Home

News: James Gosling Unleashes on Java 1.4.2 Users

  1. In this audio clip and commentary taken from Gosling's keynote address at TSSJS 2011, James unleashes on Java 1.4.2 users, describing that particular version as "the bane of my existence."

    So what should people still deploying on 1.4.2 do, according to Gosling?  "Get with the program. Upgrade! If I could have I would have pushed an update out that would have caused it to just die. But so many people would just not let go of one four two."

    Read the full article and check out the audio clip. It's a short two minutes, but it's moderately amusing:

    James Gosling Unleashes on Java 1.4.2 users.


    *********************

    Richard Mayhew's Java Version Survey Results

    James Gosling's Blog

    TheServerSide Java Symposium

  2. So what should people still deploying on 1.4.2 do, according to Gosling?  "Get with the program. Upgrade! If I could have I would have pushed an update out that would have caused it to just die."

    While you were at it you could have made those hundreds of deprecated methods disappear (the ones that irritate many many people, and many of these have been deprecated since 1.1), oh and that awful attempt at a Date class ...

  3. Funny you should mention the deprecated methods.

    I had a chance to speak with Rod Johnson from Spring fame, and asked him to contrast his approach to aggressively deprecating methods, with the Sun approach of letting them linger on forever. Basically, his opinion was that deprecated means deprecated and they should be yanked out of the API as soon as is reasonably possible.

    About the Date class though...Yeah, just the mention of the Date class is enough to start a flame war.

    I'll get the quote/discussion/audio of Rod and his response to the deprecation of methods up next week.

  4. Indeed, the simplest solution is to just upgrade the darn thing to at least version five. And if you're really adventurous, jump right ahead to version 7.

    THis has got to be a joke. Whoever concocted it apparently still believes in WORA and perfect backwards compatibility, both of which have been shown years ago not to exist. For some enlightenment, head over to a Java forum (like JavaRanch), and spend some time reading about the problems of people stuck with Java 1.4. Here's a little preview: not upgrading has nothing to do with an unwillingness to do so.

  5. Indeed, the simplest solution is to just upgrade the darn thing to at least version five. And if you're really adventurous, jump right ahead to version 7.

    THis has got to be a joke. Whoever concocted it apparently still believes in WORA and perfect backwards compatibility, both of which have been shown years ago not to exist. For some enlightenment, head over to a Java forum (like JavaRanch), and spend some time reading about the problems of people stuck with Java 1.4. 

    I'm not sure. I've worked with quite a lot of teams that "simply" continuously upgraded their stacks. A particular large enterprise application I was involved with started with 1.3 and pretty effortless upgraded to 1.4 and the point updates after it. 5 was a little more hassle, but nothing truly out of the ordinary.

    Taking advantage of all Java 5 features is something else of course, but the basic migration was relatively quickly done. Java 6 was effortless again, and tests with the latest builds of Java 7 (140 and 141) seem to be just fine actually.

    I do agree that if management doesn't give you an inch that it's going to be difficult, but I've seen many engineers sneaking some of the migration work into other work item. E.g. work on module A, test whether it all works on both the current and the next version. If it doesn't, make sure it does and write the extra hours off on "working on module A", instead of "migrating to new Java version, new AS version, etc".

  6. The main reason an upgrade from Java 1.4.2 is denied by management is $$$$. Their mantra "If it ain't broken, don't fix it" still rings true to them. The fact that lots of systems still running on Java 1.4.2 (even on WebSphere 5.1, which refused to die) are still running fine actually speaks of the robustness and pervasiveness of Java technology in the industry.

    However I do like the idea of an update to Java 1.4.2 that would 'kill' it, forcing them to be upgraded to a better version. But then the supporting systems (app server) would also need to be upgraded, so is not 1-to-1 switch. Anyway, this 'update' can cause a disaster that management will take notice of and they will scour the earth for the funds to upgrade ASAP LOL...

    The irony of reactive management...penny wise, pound foolish...

  7. Post-1.4.2 Java ME?[ Go to top ]

    Hmm... Is a post-1.4.2 Java ME stack available yet?

  8. Just switch to new version and solve your problems, managers are thinking about $$$, we've no problem in a very large enterprise system etc. These approaches are not good.

    The first one James Gosling approach smells like marketing. When you face some problem in any system and call the related company and they can't understand why this problem happens, they just say "Upgrade to new version and it CAN solve your problems". But the problems that you may face during upgrade is your responsibility to tackle.

    Manager are just thinking $$$. In the world of open source everything is free and a few computers and software that requires a few thousand dollars not a problem. Hey wake up. This is business. We're not playing in the sand like children. Will you work for free if they don't give money for a few months? Of course managers and boss think about money and profit. But employees must also think about that. If some system is working but there is a small bug in some small function why spend a lot of money just for a new version? You can buy new tools and equipments for your members.

    And the last one, "no problem in our system". Did you ever heard about SOX or ISO, project cycle etc.? Again this is not child play, it's a business. In our case, one minute linestop due to a system or any other problem is equal to 15 thousand dollar loss for the company. I doubt even you've developed an enterprise system. You should test every function to make sure that the new version will not cause any problem.

    People like James Goslin seems very cool when they talk. But they don't know business and they've finished a very big company like Sun. Just go and play in applet sandbox.