Slashdot Mirror
HP contributing to alternative JCK : Mauve
HP is
contributing Java technology to
Mauve, an alternative to Sun's Java Compatability Kit.
This is good news for alternative Java implementations such
as Kaffe, since the
JCK is only available for a fee.
Free Java will only be a success if people help
the free implementations. We all know that Sun
is not going to make the JDK into "free software".
That whole SCSL is a sham. It is not open and
not free. I think HP is doing the right thing
here by supporting a free compatibility kit.
How come people often use the word "technology" in association with Java? Why not use "source code" or "library implementation" or whatever term is appropriate to describe what is being talked about... That is unless you are a press release writer where your goal is to hype the product with impressive sounding terms.
If there is Java technology, there must be corresponding "C technology" to refer to C stuff. Perl technology? COBOL technology? Haskell technology? Visual Basic technology? Sure sounds silly to me.
Well, I can generate win32 x86 code from Visual Cafe. I don't see any particular reason why it should be impossible to do a gjj ;) or similar. Just that no one has done it yet.
(And that's the word: Yet)
fortunately this kind of disinformation have no place for most people. otherwise none would ever
:)
used a computer - every technology have been labeled as something bad several times in their
lifetimes.
anyway, sleep with this flame: perl is a very
crap technology, and so is linux
Thanks, HP, but your crap technology doesn't contribute. Stay away.
fyi the GCJ frontend to gcc is a part of the EGCS project
feel free to contribute to this wonderfull project
This is a little different than a front-end, but there are a lot of (insert-your-favorite-language-here)-to-C compilers out there. There's probably at least one of these for Java.
Gwydion Dylan, a free open-source Dylan compiler for unix systems, has a Dylan-to-C compiler which generates surprisingly speedy C code.
Crap technology... Do you mean Java?
// or something like that
myVector.Add(Integer(3));
Oh, the joy of primitive data types not being objects!
Thanks for the comment Scott McNealy!!
But If you noticed this *is* open-source software.
So, obviously the technology can be judged by reading the source code, and if it is "crap" (which I personally dont think it will be) it wont make it into the project if someone writes a better version...so if you think you are so good and have better technological capability, contribute to the project yourself!!!
do you have any idea what your talking about? sure primitives in java are not objects but your example dosent have a primitive in it, so i dont really know what your trying to prove. why ppl get hung up on this primitive thing is beyond me anyways. you dont have to use primitives if you dont want to. they are there primarily as a comprimise for speed and efficiency. nobody said you ever had to use them.
I think his point was that he can't just do myVector.Add(3), that is, 3 is a primitive and he has to explicitly create an Integer object from it to pass to the Add method. And how is he supposed to write 3 without using primitives?
Java technology... WooHoo!
MMX technology... WooHoo!
Alien technology... WooHoo!
OH YES I AM SOOOO HORNY BABY!
ROB, PLEASE STICK YOUR JUICY COCK IN MY TIGHT PUSSY YOU BIG STUD YOU!
OH GOD, MAKE ME CUM!!!
do you have any idea what your talking about? sure primitives in java are not objects but your example dosent have a primitive in it
t 2712
3 is a primitive.
If we want to talk about technology, compiler and language technology has advanced to such a state that pure object-oriented languages can be made to perform very well. Pure object-orientation makes programming cleaner and more consistent for the programmer. See "Primitive Types Considered Harmful" http://www7.software.ibm.com/vad.nsf/Data/Documen
Java garbage collector implementations have been on the low end of the technology scale too.
The Linux community needs to wake up.
HP does not have good intentions with the CTs.
The 100% pure movement needs to stand up
against this crap; it's the thin edge of the
wedge. If you've been following this at all,
you know that HP is in the MS camp over Java.
I'll have you know that I happen to be an ardent supporter of crap technology, and don't appreciate your looking down on it. Just like Java and Linux and C and internal combustion and pornograhy distribution, crap technology has it's place in the world too.
For an example, my bathroom is a very modern and comfortable one, thanks to crap technology.
> Crap technology... Do you mean Java? // or something like that
> myVector.Add(Integer(3));
ITYM: myVector.addElement(new Integer(3));
That you have to manually box and unbox primitive types is sometimes annoying, but it sucks less than absence of static type checking in Smalltalk, absence of dynamic class loading and code mobility in Eiffel and the fact that the only decent implementation of Dylan (Harlequin) ties you to Windows. Not to mention the vendor lock-in.
Crap technology: do you mean Smalltalk, Eiffel or Dylan?
HP doesn't really like java,
it likes windows and its microsoft owner.
If you like HP's Java, stay away from us you too, mr. Bill G. lover.
Why does it have to be? Sun invented Java. Doesn't it have the right to make it as open or closed as it chooses? Let's say you invented an application for, Linux let's say, and it was very popular and you are making a lot of money off of it. It is no popular in fact, that other people want it for their platform, say Windows or Mac. Now you don't have the time or interest to make 18 different versions of your program to be available on different platforms, but other people would gladly write these and maybe they even want to add new features for thier customers. They will do all the selling and everything, all they need is to base their code on yours. Now, will you give them all your code for free and let them profit of of what you made? Or would you demand some sort of royalty? This is very similar to what Sun is doing. If you want to modify their code to create a custom VM, and you then want to sell it, shouldn't they be entitled to a share of the profits? Also, say you want the people's ports of your program to be exactly the same as yours. Also suppose that this is a very complicated process, and you yourself have to do it to get it right. Doing this takes time away from your making money, so you charge for it. Is that so unreasonable? If it isn't unreasonable for you, then I dont see why it is unreasonable for Sun.
Exactly..I wonder what the good people at Sun is actually working on these days!
I too. Mere kaffe?
scrytch scribbled:
"""
The manual boxing/unboxing of types is annoying. That Java doesn't even have a list type is infuriating. What is this Vector nonsense? Hell, even most MUD server languages at least have actual list datatypes, many have associative arrays, most have list slicing and splicing literals as well. Not to mention there's usually some iteration functions/structures like map,
apply, foreach, and so forth.
"""
java.util.List
And of course, Java isn't the only language which compiles to JVM. If you hate Java, you can use e.g. JPython and write:
a = [1,2,3,4]
a = a + [5]
map(lambda x:x*x, a)
There is also Kawa which is rather good implementation of Scheme in JVM...
> erratum...
> myVector.addElement(new Integer(3));
add() is fine, if you're using Java 1.2 (since Vector implements Collection in 1.2)
Maybe you can do a duet with RMS.....
Nah
The code they are contributing will be free software. The goal is to produce a test suite for a 100% compatible Java implementation. If someone wants to extend this suite to test their own Java language extensions, that is fine with me. Isn't this what free software is all about?
myVector.Add(Integer(3)); // or something like that
Oh, the joy of primitive data types not being objects!
1) As someone else has pointed out:
myVector.add(new Integer(3));
2) Why not just use an array? If it's that important, use JNI. Or just code the whole thing in C/C++.
I need some time to sort out HP's devilish political maneuvering.
Hangover and out :)
--
--
=8^
myVector.addElement(new Integer(3));
mea culpa...
coffee... black... now...
--
--
=8^
Which story category is typically signaled
by an icon that looks like it could be Samuel
Beckett's vision of Don King's afterlife?
I'm afraid there isn't a whole lot of quality info available yet on this news. Here are some key points:
- A bunch of us have been hacking on Mauve for a while now.
- HP's contributions are pure JLS 1.0 compatability tests (for the libraries)
- They are Free - GPL'd Free - and come with no strings attached.
- It's a major contribution that benefits all of our Free software efforts.
- Anyone can contribute tests, fixes, changes, whatever to this project. Just subscribe to the lists.
- The folks at HP aren't just throwing this over the fence. They have a real desire to be good net citizens and participate in the maintenance of this project.
I, for one, am very grateful for HP's generousity. GCJ, Classpath and Kaffe will benefit enormously from their contribution.
AG - green@cygnus.com
Yeah.. Let's just all support HP's efforts. Let them release some half baked buggy source code, let the linux community fix the problems and watch HP make millions of $$$ for our efforts.
You have been assimilated.
Java is not just a language, it is also an
object orriented virtual machine and a set of
class libraries, that is why Sun refer to the
whole thing as a 'technology'. A GCC frontend
would (does) produce fine binaries, but the
result is then not portable.
The manual boxing/unboxing of types is annoying. That Java doesn't even have a list type is infuriating. What is this Vector nonsense? Hell, even most MUD server languages at least have actual list datatypes, many have associative arrays, most have list slicing and splicing literals as well. Not to mention there's usually some iteration functions/structures like map, apply, foreach, and so forth.
I've finally had it: until slashdot gets article moderation, I am not coming back.
> java.util.List
Which, as I've been saying over and over and over again, does not have any decent literal, slicing, or splicing syntax. Python's lists are an example of literal lists (I could do without its schizophrenic BASIC-inspired slicing syntax though)
As for JPython, just search for my posts whenever python comes up. I reserve even more venom for its idiotic treatment of whitespace, and its bleating hordes who beat me over the head with Microsoftian notions that it's actually some kind of feature.
I've finally had it: until slashdot gets article moderation, I am not coming back.
Good to hear that someone is filling this void. Sun's arguments against open source compatibility testing seem to amount to, "It's too complicated. Leave it to the experts. That will be $50,000. Sign here."
It's no wonder Java has been progressing so slowly in the market. And high time for things to improve.
An anonymous user wrote: Of course Sun has that right. And the market has every right to ignore Java as another proprietary technology. There is a strong business advantage to open source---independence from a single vendor. If Sun holds onto Java as closely as it's trying to do, it's hardly in its customers' interest to pursue the technology....
Rob Levin
Head of Operations, Open Projects Net
"Open source, open technology, open information"
Another anonymous user wrote: Why should we care? If they're producing source, check the license and make sure it's truly open. If it is, anybody can use it, modify it and redistribute it. If you don't like their compatibility suite, write your own or modify theirs.
The real problem here is that Sun has no interest in freeing up their implementation of Java; as long as there is no serious alternative to their work, who is really going to want another proprietary technology? So, if HP really wants to put out an open source compatibility testing suite, kudos to them!
Rob Levin, Head of Operations
Open Projects Net
"Open source, open technology, open information"
rcooper wrote: There's something wrong with HP making money? I'd much rather see someone profit from releasing open source (that I can continue to use and update) than from releasing software under NDA's and per-use commercial licenses. Give me something I can use and I'm happy as a clam to pay for it.
Rob Levin, Head of Operations
Open Projects Net
"Open source, open technology, open information"
Sure, there are those who say Java is what C++ should have been...
but then why not free Java from the portability hype, and the internet hype, and the compatibility hype, and just make a gcc frontend for it, just like C++.
Why not? because then java would be just another language, and not a bunch of hyped "technology."
I've been waiting for IBM to make a major Java/Linux announcement. While there is always Sun's proprietary (half-baked opensource to come soon apparently, but it's not Freed Software) JDK 1.2 out for Linux (I believe it is out), we need freed software alternatives. This is an excellent opportunity for HP to show that they care about GNU/Linux and freed software in general.
Cheers,
Joshua (who is going to free the Free Software Song in RealAudio, MP3, and a couple RIFF/WAVE formats soon--people need to hear that song!)
--jon. Postel is dead. May we all mourn his, and our, loss.
i know its slightly off subject... but how come so many people "Linux ports" are simply java?
sorrie just something that was anoying me, i`ll go back into my hole now
Scott Aaron Bamford (vpp) "We`re giving you the chance to skrew it up in a whole new and exciting way" sab@clara.net
"Pure object-orientation makes programming cleaner and more consistent for the programmer."
a per.html
..
..). Current native compilers get most of their speed out of the fact that they don't have to compile at runtime.
And not necessarily slower:
http://www.javasoft.com/products/hotspot/whitep
In short the whitepaper claims that Java will perform only slightly worse or even slightly better than C++. The hated Garbage collection stops are no longer an issue, nor is object reference (because it is now implemented as a C pointer), nor is thread synchronization,
Suppose they actually deliver half of what they are claiming. What reason would there be to want to natively compile your code (no more dynamic optimization, crappy garbage collection,
I know it's a whitepaper, it's from SUN and could hardly be considered objective. But still, hotspot is based on research that's been going on since the mid eighties. Check out Self for instance, a highly advanced delegation based language that performs well too (thanks to a predecessor of the hotspot technology).
Apparently hotspot is going Beta soon. I think it will permanently shut up people who claim that Java is slow, too slow and will never be fast enough. Sure it will take a while before it is actually useful (compare jdk 1.1, it only became stable enough at version 1.1.5 or so).
Jilles