JDK 1.4 codenamed Merlin to be released this month


News: JDK 1.4 codenamed Merlin to be released this month

  1. JDK 1.4 codenamed Merlin is expected to be released this month. It is developed under Java Specification Request(JSR) 59. The package will be bundled with XML, Logging API, JDBC 3.0, Security, Preference API, Assertions, custom remote reference for RMI and many other features.

    For details check out this article
  2. Very nice features indeed! Assertion facility and logging API are nice to have. I would really have liked to see Generic Types support to Java Language already in this release (JSR-14), but... I guess we have to wait...
  3. Doesns't Java already have generic type support at the Object level?

  4. Never mind. I just read JSR-14. It is describing the kind of generic type support that I was accustomed to in my Ada days.

  5. Please read a litte bit more carefully next time. J2SE 1.4 (*) release is planned for fall. This month you get only a beta release.

    And before the release of 1.4 we will see a bug fix release for 1.3. (Which looks at least for Linux very promising.)

    (*) For nitpickers: JDK ist the old name for the software development kit for the Java 2 Standard Edition (J2SE).
  6. RMI Custom Remote References and RMI security are not going to be in Merlin. After approving the objectives of the JSRs and participating in the expert group, several EC members decided that the objectives were counter to the interests of J2EE and voted not to release the spec for broader community review. Sun, in its defense, had come up with specifications which met their stated objectives and addressed the needs of RMI developers but had to bow to the will of BEA and IBM. JSR 76 and 78 never saw the light of day.
  7. Regarding JSR76/78. The proposed new API's did *not* address the needs of RMI developers, which is why they were (IMO correctly) voted down by a majority of the EC members (including BEA and IBM, but also many others).

    These proposed new API's would have hindered RMI's evolution severely, since it would have been very difficult to address the basic problem with RMI today if they had been accepted: pluggable wire transports.

    When these problems with the proposed API's have been fixed, and they will be, RMI will be in a much better shape than if these JSR's had been accepted right away.

    Bill, do you think I'm wrong? If so, can you please clarify why you think the rejected votes were bad decisions? Do you know something I/we don't?