Hacker News Reader

Comment by olefoo | Hacker News Reader

This puts a period to the age of Java as a vital language and puts it firmly in the category of a legacy tool.

I haven't been active with Java for more than 5 years, and I haven't been active in the ASF for longer than that so I may be missing some of the nuance here, but isn't it possible maybe even probable that this is simply the point where Java and/or the Java VM "forks" into Oracle's proprietary version (likely to die on the vine) and one of more versions from some combination of IBM, Apache, RedHat, Eclipse, etc.

It seems like there are a large number of Java developers, a large number of organizations heavily invested in the Java stack, and a number of vendors with both the capacity and potentially the motivation to serve those organizations.

I find it hard to imagine that if developers and organizations abandon Oracle's Java in droves that no one will make an effort to fill that void. (And fill it with something more "Java-like" than not, at least at the start.)

What good is Java if there's no longer a spec to help ensure cross-platform compatibility?

Also, there are major trademark, patent, and copyright issues that need to be settled before anybody can safely fork Java and stop adhering to the spec.

"Mostly compatible" (at the source code level) would be good enough for quite a few applications. For example, take a look at a lot of the C and C++ applications out there, which use OS APIs that often vary widely. Despite that, C and C++ applications are still written, and very popular.

The problem (as I understand it) is that without passing the TCK (which Apache can't get without field-of-use restrictions), you don't get the patent license required to safely run Java in a business environment. I don't think the line-of-business app developers that have driven the bulk of Java usage over the years can afford to risk getting sued by Oracle, so they can't really depend on a non-licensed runtime.

I thought that went the other way: you need a patent license to author and/or run the TCK. You can create a Java compatible VM (many have) but you can't run the "official" compatibility test without Sun's, now Oracle's permission, and there is some concern (at least at Apache) that you cannot easily create a "cleanroom" implementation of that TCK. (Although I think that latter concern is more social than legal--if I remember correctly conventional wisdom said that there were too may Sun folks already involved with Java projects at Apache to have any reasonable expectation that a cleanroom would remain clean.)

Both C and C++ are ISO standards. Oracle doesn't own them and can't buy them.

> Also, there are major trademark, patent, and copyright issues that need to be settled before anybody can safely fork Java and stop adhering to the spec.

One would certainly have trademark issues with the name Java, but I'm not certain about the rest of that statement.

Sun/Oracle have already provided a GPL'ed version of the JDK and core libraries at http://openjdk.java.net/. There are/were independent attempts to do this also, see http://en.wikipedia.org/wiki/Java_Virtual_Machine#Licensing and http://en.wikipedia.org/wiki/List_of_Java_virtual_machines for example.

I think the only real issue is around the TCK, but that's just blessing an implementation as "Java Compatible(tm)", or at least I think so. Couldn't one take one of the existing JDK/JRE implementations, rebrand it IcedTea, and create their own IcedTea compatibility test?

Absolutely. Just because Oracle owns the test suite that "certifies" Java does not mean it won't go on. There are a lot of heavy hitters that are stakeholders, and they are very incentivized to protect the stack.

> versions from some combination of IBM, Apache, RedHat, Eclipse, etc.

To expand: don't forget about VMware (SpringSource), JBoss, Azul Systems, Qualcomm, Siemens, Hitachi, Mitsubishi, and, of course, Google (Android).

Oracle will hear from them, as it has seats on most of the major orgs, including OSGi, where most of the Telecom players live.

Google's investment in Java goes a great deal further than Android - internally many important webapps, including AdSense are written in GWT.

Far be it from me to defend Java - I don't use it much, and indeed, I'm an ASF member myself - but I'm not sure it's quite that simple. The C programming language for instance, is not exactly a hotbed of "cool new language features", and yet it's still very much used, even for new projects. Redis, for instance, just to grab one that comes to mind.

Java (the platform) becomes a legacy tool when there's a clear and open alternative to the JVM, and no, Mono doesn't cut it.

Exactly. The Java ecosystem is vital and people will not move until there's something to fill that void.

Seems to me that a heck of a lot of these "periods" have been declared over the last 12 months.

They all kinda run together into one long ellipsis. However, I think that together it might actually mean something. It seems clear to me that Java is in the waning phase of it's life.

Waning, perhaps, but nowhere near the end of its life. Despite any waning, it still holds a large enough share of developers and enterprise deployments. Oracle does have the potential to screw everything up (they've already screwed up plenty, imo), of course, and anyone involved in Java right now has to be concerned.

Oh, I agree, Java has a very long life in front of it, just because of the installed based. However, I think that Java's best days are behind it (much like Sun). The only thing that is left is Oracle trying to extract as much value out of it as possible. (And I think that's a pity)

Reply on news.ycombinator.com