Oracle won't pull plug on Java SE 6 until 2013 Oracle is extending the official end-of-life date for its aging Java SE 6 software development platform a second time as it struggles to get the Java language development process onto a consistent, two-year release schedule. Had it stuck to its original roadmap, Oracle would have stopped providing public support and software …
The problem for most companies to switch from Java 6 to Java 7 has nothing to do with changes in the language, but mostly with having a new version itself. I am not aware of any language change that is not backward compatible (as in code written for Java 6 will still work in Java 7).
Most of the problems are with resistance to change or lack of corporate reasons to change, and platforms that only (formally) support running on Java 6. Also for some reason companies are a lot less weary of updating point release than updating to major release. Add to that the lack of exciting new features in Java 7 (with the exception of try-with-resources), and there is no real good reason to spend time and effort to switch.
All-in-all Java 7 was a disappointment, and now it looks some of the exciting things which mere delayed to Java 8 will be delayed to Java 9....
Bull pucky. I have had perfectly functional Java applications from major software vendors suddenly cease to function after a minor revision update in the JRE. Backwards compatibility in Java is a joke, and moving to 7 will be a complete nightmare for most organizations with any dependency on Java apps whatsoever.
Java7 introduces some changes in bytecode, and this can mean that things that functioned with JRE1.6 don't with 1.7. Obviously there is some argument you can pass to the JVM that means it treats the bytecode like in 1.6. But third party bytecode tools are still being updated to use this 1.7 bytecode, and it will take some time before all are upgraded.
Nonsense.
Just to change some apps from java 5 to 6, we had some 50K pounds spent for each of them..at least!! so if you have some 100 apps that would mean a minimum os 5 million pounds.. and no benefit!!
Not counting that big apps can cost 100.000 pounds or more just for them to work as they did with the previous version...
Really? 50K for migrating from 5 to 6?, either you have some very rich consultants, or really slow developers? There is not that much of a difference between 5 and 6 and being 5 compiled jars run fine in 6 I can't see where your money went.
Fluff. Going from JDK 6.18 to JDK 6.WellAnythingBiggerThan18 was trouble enough for us due to a Bug introduced into the JRE, a nice well-documented bug that Oracle don't want to fix (still) that buggers up JWS installs (null pointers due to jars being put in a weak-cache). Luckily a couple of bright sparks in the world built a feasible workaround, but we fought random application failurs (yep, random) for too fluffing long.
And yes, there are a few things that give you run-time errors in Java 7 that worked perfectly fine in Java 6... We've not found many.
Don't get me started on IDE compatibility - who would have thought that Oracle's OWN IDE would fail to offer Java 7 support for this long...
I can imagine that subtly different behaviour in libraries, tricky multithreading dependencies and API ambiguities in programs that "used to work" (but were actually buggy) will cause trouble on upgrades.
Developers hope that they know what a library does, but they often don't. Ambiguities exist, complexity is high, edge cases are numerous and assertions are missing everywhere (I have seen people afraid to use assertions because "it might break the program at runtime"). This is not helped by unelegant APIs and the extreme complexity inherent in trying to make Bad Ideas Simple To Use. If you have ever seen someone faff around with JAXB .... it's not a pretty sight.
Upgrading the JVM could possibly be considered in the same league as upgrading the OS, the compiler and the runtime libraries all at the same time.
Lotus' founders tell a different story.
I love this quote: "We had to make changes to DOS to help some very old applications that were doing some very bad things."
And they just finished migrating our most troublesome and somewhat critical app from 1.0.5.16 to 1.0.6.29 last month.
I was really hoping for a break in the "critical app requires non-supported java" support calls. Oracle needs to figure out that most business can't support that sort of release cycle. Sun probably did us all a huge favor in keeping a (mostly) stable platform for about 5 years.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
