Discussions

News: Day One of JavaOne 2007

  1. Day One of JavaOne 2007 (17 messages)

    By Frank Cohen Day 2 and 3 Coverage The opening keynote session was a reminder of where Sun has been over the past ten years. On the stage at various times were Jonathan Schwartz, Scott McNealy, James Gosling, John Gage, and many others. This is a good time for these veterans to bask in a little bit of glory because of the problems Sun has had from the bad times, losses, layoffs, and bad technology decisions. Two years ago I asked Jonathan why Sun continued to pour money into NetBeans when they were laying-off long time employees. This year things look up, for at least a little while. I felt generous towards Sun, and I believe Sun feels upbeat itself. For instance, there was a lot of banter going on between Jonathan Schwartz and Rich Green, executive VP of software at Sun. Rich tried to get Jonathan to use the word "people" instead of "consumer." These two people seem to like each other. John Gage also announced Sun's effort to make the conference carbon-neutral. And the morning keynote included announcement of an United Nations collaboration on an effort to establish an "Engineers Without Borders" service effort modeled after the successful "Doctors Without Borders." The upbeat feeling is making me feel good too. No skeptical questions from me this JavaOne! Rich kicked things off this morning by talking about goals for the Java platform: 1) faster, faster, faster, 2) Make Java more relevant in deploying applications that people want to use, and 3) expand the tent to let other developers onto the platform. Rich announced efforts to streamline the JRE code base, make the JRE downloads smaller, and make Java more modular to increase deployment flexibility and reduce deployment size. He also discussed the JavaFX consumer focused family of Java technologies to provide scalable rich interfaces. At first, JavaFX appears to be an answer to Adobe Flash and Flex, Microsoft SilverLight, and Oracle WebCenter (part of Oracle Fusion Middleware.) But that doesn't quite hold. Sun has advantages over these others: a rich object-oriented programming language with lots of supporting libraries and 6 million developers that know the language, a runtime that is already deployed to billions of mobile phones and desktops, and a consumer/service architecture that supports XML messaging between consumer and service. These are big competitive advantages in favor of Sun. Of course, the downside to the JavaFX announcement is how last-minute, and thrown-together, the presentations appeared to be. Missing are GUI development tools for people not wanting to learn JavaFX Script but you do get yet another functional domain specific scripting language, and a way to reuse the user interaction already built for JSF, jMaki, and the other Ajax initiatives at Sun in a JavaFX application. Rich said they are focusing on getting the JavaFX Script language right first. They will roll out a roadmap for tools around JavaFX sometime in the future. Rich said their tool goal is to make sure that the average Flash developer does not need to learn Java 6 and JavaFX Script to be productive creating JavaFX rich interfaces. Rich said JavaFX is a binary runtime that Sun will license to network operators. Look to http://openjfx.org which redirects to the dev.java.net site for Eclipse and NetBeans plug-ins to develop JavaFX scripts. The opening keynote covered many news topics: - Rich introduced Martin Harriman, vice president of marketing and business development for Ericsson multimedia. Martin told attendees that Ericsson wants to increase adoption of IMS. Ericsson will contribute IMS technology to Glassfish to make the Sun Java System Communications Application Server. Combined with JavaFX the resulting integration could be excellent. - JSR #1, the RealtimeThread, is now available and in the Java VM. RealtimeThread replaces java.lang.Thread to provide real time services. Anna Ewing, CIO for NASDAQ, talked about how vital RealtimeThread is to NASDAQ. - Sun announced two open-source related topics: Completion of the open-source Java, named the OpenJDK. Sun also appointed the interim governance body for OpenJDK, including Doug Lea, State University of New York, Fabiane Nardon, CTO VIDATIS, and Simon Phipps and Mark Reinhold of Sun. OpenJDK is open source under GPL version 2. When asked about portions of the JDK that include contributions (for example the Java 2D libraries,) Rich said Sun has contributor agreements from many of the contributors to the OpenJDK and where there is no contributor agreement there are binaries. - Sun also offered that they would make it easy for the OpenJDK project team to use the Java Compatibility Toolkit (TCK.) Sun's adoption of open-source still seems colored by corporate interests. For instance, Rich said GPL forces everyone innovating with Java to share the improvements, Sun will encourage its maintenance, GPL will cover the co-ommission errors, and TCK takes care of omission errors. Sun's focus is around building a single OpenJDK open-source community that discourages branches. I wonder if this will also discourage innovation. For instance, Rich gave a completely nebulous answer about making the TCK open-source. The TCK is Sun's control over who gets to call their technology Java. Rich later said in a press conference to more-or-less "stay tuned" on open-sourcing the TCK. From my perspective this non-answer on the disposition of the TCK is the last task Sun has to really deliver an open-source Java. One cannot be a little pregnant, you either are or you are not. Yet, Sun seems to want control over who is designated Java and who is not. If Simon Phipps was correct in yesterday's unconference, then the OSI should be shaming Sun over this TCK non-decision. Busy Day for PushToTest I had a busy day showing off the new PushToTest TestMaker Version 5 to a lot of people. I got a lot of questions from attendees on building unit tests with the JSR 223 script engine driver and turning unit tests into scalability and performance tests. I also found many people that are already using the new soapui 1.7.1 utility that is integrated into PushToTest Version 5. The Java.net folks hosted my demonstration to attendees and they recorded a podcast of the demo that will appear on the Java.net site. I also did an interview with SysConTV on the new test automation platform. Keep Java Pure There is a general nervousness on the topic of keeping Java pure. It used to be that the developer community could count on Sun to protect the core Java platform from attempts to pollute and derail Java, in the way Unix splintered. Now, Sun's support and adoption of dynamic scripting languages makes the community nervous again. As one attendee phrased it "the scripting languages seem be gratuitously different" and this could be a way for the Java core platform to be made more fragile. Rich and Jonathan answered the question by saying "People are gratuitously different." Jonathan said including dynamic scripting support works to that diversity. I don't think they realize the nature of JSR 223 ScriptEngine's use of native calls to make this work. Jython compiles to Java ByteCode on the fly so everything happens within the VM and HotSpot Compiler. Most of the other scripting languages are called through a JNI call as an external process. This has the potential to create memory leaks, extra code to realize calls from Java objects to specific interfaces (methods) and data transformation errors (Java HashMaps to Ruby dictionaries for example.) I credit Sun's complete turnaround on dynamic scripting language support to Tim Bray and his minions, including Charlie Nutter and Tom Enebo. Consider James Gosling in March 2006 saying "None of [the dynamic scripting languages] attempt any serious breadth in the application domain and they both have really serious scaling and performance problems." This morning James in the keynote session said, "What makes scripting languages so useful and charming is their ability to do things like JavaFX Script does." Sushi Everywhere, at least Maki Arun Gupta demonstrated jMaki running on the Java language, but jMaki also runs on Ruby, PHP, and Phobos (server-side JavaScript). Next week Sun is rumored to be showing jMaki running on Ruby On Rails (RoR). Look for details at http://blogs.sun.com/arungupta and a screencast of the demo he presented today, look for "#web1". JMaiki is deployable on Glassfish v2 and v3. Bob Brewin, Distinguished Engineer and CTO, Software, at Sun Danny Coward, Java SE Platform Lead, at Sun led the afternoon technical keynote. Bob described research into packaging and deployment. For example, at development time look for something new called superpackages, and at deployment time look for superJARs. Both are using OSGI as the primary format that Sun would like to be interoperable with. Bob introduced Charlie Nutter to show JRuby on Rails deploying the Mephisto Blogging application to Glassfish application server. Thor showed JRuby integration with NetBeans 6 by showing Ruby code completion, integration of Java objects in Ruby scripts, and finding code source. Bob then focused attendees on the new NASA World Wind geospatial Earth browser. World Wind is similar to Microsoft Earth and Google Earth but written in Java and open source. The demonstration showed the WorldWindowGLCanvas component in NetBeans 5.5 receiving event notices to handle user selected information (like city locations) and connected it to List selection listener to receive mouse clicks. Details and downloads at: http://www.simulation.com/javaone and http://worldwind.arc.nasa.gov One presenter on a lark suggested to Bob that a Solaris port for ARM would mean that JavaFX Mobile would allow anyone to build mobile applications that could be run on a variety of mobile devices, some based on Solaris kernel and others on the Linux kernel. Michael Kolling demonstrated the Greenfoot teaching environment. Greenfoot is a learning environment to teach object programming to kids. Michael's example talked about how his students use Greenfoot to create games. The operating environment is like Ruby On Rails in that everything is available to the user but all the complexity is hidden. Greenfoot's runtime is available for free, it is not open source. Next, I'm off to the Tangosol party... perhaps the last Tangosol party at JavaOne now that they are wholly-owned by Oracle. More tomorrow! -Frank

    Threaded Messages (17)

  2. Re: Day One of JavaOne[ Go to top ]

    Good summary, Frank...
    At first, JavaFX appears to be an answer to Adobe Flash and Flex, Microsoft SilverLight, and Oracle WebCenter (part of Oracle Fusion Middleware.) But that doesn't quite hold. Sun has advantages over these others: a rich object-oriented programming language with lots of supporting libraries and 6 million developers that know the language, a runtime that is already deployed to billions of mobile phones and desktops, and a consumer/service architecture that supports XML messaging between consumer and service. These are big competitive advantages in favor of Sun.
    I actually think that Sun doesn't necessarily have an advantage in this area. I think Flex is a great programming model, combined with AS3, which allows you to use the Flex declarative model transparently with AS3 is awesome. But besides the programming models, Sun failed in this area because of the their distribution model. Flash is installed on 98% of browsers and the reason for that is mostly the ease of installation and their distribution model. Flash plugin seemlessly installs in most environments without having to download and install a JRE outside of the browser. ActionScript is actually not very far fetched from Java itself and should take a very short time for any java developer to master. We recently used Flex/AS3 to provide a front end to a stateless SOA backend architecture and I loved it so much that I doubt I'd will willingly any time in the near future use any other technology for front end development. I think the web 2.0 AJAX technologies are still relavant for any application thar requires a internet facing front end for search crawling, etc..., but for anything else, I think Flex serves its purpose and provides a great programming MVC model. It's also very relavant with the advent of stateless SOA architectures, since you now basically have a stateful front end GUI that connects to stateless backend services. IMO that scales very well. Ilya
  3. Re: Day One of JavaOne[ Go to top ]

    Also, I don't see much wrong with the TCK and Sun keeping some sort of control of Java. I'm all for open sourcing it, branching it, and see more development in this area. I'd much rather though see a more viabrant central java community with open source developer participation. I'd hate to see 20 different java clones that are all incomplete and have some benfits, but more drawbacks.
  4. The downside of control[ Go to top ]

    You make very good points Ilya. I see a downside to anyone trying to retain control over an open source community when that control is used to squash innovation. There is a philosophy behind the open source license that makes it difficult to stay in control when the community wants to go somewhere that the leadership doesn't. For instance, look at what has happened to business integration with Sun. The community could have developed JBI years ago without getting encumbered with Sun's consumption of the SeeBeyond purchase. -Frank
  5. Re: Day One of JavaOne[ Go to top ]

    I actually think that Sun doesn't necessarily have an advantage in this area. I think Flex is a great programming model, combined with AS3, which allows you to use the Flex declarative model transparently with AS3 is awesome. But besides the programming models, Sun failed in this area because of the their distribution model. Flash is installed on 98% of browsers and the reason for that is mostly the ease of installation and their distribution model. Flash plugin seemlessly installs in most environments without having to download and install a JRE outside of the browser. ActionScript is actually not very far fetched from Java itself and should take a very short time for any java developer to master. We recently used Flex/AS3 to provide a front end to a stateless SOA backend architecture and I loved it so much that I doubt I'd will willingly any time in the near future use any other technology for front end development. I think the web 2.0 AJAX technologies are still relavant for any application thar requires a internet facing front end for search crawling, etc..., but for anything else, I think Flex serves its purpose and provides a great programming MVC model. It's also very relavant with the advent of stateless SOA architectures, since you now basically have a stateful front end GUI that connects to stateless backend services. IMO that scales very well.

    Ilya
    I have not got the point of JavaFX as well. If JavaFX can be integrated with a browser, then demoing it with JNLP standalone apps is not the best advertisement of the technology. If JavaFX cannot run on currently available Java plugins and requires 15MB library instead, then it already is dead. Flash plugin is small and works everywhere, Avalon aims on seamless integration of desktop and web, where JavaFX aims to? Used in a standalone JNLP app, JavaFX appears as a mere replacement for Swing. I don't see the point, especially after a decade of arguments that Swing is good because it is all code-based and does not use declarative definitions of menus, windows and other OS-specific resources. Surprise, surprise, now it seems that Sun turned 180 degrees, and resources, in this case XML-based, are good. Hypocrites. On the other hand, Flash is not a standard component of web, it has problems with accessibility. Also, I am not buying the concept of fat client Flash-based app as long as browsers have refresh button. Even Google Finance sucks in this regard, they use Flash object to show stock charts. If you refresh the page, your settings and time range are gone. Until browsers have standard and universal way of making browser app stateful in-client, that is, standard ways of controlling back/forward/refresh functions and a standard API for storing client state after browser finishes a client-stateful browser app will remain a hack.
  6. JavaFX...[ Go to top ]

    the flash runtime is less than 1MB ,silverlight is 1.5MB and this small plugin download is a key requirment for wide deployment on the browser side,java JRE6 is about 15MB or more this is without the JavaFX runtime(interpreter or compiler) can Sun released a JavaME for JavaFX deployment ? believe me:NO END USER WILL DOWNLOAD A PLUGIN OF 15MB
  7. Re: JavaFX...[ Go to top ]

    the flash runtime is less than 1MB ,silverlight is 1.5MB and this small plugin download is a key requirment for wide deployment on the browser side,java JRE6 is about 15MB or more this is without the JavaFX runtime(interpreter or compiler)
    can Sun released a JavaME for JavaFX deployment ?
    believe me:NO END USER WILL DOWNLOAD A PLUGIN OF 15MB
    Very true, as I mentioned the popularity of flash and its ubiquity is due in large to its distribution model. There are other reasons of course like the awesome macromedia tools that eases the pain of developing with flash, but nothing overshadows the distribution.
  8. Re: JavaFX...[ Go to top ]

    believe me:NO END USER WILL DOWNLOAD A PLUGIN OF 15MB
    Probably Java is already on user's computer when the user got the computer: I believe many or most of computers shipped with Java. Sun just needs to distribute additional JavaFX, or JavaFX will be part of Java, pre-installed on the computer. Wei Jiang J2EE tools
  9. Re: JavaFX...[ Go to top ]

    believe me:NO END USER WILL DOWNLOAD A PLUGIN OF 15MB


    Probably Java is already on user's computer when the user got the computer: I believe many or most of computers shipped with Java. Sun just needs to distribute additional JavaFX, or JavaFX will be part of Java, pre-installed on the computer.

    Wei Jiang
    J2EE tools
    That's not the point. What happens when a new JRE comes out? Still have to reinstall the 15MB worth of needless stuff. Flash runtime is small and upgrades are very easy. Within 3 months of flash 9 release, over 80% of machines were upgraded. That's in fact due to the wondeful seamless distribution model. When I speak with someone and am demonstrating a flex app, if they don't have flash 9, the browser asks them to install it, they click ok and within a minute it's done. When a user doesn't have the right jvm or one at all, I'm in for a half an hour of headache. Ilya
  10. Re: JavaFX[ Go to top ]

    Ilya, you point is clear. I just wonder when Sun is talking on re-packaging of jre it might solve the problem of size and limit dependencies. Are there any details on that? Gennady
  11. Re: emergence of Glassfish[ Go to top ]

    It would appear that the full weight of a legit enterprise vendor is being corralled in to the initiative that is Glassfish, as WebLogic continues to lose its raison d'etre for Sun accounts, even JBoss has some uphill obstacles to being deployed on Solaris, it would seem that the Sun sales force will now be going in to accounts with Sun Java System Application Server 9x and +... Beyond development tools, where something like Spring can always come up with something to compete, there is the deployment platform that JEE has perfected. Distribution, standards, frameworks, tools, apps, and customer loyalty are all embedded in the app server market, and Glassfish is now the vanguard of this effort. It is now a race to the ultimate prize of v.3 of the deployment scenarios, where v.1 was ISV deployments, v.2 was custom integration, and v.3 will be standard integration of v. 1 and 2 via SOA... Who will win? If you believe the reports out of NPR and Ian Cringley, it would seem that IBM recognizes that their bloated GS org. is inhibiting the WebSphere org. from competing effectively, and will need to re-invest in product development or else GS will become experts in deploying Sun's app server, and lose margins in the process. BEA has AquaLogic. JBoss will do well on Linux, and Oracle will do their installed base... But only Glassfish intrigues at this point, with both an OSS model for developer inclusion, as well as the influence of the Java franchise behind it, and a concentrated R&D budget, plus acquisition options, plus JBI to make this competitive battle less than fair. A SIP App server with Glassfish container comes close to closing all gaps that have been outstanding for 4+ years, especially considering the teleco market is hugely important to Sun-proper... It's been 2 years since Glassfish was launched and it is pretty clear now that it is in a very strong postition, irrespective of dice.com job postings, to make inroads in to the app server installed base, alongside JBoss' model on Linux; the question will be who gets to lead the portable apps model that will be key to SOA inter-enterprise efforts, and how that technically shakes out...I have been looking forward to this for 5 years, and it is finally coming to fruition...
  12. JavaFX was F3[ Go to top ]

    JavaFX was formerly known as F3. Checkout cool demos in Chris Oliver's weblog. Adi
  13. Not the last Tangosol party ..[ Go to top ]

    Neelan and I have already agreed to pay out of our own pockets to keep the Tangosolarmetric party going if we have to .. .. but Oracle's already trying to get it in the budget for next year :-) Peace, Cameron Purdy Tangosol Coherence: The Java Data Grid
  14. A prince among partiers[ Go to top ]

    The Tangosol party was terrific last night. I just wish the sun was not so bright today. A few newbies came along with their boyfriends and asked me to tell them who all these people are. I told them, "These are the super geeks, the freeks that are changing the world." Thanks for brining us together Cameron. -Frank
  15. Re: A prince among partiers[ Go to top ]

    The Tangosol party was terrific last night. I just wish the sun was not so bright today. A few newbies came along with their boyfriends and asked me to tell them who all these people are. I told them, "These are the super geeks, the freeks that are changing the world." Thanks for brining us together Cameron.
    It's a great industry to be a part of, and I hope the next seven years are as fun as these last seven :-) Peace, Cameron Purdy Tangosol Coherence: The Java Data Grid
  16. Re: Day One of JavaOne[ Go to top ]

    I see Java guys huddling for an answer to WPF/e(Silverlight). Are there any works in Java community to leverage the new Engine microsoft provides for WPF ? With WPF/e(silverlight) Microsoft did its homework and Sun seems to have hurried it (Vaporware?).
  17. I don't see the JavaFX as vaporware. Sun has a track record with the cell phone network operators and this improves their offering. Midlets are nice, but the interaction of JavaFX interface is appealing. The way in which Sun presented JavaFX makes me think they were opportunistic at taking advantage of the timing of JavaOne. However, it would be a shame if JavaFX bumped off the schedule the backend integration technologies (SCA, JBI, ID server, ESB, etc.) None of these were present in a big way yesterday. -Frank
  18. About Silverlight[ Go to top ]

    A couple of remarks: 1. Silverlight is similar to Java on the desktop/JavaWeb start: you get a full CLR runtime with a limited set of base classes that are mainly focused on XAML/WPF. 2. It is only available for Windows and MacOS (granted, this is the majority of clients, but a lot of Linux users are left out). 3. It is NOT open source - the only part of Silverlight that is covered by the Microsoft Permissive License (BSD-like) is the DLR and IronRuby & IronPython. These *require* the CLR to be present. 4. Silverlight is still in its very early alpha stages. Best, Gerald http://blog.beuchelt.com/