA risk activities. And at some point, it was pretty clear that if a distribution was going to happen and if it was going to happen in this kind of a community way, that Debian was probably a really good way to make that happen. And so when HP engaged directly with the Debian project to pursue the PA risk Linux port, there were some things that were very clear that this needed to be a real port. At that time, Debian had been ported to multiple processor architectures, but it had gotten nowhere near the state that it is today where more architectures are supported than by any other distribution. And so there were lots of questions about how this could happen. And it was very clear that we wanted this to be a real Debian port maintained and merged fully into the Debian mirror network. And this is accomplished by hiring some more additional Debian developers. And at some point in that process, I ended up being one of those people. But also by encouraging some existing HP engineers to go become more active in the Debian project and eventually to sign up and become registered Debian developers themselves. There are also some select investments made in specific features that were of interest. I won't bore you with the details, but there's actually a relatively long history of HP being willing to step up and put a few hundred or a few thousand dollars here or there to help make various things that needed to happen, happen. And then over the long haul of time, the experience that HP gained working with Debian on the distribution port for the PA risk processor sort of provided the context to make it easy for us to do Debian related work for our our itanium based servers when they were first being introduced and for many other things since then. The other thing that I think is really interesting is that the experience of building a relationship with the Debian project and figuring out what it took to be successful in engaging in a community oriented distribution activity had a lot to do with the attitude that HP's taken since then on how to approach the whole open source community development effort. It's a very conscious decision at HP to participate directly in open source development communities. In HP we don't talk about working with the community. We talk about being part of the community and as a result we think of ourselves as citizens of this larger community and we think about what the appropriate behaviors for citizens are who want to engage in social processes around this kind of community development. We try to support in line with existing community values and behaviors while focusing on developing those features and capabilities particularly enterprise robustness features and capabilities that are particularly important to our customers. And I think one of the really sort of interesting subtexts here is that HP has so far never felt the need to create a new open source license just for us. Many of our competitors in the open source business world have created new licenses and used new and different licenses that people have had to go figure out and understand when they chose to release interesting pieces of source code into the open source community. This approach of aligning and working with existing projects and making contributions to existing projects under the licenses they've already chosen has actually worked really well for HP. It means that we are one of the largest distributors today and one of the largest users of GPL version two within the enterprise world and that's part of the reason that we've taken such an active role in the GPL v3 process with one of HP's senior legal counsel serving as the initially as the co-chairman and more recently as the chairman of one of the discussion committees there. It's a reasonably significant commitment on the company's part to that process and while that process has not finished yet we are all you know hopeful that the end results will be good for for us and for everyone in the community and then we work with our Linux distribution partners Debian and commercial distributions and with independent software vendors to ensure that in the end we're delivering innovative and effective solutions that our customers can use. I went back and dug this slide out you don't need to pour it all the details but this is a slide that I took out of an old HP internal presentation that was used by the R&D manager of our internal Linux support organization to explain why it was that the Debian was such an interesting thing for him to work with and you know trying to decode pictographic things is sometimes difficult but the concept he was pointing to here is that when new hardware features are being developed and when new software functionality is showing up in the open source community there is a time delay between when new things get invented and when they show up in a commercial stable distribution release. There's just the normal cycle time of when a red hat or a Suzer or someone like that schedules their stable releases. When you work back from those stable releases there's some amount of time that's required to bring new software functionality in whether it's device drivers to support new hardware or new software features get those integrated and stabilized and prepared for a release and that there was a market opportunity represented by that time window between when new things first became available and when they showed up in commercial distributions and as a hardware company for HP it would be very valuable to be able to provide working Linux distributions and related tool chains to developers before our commercial distribution partners could necessarily get there with their completely stable releases and that there were opportunities when we were investigating and supporting new software to be able to deliver that faster if we had a lighter weight process for getting it out into the community and then there's this notion that as the market matures that gap might narrow because either the pace of innovation of new functionality would change or the rate at which things got integrated and smoothly delivered to customers might change but this was actually sort of another flavor not just this strong developer community relationship piece but a real sort of sense of there being a time value of money market opportunity window for making new functionality available by putting it on top of a community distribution that HP could be engaged in delivering just as quickly and efficiently as anyone else while we were waiting for our commercial distribution partners to run their normal cycle times. So in terms of things that have actually come out of this history of development relationship if we run the clock way back to when the titanium two servers were first being introduced they used a new processor chipset called ZX one that was developed at HP and it took quite some time to get ZX one support fully integrated into the kernel dot org tree and for all of the 64-bit isms it turns out that for various reasons titanium systems put more stress on some 64-bit porting issues than alpha and other 64-bit architectures had previously partially just because we were building systems that could handle gobs more memory than ever before and partially because that memory was often organized in HP servers in a sparse way across the address space where we actually ended up tickling higher order address bits faster and more furiously than some previous 64-bit systems but the really intriguing thing is that we had full ZX one chipset support and therefore Debian was installable and usable on HP's initial titanium two server offerings in the woody stable release and that release was actually made before any of those systems shipped. That's something that we have never succeeded in doing with any other operating system Linux or otherwise is being in a situation where the operating system is actually shipping in a stable release form before the hardware starts shipping. Now you can question the value of having something that's in a stable OS release before the hardware is even available and in fact our Debian leadership has established a set of criteria for adding new architecture support to the archive in the future that suggests that there needs to be hardware available that people can actually run this on before it's relevant but I point to this as an example of something that it was possible to do through the tight development relationship that HP was maintaining with the Debian project even at that point in history that would have been difficult for us to accomplish any other way. The first time that we set the I believe that was the first spec FP 2000 record ever set by Linux system was set using Debian on an titanium two server. We have of course since then focused most of our specsmanship work on the commercial distributions that we work with but the first time we set a spec FP record with Linux on an titanium two server was done with Debian before the commercial distributions were available with support for that chip set. There was a thing that I was actually involved in helping to invent which was an installation and software recovery tool set for titanium two systems that was based on Debian. It was actually a Debian boot floppies instance with the RAM disk expanded and the hood jacked up and a few extra tools slid in an extra layer of menu at the front end. But people discovered fairly quickly that you could actually use one of those. And if you knew the right magic words you could successfully use it to do a Debian base install. And so I was very pleased that in later versions of that we explicitly added do a Debian base install as a menu option. Some of the earlier carrier grade servers and HP's first entry into blade servers and various early access titanium two servers were all shipped with Debian pre installs. That was often the default Linux that was made available for developers. In almost all cases those products have now been replaced by other more mature products online and many of those ship with commercial distribution support these days. But there is still a substantial solution that HP has made frankly a lot of money selling into the telecommunications market that is Debian with an HP customized kernel and a few additional utilities running on top of a customized version of one of our itanium server models that's now helping to make things like cell phone calls just work all over the world. And then Debian's components have ended up being used in many products that HP ships that embed Linux technology as a member of HP's open source review board I have the ability to look into the database and review all the different places where HP has had our products intersect with open source and there are currently a couple hundred products shipping from HP that embed Linux and open source technologies of one kind or another and in places where that's more than just a kernel or more than just one application being loaded on some other operating system platform. It's surprising how many of those are using Debian as the source for picking up the packages they need. One really good example of that is in the thin client product line and you'll have to pardon the bill. This is a deck I pulled from. You can tell the ones that were put together in my marketing people but in fact a couple of the thin client folks are here. Michael's in the back room you know one away of Michael he's one of the architects from that team and he's actually here this week to soak up as much as he can and to I think we're hoping that sometime in this coming week some of this model the T5725 will actually be here for folks to take a look at play with look under the hood with and so forth. But these are sort of you know self-contained no moving parts desktop or whatever client computing systems that have Debian in their guts and support customization of the image of the 5725 some of the other more embedded products are also using Debian pieces in their guts but at the same time they are doing it in a way where at the end of the day that's not part of what the customer worries about it's just a product that works really well and solves a particular need and the fact that we happen to be picking Debian technology to make that happen is you know reasonably irrelevant to the customer but the reason that they've ended up picking Debian is the same reason lots of other people do it has a lot to do with the availability of all of these packages of stuff on any given day nobody cares about all 15,000 of those packages but the fact that there's somebody in the Debian ecosystem who does care about each of those packages and maintains it means that we all get to pick the subset of those 15,000 that matter to us and they're all there and they're all available. It also helps that there's such a strong international developer and community support experience around the Debian project because it means that the internationalization support in Debian is much better than it might be otherwise and of course it doesn't hurt that at the end of the day products that we build around Debian are able to be shipped royalty free. So let's fast forward to August of 2006. This is since the last time we had a Debian developers conference but it's not the first time I've talked about this in public of course. HP announced support for Debian on our ProLiant server family and they're in fact also here this week a couple of folks from our ProLiant server organization I don't know if they want to smile and wave or not but they had a lot to do with these cool shirts that I'm wearing and your organizers are wearing today in which I hope many more of you will get a chance to see before the end of the week. This was I think a really significant moment certainly in the history of the Debian project because you know this is a first tier server vendor offering real support for Debian on its principal family of servers that are of interest to Linux customers and part of the reason that that matters is that HP ProLiants have been the market leader in worldwide x86 stuff for a long time for about a decade actually and in fact for the last nine years since they first started counting we have been the clear and undisputed leader in both number of units and revenue in sales of Linux servers worldwide as a company. So when HP decided to add support explicit support for Debian to this product line it's a big deal and you know again taken from one of our marketing decks but I think worth mentioning anyway when we talk about industry standard servers there's industry standards and then there's things that help to differentiate quality and value within that concept of standard and part of what makes the ProLiant stuff interesting is that HP's made some significant investments in technology everything from acoustics and thermal management to driving density and better cabling and all sorts of interesting common components across the different product lines that allow us to drive volumes up and therefore get high reliability and reasonable economics across a wide range of form factors and different models and when this Debian enablement activity started there were really sort of two things that were going on one was this notion that we wanted to make sure that if you were already a customer or an organization that had decided to use Debian that there were no reasons that you would want to say well you know we have to go somewhere other than HP to get this because you know it's not being officially supported but then at the same time we recognize that there were lots of customers that were also in heterogeneous environments who'd already made decisions to use some of HP's commercial management and support offerings and we wanted to make sure that it was possible to use Debian systems well there so it's sort of a two-pronged motivation one is if you're a Debian user we want to make sure that there's no reason that you don't also want to be a ProLiant user but at the same time if you're already a big HP customer we want to make sure there's no reason that you aren't comfortable also using Debian and so there are all sorts of value-added components that come as part of this product these are sort of normal things that the ProLiant folks do and what it really comes down to is making sure that all the device drivers and management agents and so forth are available and tested for the combination of these different systems and Debian and one of the things I particularly wanted to point to is this thing called Insight Control Linux Edition for those of you who follow the events in terms of mergers and acquisitions and so forth a while back HP acquired this company called RLX which was one of the first companies in the Blades Marketplace and they had a product called Control Tower which was a management interface for managing Blades the reason I wanted to mention this is not only is Insight Control Linux Edition an interesting tool to use for getting a management view of a bunch of Linux systems but it's actually based on Debian Insight Control Linux Edition is a Debian, mini Debian system with all sorts of other things running on top of it which folks don't always notice it's another one of those examples of a place where this is the right technology for putting a product together that works well and solves a real problem and at the end of the day it's not for the end user about the fact that this is Debian but for those of us in the Debian community it's really cool when we notice that the stuff that we've worked on for all these years is helping to solve yet another interesting problem and so there's all sorts of things that have happened there's been a process of getting through the qualification testing with Debian Sarge using two, six kernels that's pretty much complete now and effort has transitioned to working with etch I understand that you should watch our website starting probably sometime in July to start seeing etch qualification results showing up that's just the typical lag time after a new release that it takes our testing organizations to get things scheduled and done it's not that I'm aware of anything there that's a problem many of us are actually running etch quite happily on these servers today and have been for some time but I wanted to point out that even things like the ability to do firmware updates on the servers from a running system and other things that aren't always part of sort of the base process of making a distribution work on a box or things that we're trying to support and so if you want to know what's available today I did a little quick cut and paste of bits and pieces from the web page this morning doesn't look exactly like that in other words but if you go to that hp.com go slash Debian there are a bunch of places you can from there click off to and one of them is a support matrix that will show you for different models of HP ProLiant server what pieces are and aren't available what things are coming soon and so forth there's another site that I wanted to really point out and again this presentation slide set will be available for you to download or look at online later if you want you don't have to scribble nodes but there's actually an HP ProLiant page within the Debian wiki structure which is an effort at aggregating community information about these systems as well HP folks contribute to that but others within the Debian larger Debian community are more than welcome to contribute experience things there as well and there's various other stuff that's downloadable from HP and then beginning on the 1st of December last year part of this whole support offering is that it is actually possible to buy from HP the same kinds of support products that are available for other Linux distributions and for other operating systems on these hardware platforms and to buy them specifically for Debian as well I'm not gonna spend any time on this but in this presentation deck are these slides that have the specific part numbers that you can buy from HP to get those different kinds of support if you want to you can find all of those from the hp.com go Debian website if you follow that particular path of mouse clicks and rodent droppings but I thought I'd stick them in here to make them easier for you to find and I mentioned earlier this HP product which is Debian GNU Linux plus HP Telco extensions I wanted to mention this again just because even though you may never actually physically see one of these systems because they're mostly being sold to communication service providers we're putting them into functions and their core networks and as a result these are not the kind of things end user see there's about a 30% chance that any cell phone call that you make anywhere in the world today is somehow going to touch one of these systems in some way I just think that's pretty cool I think it's cool that Debian is a core component of how many of the world's telecommunications providers are now building and deploying services to them it's not Debian per se that's important it's the fact that there's a large number of nines high availability Linux based solution available on hardware that meets their expectations but along the way somebody noticed and discovered that Debian's distribution was a good thing to build that on and that it would work well and be able to solve the problem and as a result there's an awful lot of it out there so once again if there's only one thing that you remember from my standing up here today other than HP's done some kind of cool shirts it's this HP.com slash go-debian this is the place to go to look to see what we're working on and what we're doing there are lots of other HP externally facing sites like opensource.hp.com where you can get a wider view of the various activities that HP's engaged in the open source community but when it comes to what we're doing with Debian and how that's turning into specific product and support offerings that are relevant to our customers and they are relevant we've sold some fairly interesting support contracts to customers who are running Debian and just last night I got taken out to dinner by a customer who just bought oodles of our blade servers to run Debian on and is looking forward to spending some more time this week talking to us about how we could get that stuff even better and so you know this is a good place to go to find out that sort of information if you want I'm gonna close with one thought I you know as someone who's been involved in Debian for a long time and tried to provide some vision and thought leadership at various points there is a challenge that we have faced through this whole process and that is that Debian doesn't have a certification test suite this kinda makes it tough for companies that are used to working with commercial distributions to figure out exactly what it is they need to do to quote-unquote support Debian and I think it's this work that Ari and Vandevent at Intel's done on the Linux ready firmware developer kit has been really interesting in this regard this how many of you are aware of this have you heard of this developers kit before anybody other than Keith okay well I'm glad I mentioned it then this is really neat this is something you can go download and run on a system and it will tell you whether that system's BIOS is doing things the right way to support Linux or whether there are issues and at the end it's basically an automated set of tasks and gives you a report this has been amazingly empowering to BIOS developers at hardware companies all over the world because now all of a sudden they can go oh that's what I need to do there's no longer any sort of funny questions about what does it mean to build a system that's gonna work with Linux they've got a simple thing they can run and get the answer so it makes me wonder whether Debian should engage in offering some specific additional tests to that suite if that's something BIOS developers and so forth are already going to be running and using and it also caused me to wonder from a sort of Debian PR kind of standpoint if it would be worth creating a Debian branded derivative of something like that toolkit with more tests that are specifically relevant I don't know this is an idea I'm actually throwing out more to the Debian developers in the audience today than to those of you who are just here to figure out what we're all about but this is the kind of thing that I think could help to make more companies that are involved in providing hardware and support services that might be interesting to have as part of the larger Debian community and ecosystem in the future understand what they need to do to be able to do that so with that hopefully we have a few minutes for questions thank you for your attention does anyone have something I'd like to ask and I have a, there's a mic in the center of the room and I have this one I can hand out either so go ahead so could you summarize the different ways that H... Oops, hello, check so could you summarize the different ways that HP makes money with Debian? Yeah, so how does HP make money with Debian? A lot of it's indirect a lot of it has to do with the fact that by using Debian in an embedded product for example if the alternative was to go to some other Linux distribution or some non-Linux operating system where we would have to pay some amount of upfront money to acquire the software or where we would have to pay royalty then there's an operational expense improvement by not having to pay that by using Debian in the case of these support offerings it's that we hope we're selling more units of hardware than we would otherwise sell that people will buy from HP who wouldn't otherwise have bought from HP because this is something we've done we'd love to directly sell more support contracts if people are putting Debian servers into real IT production use we're having somebody clueful who's willing to answer the phone 24 hours a day any day of the year anywhere in the world would be a useful way to raise that customer or that IT shops confidence in using Debian to deploy that and that could be a valuable thing and worth paying somebody like HP for and then in places like this Telco thing it's a key component of enabling us to pursue a market opportunity that we might not have been able to afford to pursue otherwise that's an example of where we went to our commercial distribution partners and we said, hey, you know we see this market opportunity we're gonna go build this custom piece of hardware to go pursue this opportunity it requires building a kernel that has a bunch of stuff that the kernel.org guys don't really like and yet, you know, we can make money on this do you wanna go there with us or not and they said, well, feel free and if you're successful maybe we'll follow you there someday so those are the kind of places where working with Debian can actually bring real money into the company even if it's sort of indirectly. So obviously GPL v3 isn't finalized yet but do you have a sense as to HP's corporate direction regarding v3? Well, I think there are things that are concerning for a company with a large patent portfolio in v3 and we have raised our concerns through the process and frankly I'm very pleased that the language in the final draft that we've seen the final call draft is better for us than some of the preceding ones but at the end of the day I think it's gonna be another interesting license in the landscape I don't think we're gonna have any strong reason to try and drive people to adopt or not adopt the license. Frankly our history has been largely with GPL v2 and unlike a lot of other folks we've been very comfortable using v2 so some of the pressures that drove people to want to create a v3 are pressures that I don't think we really felt and yet if the community wanted to pursue a new license we thought the right thing to do is to try and participate actively and have that license in the end be as useful and as good a license as possible. I will comment that I think there was this hope in the beginning of the v3 process that somehow when we got to GPL v3 that there would be this wonderful unifying consequence of revisiting the terms of the GPL and I'm afraid that what's happened instead is every time some specific issue came up for discussion people who previously were willing and able to work with each other within a community even if they had differences of opinion on that particular point. When that point came up for explicit discussion now people had to take sides and decide how they felt about that and so part of this process has actually been a series of divisions where lots of people who sort of work together under the GPL v2 umbrella have had to say yeah I don't know if I'm gonna like this better or not and I'm hoping that eventually a lot of those sorts of divisions get rehealed that once the license is stable and solid and people start to use it and understand it some of that sense of community will reform around it but I think it's way too early to tell what's gonna happen there and of course the final call draft we don't know I hope there are no further substantive changes in the text before the final but this is the FSF and RMS we're talking about so who knows. That's right. Are there any thoughts about taking HPUX and making it open source? There are thoughts about that on a regular basis and I would not expect it any time soon. I mean to be honest with you that we find ourselves in this interesting situation where we have a market leading position in a Linux market and a market leading position in the Unix market and there are actually different markets with different expectations and that would be a whole other talk and could be very interesting but it's not one that we have time for today since they're waving hooks at me from the back of the room so no I don't expect that to happen anytime soon and frankly personally I'd rather see our Linux offerings just continue to grow in relevance and both financially and technically till that question's just not relevant anymore. Okay well thank you very much for your time and your attention. I will be here all week. In fact my family's joining me on Tuesday sometime and we're all gonna be here towards the end of the week so if you have other questions or would like to ask me more about the history of either the WN project or those parts of HP that I've been involved in feel free to find me and ask I'd be happy to chat with you. Thank you much.