Want to read Slashdot from your mobile device? Point it at m.slashdot.org and keep reading!

 


Forgot your password?
Close
typodupeerror
×
Java Programming

Java 2 on Linux 94

Posted by CmdrTaco from the check-out-blackdown dept.
EmilEifrem writes "Just in case you guys felt like posting something about Java, Infoworld says that Java2 is due to be released Thursday. Is it Thursday over there yet? " It supposedly will be online after it propogates to the mirrors.
This discussion has been archived. No new comments can be posted.

Java 2 on Linux More

Java 2 on Linux

Comments Filter:
  • I also use Blackdown Port of JDK1.1.7. Without the TYA JIT, things worked but were sluggish; with the TYA JIT, things got quite snappy. I'm able to run anything I want to including NetBeans2.0 Java IDE.

    Some of my Java apps have been going for days with no problems.

    Better to have TYA than not. I like it!
  • Posted by wri guy:

    Sorry, but I think your Java logo is lame.
    Won't Sun allow you to use the real logo?
    I would think they would like the publicity!
  • Gee, I wonder who the idiot was that suggested this fix in the jitterbug log?

    (of course, switching to Debian 2.1 is going to be my bugfix as of next week, but still.)

    Considering that the announcement explicitly says that it is a prerelease ONLY FOR THE BRAVE, this sort of "oooooh scaaaarey" dialogue has all the importance of your average War on Drugs.

    "But weird shit could happen, JUST LIKE THEY SAY!"

  • Not on Linux. JDK1.2 has been out for a while on Windows (and is very very impressive I might add) but is only due (if the rumours are true) to come out on Linux in the next few days. I can't wait!

    --

  • Why do I constantly hear things like "Linux won't natively support java any time soon". Err, it does already, and has done for some time - last time I compiled my kernel it was right there: "Support for Java binaries"...

    Have I got this confused (it works on my machine - perhaps I'm the only one) or is this FUD?
  • Oh, the maze! That's easy. You kill the cat nine times and throw it into the swamp with the cricket bat. You then have to feed the snake to the bird, which will scare off the mirror in the canyon. Then follow these directions -exactly- to get out of the maze:

    E,N,E,M,E,N,E,M,I,N,E,M,O

    Now, the next stage is trickier. You have to complete the Swing, defeat the Native and discover the magic ORB of the Dyslexic Snake.

  • Regardless, the one you want is libstdc++ 2.9, so that could be part of the problem
  • Although there are many "cons" pointed out about Java that are valid, on the whole I think it's going to get bigger, not go away. It does potentially make some projects easier. Additionally, monsterboard and dice.com seem to show an increasing demand for Java developers (while the demand for VB skills is decreasing).
  • Check JitterBug first! Bug #424

    http://www.blackdown.org/cgi-bin/jdk/incoming?id =424;page=14;user=guest
  • Java is a programming language, not a game!
  • So unfortunately what you write is not very accellerated.

    However, you can license Magician classes, which provide an OS-agnostic 3D interface in Java, and works on Win32, Irix, MacOS... not sure about Linux.

    The program I test is Java-based, using these extensions. I'm running on very nice hardware so I couldn't describe how it runs on standard systems without geometry accelleration and hardware OpenGL support.

  • The readme file states that a JIT is included
    and used by defualt. And that it is based on
    the Solaris x86 version. This is great news,
    I think. Anyone notice any speed improvements?
  • From the status page [blackdown.org]:

    Most people already know why, but in case you don't: the source licensing agreement that the Blackdown Team has with Sun says that we cannot release our ports, even in pre-release testing form, without passing the JCK. So, even though our port has been running quite well since November, we can't release it because there are some JCK tests that it doesn't yet pass.

    So what's this? Does this mean that this build does pass all the compatibility tests (sounds like it has to, or they couldn't release it at all), but does not run stably (you'd think general stability would be among the criteria that Sun insists on)?

    By the way, funny story: two days ago, a friend asked me about Java 2 for Linux. I hunted around a little and pointed him to blackdown.org, which had 1.1.8 available, and 1.2 "Real Soon Now". Then yesterday at Linux World, I saw that the Sun booth had "Java 2 SDK for Linux" signs up, so I asked the guy about it. He said that an hour earlier Steve Byrne had been there and showed him that the size of the file on his web page was increasing as it was being uploaded from his machine at home over a slow dial-up. Now this. So this is not the real release yet, but still, a lot of coincidences. For once, "Real Soon Now" really is soon.

    David Gould
  • BASIC [wins.uva.nl]

    BASIC /bay'-sic/ /n./ [acronym: Beginner's All-purpose Symbolic Instruction Code] A programming language, originally designed for Dartmouth's experimental timesharing system in the early 1960s, which has since become the leading cause of brain damage in proto-hackers. Edsger W. Dijkstra observed in "Selected Writings on Computing: A Personal Perspective" that "It is practically impossible to teach good programming style to students that have had prior exposure to BASIC: as potential programmers they are mentally mutilated beyond hope of regeneration." This is another case (like Pascal) of the cascading lossage that happens when a language deliberately designed as an educational toy gets taken too seriously. A novice can write short BASIC programs (on the order of 10-20 lines) very easily; writing anything longer (a) is very painful, and (b) encourages bad habits that will make it harder to use more powerful languages well. This wouldn't be so bad if historical accidents hadn't made BASIC so common on low-end micros. As it is, it ruins thousands of potential wizards a year.

    David Gould
  • hmm Netscape does that anyway even with no java
  • You're comparing apples to oranges. Blackdown's JDK doesn't include a JIT. Solaris, Windows, OS/2, and others were all tested with the JIT. If you disable the JIT, you'll find that Linux's port is quite nice.

    As for the scalability, I believe they were testing a version running green threads. The native threads included in Blackdown's 1.1.7 ought to help. The 2.2 kernel also should help, as it has much better support for threads.

    There's no defending the performance of TYA, though. I keep hoping that IBM and/or Sun will release a JIT open source. That would be really nice...
  • I didn't mean that TYA was slower than Blackdown--it's not. In fact, TYA is about 2x faster, according to my own benchmarks (real world code to load data from disk, process data, then manually create an image (i.e. no hardware acceleration through native code)).

    However, as a JIT, this simply sucks. Most JITs are about 10x faster than interpreters.

    That said, the Blackdown port is a little bit faster (~10%) than Sun's interpreter (ooh, now I'm comparing apples to oranges--the Sun interpreter is running under WinNT).

    However, for complex applications (I write oil industry apps & tools, not animated icons for the web), interpreted code is simply unacceptable. Furthermore, TYA's performance is also unacceptable. What is acceptable? The Symantec JIT (the one that ships with JDK1.1.7) is awesome. The MS JIT is also awesome (with my code, it's with 5% of the Symantec JIT; sometimes faster, sometimes slower). I don't have access to IBM's JIT, but I've been told by reliable sources that it is about 2x faster than MS' and Symantec's. I've also been told that a group inside IBM turned off a few of Java's features (most notably, bounds checking), and received an *additional* 10x speedup on mathematical code (I write seismic viewers, and must perform stuff like filtering, normalization, and scaling--this stuff should all benefit greatly by removing bounds checking). There is some talk that the JavaGrande group (the Java supercomputing group) will convince Sun to allow the JVM to turn off bounds checking, but only after telling the user that they're about to commit an agregious act.
  • Many congratulations to the Blackdown folks for getting the prerelease of the port out. Great work, guys! We really appreciate what you've done.

    To the Linux community - Java is very important for the future of Linux, it's imperative that Java work well on it. Java is already the language for Internet programming, and it's becoming the language for embedded networking thanks to Jini. Linux is uniquely positioned to serve these areas. We have a real opportunity here, especially now that Microsoft is dropping the Java ball.

    The one problem left is the licensing model - Java isn't exactly Free. But Sun is pursuing very interesting licensing terms with their Community Source License; it's worth a look. The SCSL isn't Open Source (tm), it's more restricted so that Sun has a commercial angle on their work. But overall, it's interesting. They're pursuing this strategy very aggressively, releasing Solaris and their CPU designs (Sparc and picoJava) under SCSL.
  • I don't know the names of the RPMs, but Debian's "egcc" is egcs, and libstdc++ (as far as I know) is part of the egcs package, at least as it comes from Cygnus. As someone else has reported, this seems to already be in the Blackdown bug tracking system.
  • *Snork*

    Java has lots of cool libs, but as a language, what does it have? No list literals. No associative array type. No array slicing literals. No multiple inheritance (I like the option to have a dangerous tool). No meta-object protocol. No parametric polymorphism. No method calls on primitives (like say, 4.sqrt()). No tail recursion elimination. No first-class functions (I'm staggered by that). No named arg passing. No lambda forms, no dynamic scope. No generic functions.

    Really, the only innovation I've seen is anonymous classes.
  • Yeah. ON ERROR GOTO

    GOTO. my sides are splitting. Even MOO is more advanced than VB.

  • Y'know, I wouldn't want a nuclear reactor to run Java, and neither does Sun:

    4. High Risk Activities. Notwithstanding Section 2, with respect to high risk activities, the following language shall apply: the Software is not designed or intended for use in on-line control of aircraft, air traffic, aircraft navigation or aircraft communications; or in the design, construction, operation or maintenance of any nuclear facility. Sun disclaims any express or implied warranty of fitness for such uses.

    I'd check my VB license but I'd have to install Windows first, sorry.
  • I'm on Slashdot. I'm so happy. Can't you guys /. my web site now, just for fun? Please?

  • Lie.. (Score:1)

    by Axe ( 11122 )
    ..faster? more stable? without a good JIT?

    It is like 30% slower on my code. At the very least.

    Hope Java 2 port is better. It should be.
  • Just add a sym link to your existing libstdc++

    ln -s /usr/lib/libstdc++*[whatever you have] /usr/lib/libstdc++-libc6.0-1.so.2

    It works for me !!!
  • Hehehe. That's pretty funny.



    Seriously though, you'd feel better having a bunch of VBXs hacked together running a nuclear power station? Originally, VB was intended as a scripting language, oh, wait, it still is.



    VB is easy, it's the microsoft way of doing things: allow for flashy, management impressing demos to be whipped up at a whim, but forget about the back-end. As soon as I ever started doing serious work in VB, I ended up spending more time working around those 'black-boxes' than getting any work done. Java is *much* better thought out and allows for much, much more easily customized UI, and other little things like language security, etc. I've just spent 3 hours today tracking down a bug in a VB application that was caused by someone not initializing a variable. You can't do that in Java, which, amoung other things, makes it a much more robust language, much more naturally un-error prone.



    Anyway, VB has it's place. It is easy to throw together flashy database connected demonstrations, but it doesn't do anything that Java can't do. Especially considering all the RAD enviroments coming out for Java. And your Java executable will run almost anywhere ( and getting more and more ).



    As far as VB being more logical, that's interesting. However, I'm not going to comment as anything I would say would probably come across as rude.



    Finally, as anyone with any experience will tell you, Java is *the* language to use for network work. Even the microsoft oriented books I've read on COM and stuff say that Java couldn't be better designed to integrate with COM if it tried, and is clearly the language of choice for any ORB driven programming model.



    BTW, does VB actually compile to real code now? Isn't it done with p-code, but cleverly link in the run-time stuff?



    CraigL->Thx();
  • He knows. that's sarcasm.
  • It's actually just a neat hack that checks to see if you are trying to run a .class file, and if you are it chops of the .class extension and passes it to the Java interpreter, which of course must be installed in a standard location. If someone out there is looking for something to do, how about adding support for executable jar files? I made a stab at it a few months ago but didn't get very far (not really knowing what I was doing). Here's a URL to look at : http://www.tools.de/java/ . Send me some EMail when you've got it working :)
  • I'm sorry to say that, like the VolanoMark (not 'Volcano', actually), my Java server code gets much poorer performance under linux than under NT. And on linux I bog down terribly after threading to 100 simultaneous TCP connections. NT allows me to handle 1,000 clients pretty easily.

    This is very discouraging, and I hope the new Blackdown release (which advertises native threads) will improve performance there.

    Note that the VolanoMark benchmark is essentially a chat server, so if you do other kinds of Java work, your results will vary.
  • according to the pre-release announcement [wisp.net], Java 3D is expected "in another couple of weeks or less", with JMF and the others to follow later.
  • Sun realease it months ago but we were still waiting for the port to linux, i can't wait for the day when it gets released the same time as the Win32 and Solaris JVM's
  • > Java is already the language for Internet programming

    Nope, that's Perl!

    /Alex
  • From what I've been able to tell, VB does compile to real code, as of VB5. However, this VB executable still requires a big honkin' library DLL. Sort of like MFC, only you can't possibly link the libraries statically.

  • As much as i hate to admit it, I do a bit of VB programming. I am currently in the process of learning Java, in the hope that it will serve some of my light programming needs better than VB. Admittedly, I'm still running 5.0, so I suppose some major changes may have been incorporated into version 6.0; maybe you can tell me if they have.

    So, has VB added decent networking capabilities since 5.0? I'd be surprised if Microsoft has provided anything better than the poorly documented, unreliable ActiveX controls that came with 5.0.

    How about error handling? Is there any way to trap and handle errors other than continuing execution until you feel like checking error numbers? I'd feel much safer around a nuclear reactor controlled by a language with exception handling.

    Every now and then, I'm amused to hear someone refer to Visual Basic as an object oriented language. I suppose aggregation could be considered a crude form of inheritance, but does VB have anything remotely like polymorphism?

    I'm not going to go into platform independence in any detail, because the very words make Microsoft supporters curl up and whimper. Java may not have reached the goal of platform independence, but at least it considers it a goal.

    Finally, I suppose that with the way you write and spell, you need a programming environment that will complete function names and check syntax on the fly. The more literate among us prefer the ability to choose tools to meet our needs.

    P.S. I'm sorry, that last bit was a cheap shot, and I'm better than that.

    P.P.S. No, on second thought, I'm not better than that.

  • D'oh! You're right, I was getting VB, which has the "on error goto" syntax, confused with VBScript, which only has the "on error resume" statement. I apologize on that particular count. However, without getting into the problems of the "goto" statement, I still prefer Java's exception handling for passing additional error data in a fairly standard way.

  • hava (Score:1)

    by davek ( 18465 )
    Alright, perhaps now I can convince my boss to allow us to support linux. Ah who am I kidding, he's a suit.
    -davek
  • No, no. You have to support the guardian with the servlets class, not kill him. Use the JSDK2.0 as a tool and extend the hell outta him.

    Then you can gain access to the next level, and eventually come face to face with the Jini.

    [sarcasm is good for the soul]

    MidKnight
  • Sure, Perl works... but internet applications are well on their way to being full-blown apps... ever tried to create a UI with Perl?
  • No associative array type.

    It's called java.util.Ha shtable [sun.com].
  • Wasn't the internet written in BASIC?

    I remember reading the code for it a ways back on my TRS-80...

    But it seemed rather slow; I didn't know why they didn't write it in machine code.

    I also read an article about converting salt into a clock radio


    ------------- Linux: Welcome to a GNU Generation -------------
  • SUN's Community Source License is indeed worth
    looking into. Especially if SUN speaks truly
    about Java Workshop going "Community Source".

    http://www.sun.com/workshop/java/index.html

    Seeing is believing...

    /Jocke

Slashdot Top Deals

You mean you didn't *know* she was off making lots of little phone companies?

Close