 For starters, let me just say that last year at Wasta Peck, I had no idea that the day after I left Wasta Peck and went to Java 1, that Jonathan Schwartz and Rich Green would announce that Java was about to be open source. In fact, 99% of Sun employees were caught completely by surprise by that. But we were very excited about it at the same time. And then that started several weeks and actually months of debating about how to open source Java and what license to choose, which ultimately would lead to our early launch on November 13th in which we announced that the license for Java was the GPL. And we released the first parts of the code, the Java hotspot virtual machine, the Java C compiler, and Java help and some other pieces, and had a broad roadmap for open-sourcing the rest of the platform. And then throughout the beginning of the year, we had a chance to talk with many people in the community at Faust N, then again at Fisle, and then at Java 1 this year, which was in May, we did the full launch of open source Java and released the remaining parts of the code for which we had copyright and could release the code. And so now that gives you sort of a brief history of open source Java. And with that, let me turn it over to Simon to say a few words. So, hi, I'm Simon Phipps, and I'm afraid the bar is closed. Stand over here and serve up what I can. So, you've got to be a man in these slides. They're great, but hey, doesn't matter to me. So, I joined Sun in 2000. I was hired at IBM where I just spent five years introducing Java into IBM. I was part of the team there that introduced Java into the company. And so I had an outsider's perspective on Java for a considerable time. I was hired into the company by the then executive vice president of software who said to me, look, I'm hiring you so you can get Java open source in 2000. So, you could consider me an extreme under performer until 2007 to get it completed. The deal that I've had with Java for a while is to look upon what's going on with, what has gone on with the Java platform as one of the many parallel processes in free software. If you look at the history of the broader free software movement, there have been lots of parallel activities that they'll branch off. Some of them do actually happen in parallel with each other. The use of BSD in the way that it was licensed in the beginning of the 80s was actually completely independent of what Richard Storm was doing over largely because of EMACs. And those two things were both encapsulating the spirit of their age. It turned out that Sun was party to the BSD side of things. It was very much a party to the view that things work best when everyone can get at the code. It's best if you don't mess too much with the rules to how they touch things. And as time went on inside Sun, that BSD kind of spirit has been a very strong thread in the company. The way Java was set up in 1995 was as close as people who were involved could make it to a BSD spirit except they were in a market context where there was an extremely hostile and dangerous monopolist who had a known practice of finding key technologies and embracing and extending them in a market place and using their market power to snuff things out before they could get going. And so Java was set up with all the source code available and with a very liberal set of rules about what you could do with the source code and with a license that was carefully crafted to make sure that it couldn't be embraced and extended when it's a completely open camp with bear traps that were around who were in the outside. And that was all happening in a context that was well clear of the commercial popularity of free software in the DUPL and FSF sensor free software. That really rose to a summit at the end of the 90s as Mozilla proved you could actually take a piece of code and that the world didn't end when you licensed it under free licenses. Now Sun screwed up at the end of the 90s because actually there was a mutual pop-up but we can discuss that later. We didn't keep track of where the free software communities were going with Java because we were too busy dealing with the success of Java. We had loads of companies coming and asking for licenses to the Java technology but it had succeeded beyond their wildest dreams and we really didn't have the scope to go fix the licenses. It was kind of priority 12 or 14 or something like that. There was no actual hostility to free software. There was just the sense that it was something that we didn't have time to deal with because there were really very few people employed by the company to deal with it and there were lots of very big corporations pressing in. Now my feeling is that in an ideal world it would have been free software a lot sooner. I ideally would have liked to become free software in 2003. I think that was the earliest point that we would have been free software. As it turned out it got debated inside the Sun every year. Every year it would be discussed, considered and would nearly make it into the Enantus for Java 1 until somebody in the last minute would say no you can't do that because and we finally got to 2006 and the context changed. Lots of people went into Java 1 2006 assuming the debate had happened and that the story was over and that it wasn't going to be free software and then to everyone's immense surprise our chief executive announced that it was going to be free software under the GPO as part of his keynote address and set the line for what was going to happen. So in the last year the Java team has been very hurriedly working out exactly how to do what they were told. The essence of this is that Sun genuinely wants the source code that it's worked on before the Java platform to be free software under the GPO and has told everyone to go do it and as frequently happens as things go free software inside Sun the group that had been vigorously opposing it before suddenly become the chief advocates or it worked that way with Solaris. All the pushback that there was about making Solaris into free software and then turning it to advocacy on past the goal and past the finishing line. And the same thing has happened with Java. We've now got people vigorously advocating free Java and we've talked with the chief executive here. So Tom does describe the state that we're in at the moment. The Java platform consists of a core piece of Java, Java SE which is the Java runtime and then there are two other what used to be called profiles of Java there's Java Mobile and Java Enterprise. All three of those have Sun's implementations of the standards available under the GPO B2 and where it's smart to do so we've added the class part exception to the GPO B2 so we're exactly mirroring the great stuff that people have done that I bought and marketed with class part. It turns out that so we've got all of all of Java EE we've got all of Java all the code we wrote for Java ME release with Java SE the Java runtime it turns out we bought code from a bunch of people down the road and they didn't give us sufficient rights to make code open source and about 4% of the code Java SE is under we have got the terms that we are subject to don't allow us to place it under a free software license and so that work is now available with binary lumps and we've put the Java SE code under GPO with an exception that allowed it to be combined with temporarily with these binary lumps while we and others get on with the work of reverse engineering the code that is in the binary lumps and making those free software as well and so Tom can comment on the time scale for the binary lumps and I'd also like to hear from some of the people who've been working on the binary or at least the focus ability to work on binary lumps and that's why there's people out of the board here and also Andrew is over there well let me throw back to Tom I'm happy to talk around the issues of Java being made into free software the thing that I want to underline for you is that it is some's intent but some's implementation to be released as free software and that we create around it a community that is independent and that is fully capable of driving that code in whichever direction it has to go and we are happy as a company to be a peer member of that community rather than insisting on being the benign ruler of that community and that's the direction that I'm intending that I'm dedicating to Tom So, with that, just to elaborate on what Simon had said for the code that we don't have copyright for we can't releasing the GPL we have these binary plugs that you can download today if you go to OpenJDK.java.net you have access to all of the source code and instructions for how to build and so currently what will happen is you can use an alpha version of JDK 7 to complete OpenJDK to have a fully functional run time today however that fully functional run time of course is entirely on free software unfortunately so this is really intended as a stopgap measure and just to be clear we really want to make OpenJDK 100% free software and 100% functional and when we do that then we will no longer have to have these binary plugs so what are these things? Well, there is a color chooser there is a font rasterizer there is a color rasterizer graphic rasterizer there is a little bit of crypto code and a few ancillary parts audio yes and so one of the things that we've been hoping for and very excited about is working with the class path community because we're license compatible to see if it's possible to leverage some of that work and collaborate together so in the past year I've spent a lot of time with Mark Villard and meeting a number of class path people and talking about what they have and so what basically has happened is the class path developers that have skills in these areas have offered their code for review for son engineers and so my role is to try and make sure that the son engineers review that code as quickly as possible so there's a couple of other things that you need to know in terms of infrastructure we have the site today which is opengdk.java.net from that you can get the source tar balls we also have some mailing lists one of the things that we do not have yet is a source code control repository Java is actually a very complex system comprising something like 6.5 million in lines of code and we have about 10 major integration points so it's a distributed version control system and unfortunately right now we're using a proprietary version control system and so we wanted when we open source Java to use an open source version control system and after a great deal of investigation the team has decided to choose Mercurial as that solution so that happens to be the version control system that OpenSolaris and other groups have decided to use also within son and currently I suspect that our timeline for publishing the Mercurial repositories will be sometime later in the summer I don't know exactly when that will be we also one of the other things that's fairly complex is we need to expose our bug tracker we have currently a way of taking bugs over the web through the website but that's there's sort of an intermediate step to get into our internal bug tracker and our internal bug tracker also kind of be exposed to the outside so we're going to choose an open source bug tracker and figure out how we can do the integration with all the back end tools that we integrate with the bug tracker within son and let's see what else do I need to tell you oh something that's very important is it's son as a goal of making sure that our commercially licensed version of Java the stuff that you've always been able to use from Java.Center.com is as close to identical to OpenJDK as possible we have no interest in maintaining a proprietary version of Java that is somehow different or has secret sauce we want to have effectively as much as possible the exact same code in both code bases however as you can imagine the only way that we can do that legally is if we hold copyright on all of the code that is in OpenJDK so as we solicit contributions from OpenJDK we're asking developers to sign the son contributor agreement which basically is a non-exclusive copyright assignment to son for your contribution and what that means is you get to keep full copyright rights to your code and you give son the right to have copyright on your code so that for example if we were to choose at a some later date to upgrade the license to GPL v3 we would have the right to do that and it also gives us the right to continue to publish our Java under the commercial license that we have now so as far as Java OpenJDK now and in the future at Java 1 this year we announced an interim governance board that will be set up to handle issues within OpenJDK until such a time as we can hold open and fair elections to replace them with a permanent governance board the decision was to have two members of the board from inside son and two members of the board outside of son I'm sorry three have frauds outside son I clearly haven't had my expressway in the two inside son are seven tips who you see here and also Mark Reinhold who is the chief engineer for the Java SC platform and on the outside of son there are three people one is Doug Lee who is a researcher at the University of Oswego who has done a lot of contributions to hotspot especially around concurrency you may have heard his name another person is Fabiane Ardon who is a lives in Sao Paulo she's a Brazilian woman who is a CTO for a Java EE related company and has been very active in the Java tools community and Java.net and also in the greater Java community in South America and the third outside member is none other than Delwar Tapich who is here the leader of the cafe community and so with that Delwar could you share with us your thoughts about the interim governance board and what you see happening in the next few months well that's going to be quite interesting because we basically just started talking about those things effectively Java long was just going on when the whole thing was announced and as you can imagine the sound guys spent the last months just trying to get all the source code out for the release so I'll take the Java one basically I think Mark went in location and so so with the interim governance board is going to do is to just set up the constitution and the general framework I think it's going to have to be lectures to have for the real governance board in the future and it's terminated basically to be done with all that in the next year or so and next so we should be done with the Java one at least I hope so it's going to be interesting because it's not the first time we've started doing this whole thing but I'm going to say I think it's Solarix they're all doing this whole governance board procedure there and it's going to be interesting and really good to have someone tips there interim support for Solarix community to see his experience to see how we can do things different better get it better this time get it better this time you never know and it's always interesting because the whole open dedicated thing in its breadth is different from Solarix in a way because coming from the classical community I know a lot of different people to have a piece of it there is researchers there is the regular runtimes there's time guides there's spare time makers there's professional guides using this stuff so we're in class but for the last five or so years we have managed to create a huge network of people who work collaboratively without sharing always the same goals or working the same project all the time last time we raised the back phone and in a similar way OpenJDK tries to attract both projects just want to deliver a job runtime but also academics and even alternative limitations of job if they could all do the same last time by its design it's going to have to be tailored to a suitable logic community just a product specific one that's kind of an interesting challenge to deal with regardless which I hope we can deal with but be very liberal very textured so that's that I'm coming from the Catholic class but things so we were guys who used to kick son in the nuts over the past couple of years but very clearly it was really interesting to see now getting to know people inside the company changing conversations a bit and always the change of the information we could kind of see that they were trying hard in the last year to get this done and we could work better much much more effectively because all Marvel was communicating with Mark Miller and coming from it all the time it sort of when Andrew Haley and Redditor set up ISD which is this experimental project around open dedication to work with plus but it's a substitute for the binary plus the people inside knew we were going to do that and we made sure that everybody was going to have it well let me just say without facility without here 4K all these very sensitive issues because one of the main challenges when we started talking about the social 5 years ago for real the plus but community was to overcome this huge fear of 4K in Java land that the open source Java is going to break everything it's going to be chaos, it's going to be drama the end of the world and it took a lot of effort on the class side and the deviant side and Fedora side and so on to make sure people understand it it's actually good for something there is a need to package and so on so in that respect it's important when we do things like ISD that we walk a tightrope and try to fall down with respect with more conservative members of the Java community are kind of looking at this whole thing with a bit of a fear in the shapes I would love to hear about ISD let me just introduce Andrew before he says the word just to say that last year as many of you know we got some Java introduced into the non free archive one of the DLJ license which was I think a really great thing for deviant but we knew we know that getting into non free is not nearly as effective as getting into main because in order for many Java packages to really survive and thrive into in main they must have all their build depends there in main as well and so to do that we had we knew that we needed Java to be open source and Fedora doesn't have the concept of non free so this wasn't even an issue but one of the motivations for open sourcing Java was to make sure that we could get Java into the main repositories for deviant and Fedora and in fact that's one of the reasons that I'm here again this time at DevCon is I think that that's very important and you know I've finished in that the Java development organization is very big and there are a lot of processes that are changing as this culture of hundreds of developers change to open source it's socially a very complex thing to do and one of the things that's complex is for them to appreciate the speed or the velocity of how you guys work in the open source world one of the things that I think is great about the ISP project is using the full of freedoms under the GPL it's possible to as the ISP team has demonstrated it's possible to actually bootstrap and build a portion of OpenJDK which is 100% GPL and does not have any binary plugs whatsoever and that I think is an amazing technical achievement and I was also you know gets us on a path now where we can begin to complete those plugs and then I hope eventually gets us solutions that we find in the ISP project back into OpenJDK and to say more about that let me introduce Andrew Haley thanks we've been using Java I suppose it goes back to around the 21st century something like that and it all started with GCJ the Java head of time compiler and I've been maintaining that for many years and I'm not sure how many we merged our library code base with that of class path a few years ago and that was a tremendous success so we had a reasonable implementation of the language in the library that was good enough for a lot of the stuff that we needed so when some announced that they were going to open source Java I was not particularly pleased to be doing that because that's my job and also if I'm completely honest I didn't believe it I was absolutely sure that some would release Java under some kind of brain damaged license that would be semi-free and just enough semi-free that we couldn't actually use it and we couldn't actually link it with all the code that we already had and there wouldn't be anything we could do with it anyway so there were several meetings that we had and I said to him, well I'm not going to forget it it ain't going to happen we just got to get on with our lives GPL was rather a surprise however I'm still here I haven't died of a heart attack and when the OpenJDK was released with these binary plugs I felt somewhat vindicated in my skepticism but we realised fairly quickly that although this was a great thing to have we couldn't actually use it because there's no way we could shift these binary plugs as part of the Dora so we were going to need to replace them as quickly as we could now there's two things you can do you can either technically approach it simply stumbling out things you don't have you moving the calls replacing dummy methods or you can replace things with functional equivalent for free software and we've actually done a combination of two I don't think we have anything suitable to replace sound for example and the graphics that we've actually got in the past version at the moment is kind of semi there so you can't actually do much in a way of graphic implications but the important thing about ICT alternative made for this in its current state is that it's good enough to build itself in a reproducible way so that each version of ICT can build an expression and so on we're building all to begin with using GCJ as the sort of bootstrap Java runtime environment which is plenty good enough to do a very routine job of rebuilding the JVK itself and so now we're not quite sure what we're going to do I think we're going to carry on improving the quality of ICT and making some of the replacements more functional and trying to work with sound to get a completely broken JVK in our code there's a few legal niceties to get sorted out as well to figure out things like there comes to be true agreement there's quite a substantial problem in that the class path code is owned by the Free Software Foundation and for some to integrate into their code basically would have to be re-licensed to them so there's lots of stuff going on to try and figure it out we really don't want to be in a position where we end up forking the thing so we're going to pass that code so we're going to try and avoid that unfortunately I can't really tell you about time scales whether this is going to be ready but I would personally be quite disappointed if something that's suitable and was reasonably functional wasn't in Fedora 8 not that that would introduce anybody else here interest anybody else here but that's sort of time scale what is the difference what's the difference and we did actually we've just changed the contributory agreement that we're using for open dedicated other son projects so that it is now possible for people who are the authors of FSF owned code the FSF gives you a grant back of your copyright for you to assign to others so if it contains fine ownership it gives you the right to then grant it to others we've changed the contributory agreement so that if you have that grant back there's lots of people who all own the copyright and it means that no one has a problem of being a licensing dead-end yeah it ought to be possible to get something sorted out but really until it is all sorted out the code that we're using to make nice two work isn't going to go back into son's code base so I'm just going to have to like we might be able to sort it out we'll sort it out so he's great he's the perfect guy to talk to because he's very patient he can deal with RMS so that really gives you an overview of the state of the coffee garden I just I'm pinging Keir Magnusson here as well to special waves and say there's also another independent documentation had been done on the passion by harmony but unfortunately I don't have a status update for you on that one I'm trying to find out oh the other thing I should have said is that the status of the Java trademark is such that you can't be Java until you're complete and you've passed the technical compatibility kit and we're still working out how to make all of that work and run the kit on open JDK so that the last the versions of Java actually been shipped with free software can officially be called Java we've been calling things GCJ and class path and things like that in the past and because we weren't allowed to call Java because we couldn't get our hands on the compatibility kit and even if we had our Java wasn't really complete enough to pass it anyway we're hoping in a few years time few months time, few days time that we're going to get our hands on all of that stuff so we'll be able to actually have free Java back to you that would be great and that's our goal too is to make sure that the implementations that are used by people can be branded Java compatible it's very important for us and then you do an update for Debian as you know Mandy, my colleague has been working on a meta package for Debian that includes all the build dependencies for ICT and one of the things that I hope that we can do with Mandy this week is to get an initial package for ICT proposed for Debian so we can begin working on that in the Debian world as well and one of my other agenda items for this week is I'm hoping to in anticipation of ICT no project eventually going to Maine I'd like to try and package as many cool applications as possible I have really two targets for packaging one is I want to package the refactor build dependencies for NetBeans and Glassfish because as you may know we, even though those are open source projects they tend to include a lot of dependent jars within them and that really doesn't match very well with Debian policy or Fedora policy for that matter so we're going to try and refactor those things out and have separate packages for them and have interdependencies and what I'd like to do is to help Sam with his goal is package up some sexy Java apps and so if anyone is interested in collaborating that would be really cool Are there any? I actually have a demo I have a very low end Dell laptop and I only support GLX 1.2 but if anyone supports GLX 1.3 I can show you Wonderland which is a 3D virtual world that currently is under a non-free license but will very soon I'm told be available under the GPL which if you want to think of it it's basically a client and a server for a free version of Second Life with the exception of Second Life that in Second Life you design elements that you can never take out of the world here you'll have full control over the world The answer is yes there's actually quite a lot of crew apps that's used in Java and we wrote it so that Nebius is a Java and Nebius is extremely cool it's very responsive and fast and it's 100% Java so it's interesting actually looking at that and people not realising that it's written in Java because it doesn't suck so it kind of fools you for a bit and you do realise after a while that so other questions are there any plans to thoughts of the end to more of the... oh this is a great question thank you for asking yes yes yes ok so the question is are there plans to support the end to the platforms and the answer is yes one of our goals of OpenGDK is ubiquity and what that means is more operating system, chip architecture combinations and we actually haven't completely decided on the technique but we have some code for new ports within Sun that we're hoping to release soon that will help the community get started on that a couple of specific ports and as a result of doing that we're going to define a policy for community sponsored ports so that we can do that and PowerPC is not surprising like one of the most abandoned Linux PowerPC is one of the most abandoned ports it's actually worth reflecting on for a little bit about the way the Java world worked prior to now the way that Sun made a lot of its money from Java in the late 90s was by licensing the source code commercially to companies that wanted Java on their platforms so the versions of Java that you see for example on macOS 10 are ones that Apple licensed the source code from Sun then Apple is not important and that means that it's actually Apple that owns the rights to that VM not Sun and having just because Sun has made this source code free software that doesn't mean that Apple's reworking of that code has made free software as well and so I couldn't comment on that so because I'm not supposed to identify any parties that I've had negotiations with but I can relate to that to end that discussion I have a different one on general terms I've spoken to quite a lot of people about their work on the Java platform where they're willing to re-license things that are proprietary that extend from our code and generally the answer has been no however I think that there's some interesting observations to make about macOS 10 one of them is that they use the new tool chain to build macOS 10 and I think that Apple is Java and has a very interesting version of Java so it will be an interesting interesting discussion and interesting people to watch but I mean the thing that's worth reflecting on as you see these things people say why isn't the version of Java on AIX free software and now you do build it why haven't you built the version on AIX well we don't know the version Java on AIX it's based on our source code so it wouldn't be a big deal for the DIV to be made GBL so that you could then build it from the version that is in the OpenJP community but IBM owns the version on AIX Apple owns the version on for PowerPC that's on iOS 10 IBM owns the version for PowerPC if there is one I think they've got a PowerPC version that they use so the answer to all of these questions is because of the way that Java was licensed in the late 90s it's fascinating with a forking discussion. Java is already forked all over the place fortunately they're all compatible forks and so it turns out that Java is the same thing in most places, two within a degree two within a margin of error and a helpful discussion to have with proprietary companies is when you see you have Java over there we know that it's based on Sun's code why don't you make it free software and I can have that conversation directly with companies but if you have contact with them as well it's good to have that conversation because ultimately I don't get the phone as to whether IBM makes its implementation or Apple makes its implementation free software now if they start, if these companies start using the GPL version it doesn't have viral nature and it forced them to have to accept it again I'm well trained not to use the V word but yes okay so basically I like to when they talk about this I like to smile at these they talk about the popular property of GPL but that's the reason why so one of the questions that's been asked about the governance for fngdk is why aren't any of the big corporations on the government and the answer is because none of the big corporations are in the fngdk community either how do we think of gear is going on I think it will be my connections went down so are you saying that basically Apple if they start using your GPL version they will be forced to give away their GPL they would be but the we will continue to make the code available under the existing licenses that all the licenses have got it so they have a backdoor they have a backdoor if they need to they don't have to jump to it now we would much rather everyone move to the fngdk community because it actually reduces our costs for that to happen and it also means that the other companies have an innovation so it seems to be in their interest but there's lots of companies out there that seem to think that the fngdk is bad for them so what about if you just disabled allowing them to license it or anything with the gpl well the difficulty there's two difficulties with that first of all it would result in our asses going straight to court and because we already have existing contract arrangements that don't run out for quite a long time and secondly it would mean that so there's a huge opportunity here for the free software community to leverage other platforms we don't have any other hands dirty because javroot actually is everywhere it's on cell phone it's on os10, it's on mainframes it's on old layers it's on windows and that means that it's actually possible to write free software that you are doing within the context of W and yet for it to have a user base that extends way beyond that and acts as a funnel dragging people back to the place where they were designed for I think there's an incredible opportunity for us as a free software community to begin to influence people that we've not been able to reach before for just simple mechanical practical reasons which were that they wouldn't switch within us and so therefore we're going to listen to any other arguments either, thank you I think we've got an incredible opportunity to begin to let 10 rules go out into platforms that we've previously had no influence over through the Java platform and so because of that I'm actually quite keen not to go piss off Apple yet because I think it's really good for it to be possible to write an app that is just intended for Debian but turns out to be usable on Apple but the people who want to read about its status and its update find themselves coming to Debian to do so I think the integration act of that funneling is really powerful for us as a free software community so I don't really want to cut them off just yet and I can't because they take my house in the cold and the thing is since Java has been commercially available for all these vendors for more than years I guess those business models have evolved very, very slowly so all of a sudden you have this new way doing things that's just the new machine which they can take or that they don't have to if you look at GCC in general it seems to make sense but chipset vendors do actually work on improving the performance of the thing on their chipsets and so on so it's not to predict what the future will bring there are different models where those chipset or oversets come to figure out what their optimal thing is it's something that I want to continue competing with their proprietary versions because they have proprietary enterprise servers where the 5% performance increase tends to be a lot less than 1% I don't know some other people may just say well we're making chips we want Java to be faster than chips so we don't want people to buy good Java that reminds me I heard Sam mentioned in his talk this morning about RML as an architecture I'm not familiar with that is there anything coming with it? are you okay? I'll change the ABI for the R50 R because the old ABI was all right but it had the fundamental flaws so you had an FPU back in the 80mph so there was a hauling inefficiency if you ever actually tried to execute any FPU instructions so you tried to avoid that and you couldn't mix you could compile things efficiently with GCC but there was an enormous overhead to that you can't mix the two sorts of code they're just incompatible so the new ABI lets you mix the old and new well with FP code and emulating FP code on the same stuff so better and there's various other things which are better although we don't have a floating point double representation which is different from everybody else so so it's just a better ABI really and we've changed the kernel ABI at the same time because since it was designed ARM changed from being a a combined Asian Instruction set to a separate one Cache, sorry, Cache was working by and now separate will go around so what were the users being for ARM on? so the point is that unless you have a very old ARM chip it's likely the same as old ARM did with that processor and you can now choose to use FP stuff if your CPU has some support for it so basically it's based on ARM except for at the moment it doesn't support the four instruction sets which is what you use on strong ARM chips install all the strong ARM chips out there and we can theoretically support that but we haven't done it yet so until we do old ARM is going to hang up just thinking for a moment this is not the following reason so what is the status of java ARM? seriously you don't have a port and you might want someone to do one it's basically go around that's kind of a leading question well I'm not aware that we have a port we've done them in the past I don't know the status of the code but there's something I want to look into please contact me it has a new one because if we don't we need one clearly just a little point on that subject once java is free enables the community to do fast ports perhaps at the same time as doing gcc I'm pretty sure that the first java compiler and runtime to be shipped working on the itanium for example was gcjet because we did it we did gcc2 we did it before any of the chips were actually available and we got it out of the door very very early today as soon as I believe Linux was running a tour on itanium the gcjet port was available and actually having open jvk ought to enable that for any new chips that people are watching Linux to they should have something running the java implementation using the open jvk out of the box on day one rather than having to wait in some cases I might be a year I'd say before the official java port was not available one of the pieces of code that Gary Benson is looking for is for us to release is the a really old and now out of date i864 port but it has the c++ jit that would be really helpful for bootstrapping new ports so that's one of the pieces of code that would help in the operating effort as soon as we can liberate it I mean you might talk about ports like that as being but if there isn't enough community will to write a full hotspot port then it doesn't get done but who cares you know you've got java working it's not super fast but it's going to be good enough for an awful lot of things that people need to do and it's going to be tested, it's going to be official it's going to be from vatom and again which is an independent SVM it's fast enough unless it's heavy so application what is the issue with porting at all assuming that the base publication is living in C or c++ just needs to be compiled well you'll be pleased to find that in the infinite wisdom of the java development team java is actually a build-up for java and that makes porting a challenge J.C. Jones the rest of you, yeah and that's actually that's literally how ICD does it is G.C.J. comes to the rest of you the way that you know I actually tried to do one of these I did get it super far but the way that you start is with NFS tricks where at the certain points of the build process when you need the JVM you do that on a machine that has one and just do that on an NFS system such that the right bits end up in the right place and you switch back to the native target so porting is kind of oddly the first time around but this is I think going to be only tricky for this period of time because at the point that we get you know we can use G.C.J. and we can use IST especially if we get this new C++ porting should be substantially easier G.C.J. is a little somewhat bustier now you know why I'll tell you why it's bustier now there are some patches that are in the debut there and nobody has ever submitted an upstream to me they've been in there for years and they well mark some of these patches to me so I can put them in upstream and we should just sit there and do it there yeah maybe some little shit like that is it popular or is it just that nobody has done the effort to integrate into the mainstream nobody has ever explained it to me I've been very tempted to simply apply them myself but they are too substantial for me to do with our copyright assignment so I can't touch them until somebody decides who the hell wrote these patches for debut in the first place but the stuff I've seen fix it would be a little bit of a doubt and it's probably the right side so that should be alright ok so let's go exactly that's from 2012 thank you just had a good question on a different topic maybe the current some GRE in the UK depends on the Lucida forms well the plans in terms of the OpenGDK released in terms of fonts well this is more of a general problem which is Java one of the challenges with Java is that in Linux is that it doesn't use system libraries for things for example fonts times on database are two kind of painful examples and the reason is that there are promises that it works the same everywhere and so classically instead of the Java mentality has not been to depend on the platform to provide things in a consistent way it's sort of like static and linked well there are open fonts that you can ship from all the places well I think that in fact I'm pretty sure that in OpenGDK we already have some open fonts or I know that that was discussed the name is something that's already been thought out from the OpenGDK community but if you have cash around fonts you definitely use it there's just the same potential, the issue here is to create consistency you either have to implement it inside the GRE or you have to implement it in the loader code inside the GRE that fixes the different ways it's implemented on all the platforms and the path of least resistance was to implement it in the GRE that's why there is a time zone database that's why Java had a problem when the US decided to arbitrary changes when summertime started this year because well Java had a time zone database that knew when summertime started and it was wrong and it's a pretty intractable problem actually because every platform calls time zones different things truth them differently has the database in a different format and either you have to have thousands of lines of code to fix it so it looks the same while every platform is different or you implement a database and then it's different from everybody else's I don't think anyone has come up with a really elegant solution to that so that's the lucid problem with the rendering subsystem but the fonts itself but the lucid I think was picked a long time ago before anyone even was aware that there was the potential of a free font problem and the difficulty with this a lot of things is it is and making it other is actually non-trivial in all places that are not there but I guess to make it to me it wouldn't help to depend on a free font but anyway the font problem is on his radar his payment here depends on solving it so I don't know specifically about the Demian Java packages the ones that you mentioned earlier but to say it was a bittersweet day for me when I saw it go in sweet because we got Java finally bitter because I had to read the package licensing on install and that was forced and I have hundreds upon hundreds upon hundreds of servers we've already planned for this yeah I need to get rid of that and it was specifically in that package and I couldn't get rid of that well that's what I'm pre-seeding this for that's why you had to read this but maybe we're doing something that's exactly why we had to use DevCon so that we could pre-seed it it should work we were very strong in those servers the same way it just seems to be very brain dead that if you're trying to do a mass installation of course we knew that it was designed to use DevCon so the organization could have said once and then use DevCon but look well clearly you haven't documented that for a long time hopefully it's only a short time yeah I just wondered what the relationship on going is like to be between GCJ and Sun's Java let's ask Andrew well is GCJ going will this not GCJ on the head will GCJ go from strength to strength as a result does it go to begin with I didn't know what to do and I sort of still don't really there's the layer between the Java virtual machine and the Java renderer isn't actually defining this specification anyway so it might be something not one that I've ever had a chance to look at so if you're very tempting to say I only use GCJ we'll simply take more of Sun's class libraries and just drop them in unfortunately this would have more or less exactly the same effect as say picking up a Tokyo telephone exchange and dropping it in New York resulting in something completely malfunctioning so I don't actually know I know there are some people using GCJ in some particular application areas in which it's particularly well suited so I'm hoping that they will keep going there whether I will keep going on the Linux desktop I don't actually know that really is up to all users it's sort of around my hands I mean if people wanted to keep going then they will do is the obvious question of my time and the time of other people at Red Hat and how much time will Red Hat be prepared to keep you know, spending money on people to keep going I don't really know that we haven't really decided because we don't know what the bright new world the free Java is going to look like and to be honest with you we're not going to know what it's going to look like until we've actually started shipping the version of OpenJMK in Fedora and Debian and then we'll take a look and see how much interest there is in keeping GCJ going a synthesis of some of the best qualities of GCJ and some of the best qualities of OpenJMK in some of these the synthesis is potentially very similar integration with GCC system compared to C++ which GCJ does very well you know, you can use normal deep rubbers if you don't need Magic Java deep rubbers you can use system profiling tools there's so many things that GCJ does very well but it doesn't have the ultimate speed the hotspot has with all of this magic run time in line and stuff like that is it possible to create an interesting technique with synthesis of these two approaches yeah, absolutely isn't there enough time or enough to do it I think we're getting towards the end are there any final questions maybe just a quick comment sure so I work with grid systems and we do house research I was just going to echo the big thank you for making Java actually come up in non-free because it was originally a pain to put it in manually on all these nodes and accept the license and redo the Java home path it's great when it's in there so you can be waiting for the OpenGTK version of the next iteration for those of you who are preceding or big installations I'd love to talk to you later and get the full story because that really fascinates me I'd love to hear what you're doing I'm going to be here all week I'll be here for another week I say this every time I do a talk but if we do something really stupid at some I would much prefer the first time I hear about this by sending an email to ombudsman.com rather than by reading about it somewhere else I'm having it for the second place you say it somewhere else but I'd really like if the first place you say it is to ombudsman.com because that way I can stand a snowball chance in getting it fixed before it becomes an issue so ombudsman.com is your friend and is there to help you on the other end of ombudsman.com is a member of my staff whose job it is to read the email work out if it's confidential and if it is anonymise it I would find someone in Sun to act as the advocate for that problem and that person inside Sun then behaves as if they have your problem as far as all the rest of the company is concerned so that no one knows whose problem it is and there is no chance of retribution if there's any concept and so there is actually an active advocate in some company I don't promise that we'll get everything that goes to ombudsman.com solved the way you would like it to be solved but I guarantee everything will get a reply when we consider that in general that works not that Sun does anything stupid Paris is solved well great, well thanks after coming