Announcement

Collapse
No announcement yet.

Oracle Is Looking To Offload Java EE To A New Steward

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • #21
    I think it's a bit premature to call Java, one of the most widely used programming languages in the world, dead or dying. Most software developers may not like the language, but it's about as close to dead as C++... which is to say, not at all.

    And Kotlin, Scala, Clojure, Ceylon, JRuby, Nashorn (Node.js for the JVM) all extend Java's life along with the life of the JVM because they all can call Java code easily.

    As I mentioned upthread, Oracle's Graal project takes the JVM a few steps further and will probably prolong its life. The currently open part of Graal is http://openjdk.java.net/projects/graal/ the whole project is at http://www.oracle.com/technetwork/or...iew/index.html I'm going to summarize it as I understand it, so I may get some details wrong. I got most of what I know from the video at https://www.infoq.com/presentations/polyglot-jvm-graal

    For anyone that doesn't feel like wasting 40 minutes on the video:

    1. They want to make an intermediate representation of programming languages in the JVM, so it's easier for more languages to be ported to the JVM, for more of the non-Java languages on the JVM to get near Java level of performance, and also easier for languages on the JVM to interoperate with each other. Right now, for example, it's extremely easy to call any Java code from Kotlin, Scala, Clojure, or JRuby but you have to do extra work with your Kotlin, Scala, Clojure, or JRuby code to call it from Java. Graal reduces that work to call other languages from Java.

    2. Graal supports ahead-of-time compilation, so that most JVM applications can be compiled into something that has none of the usual JVM startup overhead in time or memory. The ahead-of-time compilation doesn't support dynamic classloading, though. (Which is only reasonable - you can't compile a class file to native ahead-of-time if the class file isn't provided until your ahead-of-time program is already running.)

    But the second item is only available in the Oracle JDK. I'm not sure how it would work with the OpenJDK GPLv2 license anyway - since the AOT compiled code gets some components of the JVM runtime built into the resulting program, I imagine the whole thing has to be released GPLv2. That's fine with me, but none of my employers would like it.

    Comment


    • #22
      Originally posted by bug77 View Post

      That's one thing Oracle miscalculated. But whether it's MySQL or MatiaDB, neither has the popularity MySQL once had. Today it's more about PostgreSQL or NoSQL.
      Is that really true? I haven't seen anything about MySQL or MariaDB trending down in popularity.

      Comment


      • #23
        Originally posted by Michael_S View Post

        Is that really true? I haven't seen anything about MySQL or MariaDB trending down in popularity.
        Try using MySQL in a Microservice environment. It's not usable because of the licensing fees. Most companies go for PostgreSQL if they want a SQL DB nowadays. PostgreSQL also supports JSON as a field type, that makes PostgreSQL so incredibly convenient to use with microservices and if you have a JS/Typescript backend, you can store your objects directly.
        In academia PostgreSQL is the de facto taught and used standard now.

        Comment


        • #24
          Originally posted by Michael_S View Post

          Is that really true? I haven't seen anything about MySQL or MariaDB trending down in popularity.
          MySQL 5.0 was released 12 years ago (or so). Today we're sitting on MySQL 5.7. If that's not abandonment, I don't know what is.
          It's still used, but reason to pick MySQL over anything else in a new projects are dwindling. I was about to say I can't think of any reason to pick MySQL, but I haven't been following it for years, so that would be dishonest.

          Comment


          • #25
            Originally posted by Degra View Post

            Try using MySQL in a Microservice environment. It's not usable because of the licensing fees. Most companies go for PostgreSQL if they want a SQL DB nowadays. PostgreSQL also supports JSON as a field type, that makes PostgreSQL so incredibly convenient to use with microservices and if you have a JS/Typescript backend, you can store your objects directly.
            In academia PostgreSQL is the de facto taught and used standard now.
            But what's to stop you from using the fully open source version, or MariaDB?

            I'm not saying Postgres is bad. We use Postgres and MariaDB at work, and we're quite pleased with both. But MySQL/MariaDB is a stable product, and it was wildly popular for years. I'm not convinced that "No longer receiving lots of technology news coverage." is the same thing as "Dead or dying."

            (Edit) (!(new hotness) && !(riding the hype cycle)) != (dead)

            Comment


            • #26
              Originally posted by L_A_G View Post
              This gives me the same kind of willies that I got when Microsoft announced they were moving from their own compiler suite to LLVM. Both are obvious ploys to move the cost of development to the open source community and other companies while continuing to reap the exact same benefits as before.

              Seeing how they've pretty much discontinued all Solaris development, killed off SPARC hardware, let the open source community do what they wanted with Star Office and are now in the process of flogging off Java EE development to companies that more interested in actually paying for it, is there really anything left of Sun Microsystems?
              Even if that is a ploy to cut development cost, it's a welcome one given how shit the MVC compiler toolset can be at times, especially with VC++. Clang especially is modified to take advantage of the code generation that VC++ has for its platform. It takes the best of both worlds and is also distributed with VC++ now. Unfortunately, it's not a default and not all that practical to use with tools like CMake that use MSBuild or NMake/Jom behind the scenes.

              You shouldn't judge a company based on cutting cost. If anything, you should judge it for the consequence of cutting that cost. In this case, I don't see a negative side.

              Comment


              • #27
                Originally posted by Michael_S View Post
                I think it's a bit premature to call Java, one of the most widely used programming languages in the world, dead or dying. Most software developers may not like the language, but it's about as close to dead as C++... which is to say, not at all.

                And Kotlin, Scala, Clojure, Ceylon, JRuby, Nashorn (Node.js for the JVM) all extend Java's life along with the life of the JVM because they all can call Java code easily.
                They also develop the Java libraries which directly is beneficial to the other JVM languages. The later versions of Java std libraries are pretty good. They're also pretty up to date with all stuff now. It's one of the best platforms at the moment, along with Javascript/Browser, CLR/Mono, gcc/llvm+posix, and GHC Haskell.

                Comment


                • #28
                  Originally posted by bug77 View Post

                  MySQL 5.0 was released 12 years ago (or so). Today we're sitting on MySQL 5.7. If that's not abandonment, I don't know what is.
                  It's still used, but reason to pick MySQL over anything else in a new projects are dwindling. I was about to say I can't think of any reason to pick MySQL, but I haven't been following it for years, so that would be dishonest.
                  Actually MySQL 8 is just around the corner, with important changes.
                  Originally posted by opensource.com
                  MySQL 8 is coming

                  ...and it's going to be a big change.

                  MySQL 8 is coming and it is going to be a big change. MySQL 5.7 has been out for well over a year and has been very well received with its native JSON data type, increased security, and better performance. But there are some things about 5.7 that needed modernization and that is why MySQL 8 is on the way.
                  What happened to 6 & 7?

                  Years ago, before the Sun Microsystems purchase of MySQL AB, there was a version of MySQL with the number 6. Sadly, it was a bit ambitious and the change of ownership left it to wither. The MySQL Cluster product has been using the 7 series for years. With the new changes for MySQL 8, developers feel they have modified it enough to bump the big number.




                  Dave Stokes gives readers an update on MySQL 8's upcoming release.

                  Comment


                  • #29
                    Originally posted by danielnez1 View Post

                    Despite their current ineptitude, I wish it was IBM who had succeeded in buying Sun. IMO Java et. al. may have been in a much better place today.
                    Absolutely agree. Oracle should not "invest" in [very] expensive lawyers to defeat Google in court and start behaving like a patent troll, but with a real competition investing in [opensource] developers.

                    Comment


                    • #30
                      Originally posted by onicsis View Post

                      Actually MySQL 8 is just around the corner, with important changes.






                      https://opensource.com/article/17/2/mysql-8-coming
                      Ok, almost one major release in 12 years. Do you know of a DB that has seen less than that? 5.0 was also "just around the corner" for years, but that may have had something to do with MySQL going first to Sun and then to Oracle.

                      Comment

                      Working...
                      X