Oracle Sets End Date for Business Java 8 Updates (infoworld.com)

An anonymous reader quotes InfoWorld: Further clarifying its ongoing support plans for Java SE 8, Oracle will require businesses to have a commercial license to get updates after January 2019. In an undated bulletin about the revision, Oracle said public updates for Java SE 8 released after January 2019 will not be available for business, commercial, or production use without a commercial license. However, public updates for Java SE 8 will be available for individual, personal use through at least the end of 2020.

Oracle advises enterprises to review the Oracle Java SE Support Roadmap to assess support requirements to migrate to a later release or obtain a commercial license... Oracle advises developers to review roadmaps for Java SE 8 and beyond and take appropriate action based on their application and its distribution model.

Comments

Why not migrate

[phantomfive]

Is there any reason not to migrate to the newest version of Java? Is any effort even required?

Re: Why not migrate

[Anonymous Coward]

A lot of effort sometimes. For example many libraries like spring require bytecode modification and that is bytecode version dependant. To update spring you might need to update other libraries too and some of them may have deprecated things you use and so on.

Re:Why not migrate

[careysub]

Well, the latest version of Confluent Platform community edition - a Kafka/ Zookeeper offering - only runs on Java 8. That would be the major reason, using code - either commercial or open source - that does not run on Java 9 or 10.

Then there is the puzzling short maintenance periods being offered for all post Java 8 releases. See for example the comments here. Moving to a new Java version that will be supported apparently for six months is very questionable for any enterprise. It appears that Oracle has some sort of scheme to make everyone (businesses, anyway) to pay for using Java going forward.

Re:Why not migrate

[shanen]

Is there any reason not to migrate to the newest version of Java? Is any effort even required?

Sure. You don't need or even want any of the new functionality and you believe that the additional complexity has made the new version less secure rather than more. Even stronger reason if you think any of the new features are negative rather than positive. Effort is NOT the real question.

Too bad we don't have any such option.

I think we're going about this all wrong. Old versions should be supported as long as sufficient numbers of people are actually willing to pay for whatever support they require. If they want to deprecate an old version, the way to do it is NOT by arbitrary announcements of when the new version shall be shoved down your throat, but rather by rational explanations of the various new versions and encouraging the users of the older version to consider one of the new ones.

Now if there are too few users of the old version who are willing to pay for the costs, then that's a different question. In that case the old version may need to be more strongly retired, and the users may be obliged to make choices, but they might prefer to gather around an even older version that still does what they want, or they might want to go forward. It should be the users driving the versions based on their real needs, not the corporate cancer (Oracle in this case) driving things as part of the never-ending quest for infinite profit.

I actually think it should go down to a feature-by-feature basis. In particular, features and functions that have security considerations should be obliged to check for their own validity before executing, possibly becoming inoperable if the security threat is too high. In terms of providing more freedom, there should be options for similar functions that don't have the problems...

DSAuPR, atAJG.

Re: Why not migrate

[Anonymous Coward]

Which proves java is a shit technology. The whole point is write once run everywhere, when everywhere does not even include a later JVM version it's pretty broken. Java leaks to much of the abstraction. It should not be possible for the monkeys to cause themselves these problems.

Re:Why not migrate

[charronia]

If you want to migrate, all your dependencies need to migrate first. It's taking a lot of libraries some time to get on board with Java 9, especially since it uses a completely new class loading mechanism.

Re: Why not migrate

[phantomfive]

That just means a new recompile, right? Or is there something more?

Re:Why not migrate

[Anonymous Coward]

Java 9 does a crap load of random things that breaks things. They redid the JRE layout, they redid the standard APIs so that some parts are not loaded by default thanks to the new module system.

There are a lot of good reasons to stick to Java 8, because a lot of code does not and never will run on newer versions thanks to random unnecessary incompatible changes.

Re:Why not migrate

[CaptainDork]

I think we're going about this all wrong.

Oracle is not a "we."

Re:Why not migrate

[angel'o'sphere]

Well,
software vendors can not really migrate if the customers don't want to migrate.
I'm in a project with a few hundred customers who all have dozens or hundreds of installations.
If we migrate the software to Java 9 (or 10 soon) ... we are still on *7*!! They have to migrate as well.
But as you probably know the mantra: never change a running/winning team, everyone is reluctant to perform a change.
Technically I'm not aware of a reason. A software compiled on Java 7 and packaged in Jar files should simply run out of the box on Java 9. However we have to test it on various OSes and DB vendors/versions. In some cases migrating to a newer Java version means migrating the OS, too.
E.g. Java 9 is not running on OS X 10.9.5. So if I have to make sure my software runs on Java 9, I need at least one test machine, and while I develop on Java 8 I can not use any Java 9 features.
However: it sucks. I know companies that are still on Java 6 ...

Re: Why not migrate

[jonwil]

Just use OpenJDK. Unfortunately for Oracle, the GNU General Public License prohibits Oracle from stopping anyone using OpenJDK for whatever they like as long as they comply with the license. And the "classpath exception" (or whatever it is) means you can use OpenJDK and you dont have to publish any of your source code, just the source code for OpenJDK and any changes you have made to OpenJDK.

Re:Why not migrate

[Anonymous Coward]

No 32-bit support after Java 8.

Re: Why not migrate

[angel'o'sphere]

Actually, no.
While the parent is right, the class loading changed, because of the concept of "modules", you can still invoke the JVM on the command line and force it to use the old mechanism.
However typical modern projects easily use hundreds, yes, several hundred, libraries. The project I'm working on right now is already in itself split up into about 100 sub projects and pulls in about 900 libraries. Mostly Apache or other open source stuff, a few dozen commercial products ...
Of course we are only directly dependent on perhaps 100 or 200 libs, but those pull other libs (via maven dependency management).
So basically everyone is waiting that the main open source products move to Java 9, so they don't have to recompile and host their own repositories.
But technically recompiling should not be necessary, standard Java 8 compiled jar files should simply run out of the box on Java 9 environments.

Re:Why not migrate

[Guillermito]

Old versions should be supported as long as sufficient numbers of people are actually willing to pay for whatever support they require.

That's exactly what's happening. According to the article and the summary, customers paying for a commercial license will continue to get updates.

Re:Why not migrate

[Anonymous Coward]

One of the things that sucks balls for Java is its lack of portability and compatibility, not only did it fuck up on its promise to write once run anywhere between platforms, it can't even do it well between versions.

Re: Why not migrate

[Kjella]

Which proves java is a shit technology. The whole point is write once run everywhere, when everywhere does not even include a later JVM version it's pretty broken. Java leaks to much of the abstraction. It should not be possible for the monkeys to cause themselves these problems.

In theory the monkeys shouldn't cause problems at all. In practice any non-trivial software has shipped with buggy, flawed or broken functionality including the JVM itself. I've witnessed myself how security patches to Java broke functionality. The solution? Install the exact version it was released with and leave it there. Sure, the support agreement says patches are supported. In practice, you get to be the beta tester.

Re:Why not migrate

[jrumney]

Or to OpenJDK, based on the same source code but without the Oracle licensing nonsense.

Re:Why not migrate

[shanen]

NOT exactly, but I suppose it is close enough for you if you read with that sort of preconception or without thinking clearly. Or perhaps the fault is my own for believing that my assumptions should be obvious enough from my conclusions?

From my perspective, Oracle is essentially taking the users hostage to maximize Oracle's profits, and will primarily be driven by profit considerations in deciding whether or not to offer any support options, and how long to offer them. Oracle is balancing the relative profits from the different versions, which includes some hidden profit (AKA brand value) from maintaining control over the Java brand.

If there were true competition driven by the real needs of users, then that would also control the frequency of new releases. My suggestion is to focus on cost recovery, not profit maximization. It doesn't even bother me that there will be some freeloaders benefiting from the availability of the software as long as the per-user cost is kept low enough so that most users can decide to support the software without financial stress. That's not how Oracle sees Java or saw OpenOffice.

Re: Why not migrate

[the_B0fh]

Not when your idiot coders have things like "is jdk == version 8.123.456" in the code.

No one apparently knows how to do something like "is jdk >= version 8"

Re:Why not migrate

[tepples]

E.g. Java 9 is not running on OS X 10.9.5.

What specific error message blocks compiling and running OpenJDK 9 on OS X 10.9 "Mavericks"?

Re: Why not migrate

[sloth jr]

I think it likely that Oracle would attempt to use its newfound "Java APIs are copyrighted" court finding to turn OpenJDK and competing implementations into infringing implementations.

Re: Why not migrate

[phantomfive]

I don't think Java compiled under Java 7 will run straight under Java 9, wasn't there a bytecode change?

Re: Why not migrate

[phantomfive]

Good to know, thanks

Re:Why not migrate

[angel'o'sphere]

No idea, never tried to compile it.

Re: Why not migrate

[angel'o'sphere]

Added byte codes, but no changes in old ones, but that change was from Java 6 to Java 7 already. I'm not aware if there is a recent byte code change again.
Old byte code usually always worked. Only meanwhile I would assume that Java 1.1 or 1.2 no longer runs because deprecated APIs finally got removed.

Re: Why not migrate

[phantomfive]

I haven't used those deprecated APIs in a while and I don't remember seeing them used either.

Re: Why not migrate

[theshowmecanuck]

So you're saying the issue is too many frameworks. I agree. Especially Spring.

Re:Why not migrate

[phantomfive]

Is that a reason to migrate, or a reason to not migrate?

Re: Why not migrate

[thomn8r]

All Java accomplished was to transfer the problem of code portability from the developer to the SysAdmin

Re:Why not migrate

[ausekilis]

One of the tenets of Java is its backwards compatibility. I can write Java 6 code that runs just fine on Java 8 or 9. I have a hard time believing that they are set to a specific version (unless they have some stupid version check that sets hard boundaries).

Re:Dear Oracle

[Anonymous Coward]

Why are you calling me sexist? I'm merely telling you the logic college administrators trying to boost women's enrollment in CS classes had decided. If you feel that is sexist, take it up with them, not me.

>To reduce the intimidation factor, the course was divided into two sections — “gold,” for those with no prior experience, and “black” for everyone else. Java, a notoriously opaque programming language, was replaced by a more accessible language called Python. And the focus of the course changed to computational approaches to solving problems across science.

>https://www.nytimes.com/2012/04/03/science/giving-women-the-access-code.html

Clarification here:

[Anonymous Coward]

Sun open sourced most of Java back in the mid-late 00s a few years before its sale to Oracle. The open sourcing of Java is what lead to Apache Harmony dying at 96 and 98 percent complete for Java 5/6 implementation support. The Java 6 support actually made it to 99 percent, but no release was published after the developers on it decided to give up since nobody was really serious about finishing that last 1 percent of missing code. During that time, the guy who started Android, then called Droid, before Google took over forked the Harmony codebase and used it as the basis for Android. Meanwhile Sun made regular code dumps of difficult to build JRE/JDK code, which eventually became the IcedTea project, under the terms of the GPLv2 and various other licenses (Sun's GPLv2 support was reluctant, as their choice of the CDDL for Solaris proves.) Long story short, thanks to them licensing it under the GPL, IcedTea as a derivative implementation of Java is patent protected for any patents used in the original codebase. So long as no NEW code is implemented using patents Oracle controls that were not used in one of the GPL'd code dumps they are safe from Google v. Oracle altering anything so long as they do not implement any APIs that are not open sourced. At this point in time any APIs that aren't open sourced but do require reimplementation will be handled just like encryption was in the 90s and simply be produced in foreign countries without software patents and sane copyright laws surrounding API/ABI/etc usage.

If Oracle wins as broadly as they are trying to win, I foresee them losing out on control of Java altogether as current Java users band together for security fixes and support of deprecated JRE/JDK versions leading to LTS IcedTea releases while deprecating Java as a new development language and eventually migrating to something else. What, I don't know, since the same dangers with Java exist in the DotNet ecosystem (especially now that Microsoft wholly owns the only 'gpled' edition, which hasn't been completely GPL since at least the 2.x series of releases, when Microsoft started releasing code under non-'Open Source compatible' licenses.)

I was never a big fan of java, given just how much memory it always seemed to use for even the most trivial of tasks, as well as what I felt was the baroqueness of the language, but given just how much of the internet ecosystem relies on it now, I think this could be VERY damaging in the short to mid term.

Re:Clarification here:

[jonwil]

What exactly about Java is so important to the Internet? Java Applets haven't been a thing in who knows how long. (I haven't had a JRE installed on my PC for years because of all the security flaws and because I haven't found any software I want to run that needs Java)

As for .NET, the compilers for C# and VB.NET, the runtime environment and a lot of the base system libraries are available (as well as things like some of the ASP.NET and WCF stuff) under FOSS licenses (Apache 2.0 and MIT from the look of it).

And this is development in the open (with contributions accepted from outsiders) rather than random out-of-date code drops.

Plus with the future of the .NET platform being directed via the .NET foundation and its technical steering group (a group that includes Microsoft, Red Hat, Jet Brains, Unity, Samsung and google) and via the ECMA standards for .NET and C#, its a lot more open than Java is (the core bits at least, I do know things like WPF and WinForms and such aren't open)

Re:Clarification here:

[angel'o'sphere]

Half of what you see on dynamic web sites (versus static HTML) is run by JVM based languages, like Java, Scala, Groovy and meanwhile Kotlin, probably even 2/3rd.

Re: Clarification here:

[Anonymous Coward]

AWS, Google and Apple (among many others) use server-side Java extensively. Itâ(TM)s not an exaggeration to say that sever-side Java is 10x as big as Rails, .net, etc

What Oracle is about now?

[Anonymous Coward]

Who knows what Oracle actually is doing nowdays? I spent quite a big chunk of my developers life using Oracle DB and trying to implement multiple DoA web projects utilizing their portal and application servers, but it was quite long time ago. What do they do now, except suing Google for Java ownership?
They missed Cloud completely.
They missed mobile, but it is Ok for them.
They missed big data.

What Oracle is about now?

Re:What Oracle is about now?

[Voyager529]

What do they do now, except suing Google for Java ownership?
They missed Cloud completely.
They missed mobile, but it is Ok for them.
They missed big data.

What Oracle is about now?

You're right, Oracle isn't exactly an exciting or innovative company. That doesn't mean they aren't still making massive amounts of money.

Oracle has their own cloud offering, believe it or not. It's not dissimilar from AWS, but Oracle simply doesn't have much of a market. If you're up-and-coming, and don't have data in an Oracle DB already, you're not starting with one. You're also not starting with Oracle Cloud because "everbody uses Amazon", though you might be able to make a case for Azure or GCC if you're already in one of those ecosystems. If you've got data in an Oracle database already, you're not happy about it, and trusting a company whose installer requires a lawyer present with even more data an infrastructure reliance isn't exactly the easiest sell.

Oracle was also never going to make any real inroads in mobile. The closest they come to being a hardware company is whatever is left from acquiring Sun, and SPARC-based big iron isn't exactly a springboard for entering the smartphone world. Now, one can argue that 'Search and E-Mail' might not have been the most logical springboard either, but Solaris never hit the consumer market, and Oracle Linux is basically rebranded Red Hat, who isn't exactly taking mobile by storm either, by the by.

Big Data is something that I can't speak too strongly to, but from what I understand, they've implemented a number of things into more recent releases of Oracle DB that enable NoSQL-like functions. The thing is, though, that projects based on Big Data are recent enough that "not-Oracle" is all but assumed on the outset.

What Oracle *does* do, however, is what they've done well for a long time - play the government contract game. Bid for a project, which conveniently uses Oracle products, get in with the lowest bid, end up having the project grow to 150% of the bid, have the project end up with some terrible combination of scope creep and design-by-committee to the point where it doesn't work right but *technically* delivers, then charge to have their consultants work for double the contract length for 'maintenance and upgrades', while also charging for the software support licensing and contracts. Then, when the project is finally scrapped...the cycle repeats.

So long as Oracle is not banned from bidding on state and federal projects, and the government doesn't make an initiative to work to remove existing Oracle installations by 2025...Oracle can play that game until someone starts caring.

Re:

[kurkosdr]

Most companies have customers, Oracle has hostages.

Re: Dear Oracle

[nnull]

The problem is, it won't be. There are so many Enterprise software that uses various versions of Java out there, it's going to be around for a long time.

Flash was phased out thanks to an effort by many of the browser developers to force it out.

Good luck trying that with Java. None of these big developers for this software have any intention of improving the situation either. I still see even NEW ERP software using and requiring 1.7. This sector of the industry doesn't care, nor do they care about security much either, aka stuxnet.

Re:Dear Oracle

[Jerry]

A "sexist" is one who thinks women cannot defend themselves.
The OP wasn't sexist, you took only half of his quote. The part you conveniently ignored states
but since like C++ before it, it was still deemed too hard for women,

It's plain to see that Anonymous didn't insert those words, but you left them out. Sexist?

So, can I just jump to Java 16?

[Volatile_Memory]

Java 8 end-of-life is January, 2019. So, let's say I want to switch...
Java 9 has already ended free support as of March, 2018. Can't go there.
Java 10 free support expires September, 2018 (again, before Java 8). No need to go here, might as well wait for...
Java 11 which won't even be available until Sept. 2018.