Slashdot Mirror
Judge Reinstates Java Injunction Against Microsoft
Anonymous Coward writes " New York Times story (free registration required) - the lead paragraph: 'SAN JOSE, Calif. (Reuters) - In the latest chapter in the legal battle between two rival software titans, a federal judge on Tuesday reinstated an order forcing Microsoft Corp. (MSFT.O) to change software based on the Java programming language created by Sun Microsystems Inc.'" And as long as we're talking about Sun, check this Red Herring Story headlined "IBM wants to kick Sun butt." Whoa, baby! They play rough in the big leagues, don't they?
Yahoo: http://dailynews.y ahoo.com/h/nm/20000125/tc/microsoft_sun_2.html ://www.zdnet.com/zdnn/stories/newsbursts/0,7407,24 27342,00.html?chkpt=p1bn
ZDNN: http
Java is still a standard. The only difference is what 'name' goes on the standardized product, Sun or 'National Java Directorate'. With Sun as the controlling body, there are resources and teeth behind each and every 'thou shalt not'.
However, I don't think thats the real issue. I'd be willing to bet the Microsoft/Java litigation gives Scotty endless pleasure. MS screws with the Java bytecode implementation, Scott sics the lawyers back on them, Steve downs some Pepto. MS screws with the Java includes, the lawyers come back out and Steve switches to vodka. Boy! Oh boy! Microsoft is releasing a new O/S with a new JVM? There's an actual release date? Naw, the lawyers have heard about it. Scotty gets his laughs, the MS release date gets laughed at, and Ballmer becomes an alcoholic.
.sig: Now legally binding!
This is probably offtopic, but it is sort of related. Anyway, it is something that I felt like complaining about.
MS is not the only company that is polluting Java or making it difficult to write code on one platform and deploy it on another... many of the IDE's that I've seen do the same thing. The auto-generated code makes use of classes that "wrap" the base Java stuff, resulting in source code that is only useable by the original IDE. Visual Cafe seems to be one of the worst offenders. A coworker of mine uses VC and she is constantly having problems with the autogenerated code and deployment on Solaris (she does her development on NT). She has had trouble getting applets to run under Solaris or HP-UX, even though they run on NT. The main cause seems to be some proprietary class that Visual Cafe includes. Some of the problems may be a result of her inexperience with Java, but I do not think that is the only reason.
I use Metrowerks CodeWarrior for Java on NT at work and on MacOS at home... it seems to be one of the best development environments around. I have not had any trouble getting applets or aplications developed in CodeWarrior to run on any of our target platforms. Even when using the RAD tools that generate code for you... the generated code seems a lot cleaner than the code generated by other IDE's. I keep the use of the RAD tools to a minimum, though. I have had less experience with IDEs like NetBeans and Simplicity for Java, however, from what I have seen, they generate pretty clean code as well.
I guess my main point is that I think its also important for the source code to be as portable as the resulting classes. Building a project in an IDE that ties you to that particular development environment is as bad as Microsoft's pollution of the basic language itself.
Anyway, sorry for the rambling.
I'm sick of hearing some of these ignorant remarks.
"Ohhh, microsoft/m$/micros~1 wants to make their java the standard java".
What a load of crap, Microsoft's _changes_ to java or more specifically, J++, to the programmer just ads some packages that you can choose or not choose to use. True their are some keywords (again additions) to make it easy to access libraries without having to write stub libraries for JNI.
These things are all additions, any java programmer (unless they're 11 year old wannabes) knows what is java and what isn't. It's not as if Microsoft have changed all the keywords, or that they've restructured the language. They've added some features. If you write _JUST_ java and compile it, it will work on other JVMs. If you want to use the java languge to write windows apps (say, as an alternative to VB) you can, but it will only run on windows. If you're stupid enough to think that it will run on Linux too, then well...uh...you shouldn't be programming.
And If you use microsoft's language extensions to make it easier to access, say, dlls/so, then it won't run on other JVMs...with some exceptions (like Transvirtual's clean room java "Kaffe").
Using J++, it's perfectly easy to make 100% pure java apps, but it's also super easy to make windows apps. When you make windows apps, treat Java like a language.
In order to call something a Java(tm) compiler, and to claim it produces Java(tm) code, it must produce code that executes properly on the reference JVM from sun.
;)
MS extended the java classes to include a bunch of MS-proprietary gui & system stuff.
MS tried to exploit what they thought was a 'loophole' in the java contract.
They made it so if you ran the code produced by j++ on the reference JVM, rather than crash (as should happen), they first had code to decide what JVM it was on, and pop up messages indicating what features would not work unless it was run on the appropriate VM from Microsoft.
Microsoft says 'well, it still RUNS, it just doesn't do the same thing'. Totally violated the spirit of the agreement.
The judge decided that they had to change it.
The whole point is, you can do whatever you want with your java compiler/JVM, but what is produced from the compiler *must* run *as the programmer intended* on the reference JVM, and your Java(tm) virtual machine must run all code produced by the Java(tm) reference compiler
It's really easy to not use any of Microsoft's extensions. I use Visual J++ as my IDE and generate code that runs beautifully on Solaris because I stay away from Microsoft's extensions.
Unfortunately, they decided to use a different native method interface, which causes some packages to fail. Also, in their standard distribution, they omitted RMI (like we're all gonna choose DCOM if RMI isn't there). That makes it tougher to write applets and applications that use RMI and run under Microsoft's VM.
It has gotten worse now because Sun no longer sends MS any Java updates, so the MS Java is pretty much stuck at around JDK 1.1.3. JDK 1.2 runs so much faster, I won't use jview at all to run programs, and I'm about to give up on VJ++.
This is a repost from today's other Java discussion. Please read and go vote!
While there has been an official Linux port, there is no native Java 2 SDK for FreeBSD. Please, all members of JDC, go here and cast your votes to have Sun release this software. We are up to 2703 votes so far. It took 4551 votes before Sun released the Linux version, so we're almost there, right? ;-)
(However, I am well aware that the release of the Linux port was due in large part to the excellent folks from Blackdown.)
Here are a few of the tons of comments on the "bug report" page where you can vote for this RFE (request for enhancement).
You must be registered in the JDC (Java Developer Connection) to vote. Registration is free and quick, so if you're a Java developer or just have a general interest on FreeBSD or Java, please go sign up and vote!
Anyone interested in doing an unofficial port please mail me at javadrew@spammerslovehotmail.com (obviously remove the "spammerslove" :-).
--
I like to watch.