• Back to all articles

Linux Needs Java, and Vice Versa

A Personal Note
I'm sorry for the lack of articles on freshmeat lately. I just bought
a house, and was busy for several weeks, then was off the net for ten
days. I'm back, and the article machine is cranked up again. Keep
those submissions coming. :)

Thanks for your patience.

Java needs to not be proprietary, though
One of the catches with Java is that Sun owns it and
maintains control of it. This means that the future of
Java on Linux is tied to whatever Sun perceives its
own self-interest to be. Sun can choose in the future to
neglect Linux as a Java platform, for assorted other
possible reasons., which would mean that newer
versions of the official Java would not appear on Linux.
It also means that if Sun dies or simply abandons Java
altogether, no one can pick up where Sun left off and
continue development. One could restart from scratch
or use a Java clone like Kaffe as a basis, but that it not
the same, not even close.

Re: Java needs to not be proprietary, though

> One of the catches with Java is that Sun
> owns it and
> maintains control of it. This means
> that the future of
> Java on Linux is tied to whatever Sun
> perceives its
> own self-interest to be. Sun can
> choose in the future to
> neglect Linux as a Java platform, for
> assorted other
> possible reasons., which would mean
> that newer
> versions of the official Java would
> not appear on Linux.

In theory, this is correct.
But I think the situation on Linux is much better than on
other platforms. For example there is still the Blackdown team
around which is still trying to catch up with latest JAVA
technology on Linux, on Windows, AIX, IRIX, etc. the
alternatives are rare.
At least it is always possible and legal to create a clean room
implementation of the JAVA runtime environment and some
people already did. JAVA is tied to SUN concerning future
directions, agreed but if SUN died or stopped Linux support
one day, the JAVA platform could still be alive for long. It
depends on what the Linux community would make out of the
situation.
And, to be honest, if JAVA died with SUN, the industry would
always find ways around by chosing different technologies.
Remember, JAVA is only a tool, not a (software) religion. I
always have to remind myself, that we can also live without
JAVA- Being a JAVA supporter for more than 5 years, this is
not always easy :-)

Oliver

Java performance... the same old stories...
Whenever I read about problems with Java performance, I believe that people are still stuck in the trenches of what Java used to be. Maybe the entry-level isn't as low as PHP solutions, but it is nothing like it used to be.
Also Tomcat is not really a speed-demon... I've done some tests comparing Tomcat with other servlet engines/app servers and it came in last by a long shot. If you want speed and can afford $1500 per server, then go with Orion (http://www.orionserver.com (http://www.orionserver.com)), a full J2EE app server. If that is too steep, Resin (http://www.caucho.com (http://www.caucho.com)) is only $500 per server (no EJB container though). If you have lots of cash, Macromedia JRun and BEA WebLogic are also good choices.

Enterprise?
Funny. I distinctly remember using remote method
calls, load balancing, fail-over and transactions
without java. Between engines like corba,
databases like postgres and languages like perl it
is perfectly reasonable and, I maintain, quicker
to develop "enterprise" functionality (as you
defined it) than using a pig of a language like Java.

The faults with java are numerous. Its slow,
people are working on that but we have fast
languages already. Its portability is rediculously
overrated, both perl and (to a lesser extent)
python run on pretty much any platform that
matters in a server environment. Suns brutal
control of the technology leaves users vulnerable
to all kinds of screwups, not the least of which
is the lack of serious investment in making it
faster, something which is clearly apparent.

Not only that, but its a shit language. Lessons
have been learned from decades of computer
language design, languages like erlang, ocaml,
eiffel, jade, perl and python, have designs which
make a serious fight against programmer flaws, not
just a sad and pathetic look in that direction by
removing pointers, but serious studied design.

Functional programming which massively reduces
complex state tracking in the mind of the
programmer as well as encouraging, far more than
the sad OO in java, real reusability. erlang
contains as part of the language an entire
serialized, function-versioned operating system
within which message passing and process handling
operate in a consistant, race free manner, making
large reliable applications far simpler to develop.

Perl made a huge leap into sense when it came out
with associative arrays (a heavily used
programming feature) built both reliably and
efficienty into the language, allowing the use of
both these and standard lisp/list arrays
throughout without creating a sickening dependency
tree which would result in reinvention.

Python, with its decision to utilise whitespace to
define code blocks, avoiding the matching-brackets
hell that many languages still suffer from.

OCaml, which won last years icfp competition to
create a raytracer in 72 hours, coming out not
only with a functionally complete tracer, but one
that performed incredibly well, due to the
techniques used in its compiler and the advantages
offered by a functional language in this respect.

Jade with its contract programming concept for
bug-free interface design.

This constant pushing of java as some kind of
solution to problems such as "Enterprise"
functionality and "rapid" development is simply
sad. It is not rapid, as a language goes it is
only marginally better than C++ in some respects,
and sorely lacking in others (templates),
development speeds in java, even with the large
number of packaged components available, are still
shockingly slow when compared with perl at the
small-program end, and functional languages in the
large-program end. This pretense that only java
can operate as a serious Enterprise language is
patently rediculous, "Enterprise" functionality
isn't even well defined (check out what our author
above had to say about it, he came up with four
terms which are covered by at least 5 general
languages I can think of off the top of my head,
and then waffled off into "other stuff"). There is
a reason why communications companies such as AT&T
use functional languages like erlang, they're
faster, more reliable and have considerably more
of the required functionality available right
there in the language, where all components can
make use of it, and where it runs at a decent speed.

Its a sad marketting ploy aimed at PHBs who know
no better. They don't understand that remote
method calls are available in every pretty much
every usable language still in existence today, or
that load balancing, failover and object brokering
can be provided in an entirely
language-independant way with any number of orbs
such as corba or language-integrated solutions
such as those present in erlang. They certainly
don't understand that in every way that I can
think of, short of hype, java is substandard in
comparison to at least one of the languages
mentioned here, and often all of them.

Don't propogate the myth. Linux only needs Java
because PHBs won't take any notice of anything
else. It is not a technical need, its a marketting
need, nothing more.

Re: Enterprise?
Funny, I don't remember ever in the history of software the mad and crazy desire to rewrite all the software in the world in one computer language. Funny how when one company, Sun, wants to push its mediocre technology on the world, that this goal becomes important.

Java is a stale design of an OOP language that even after five+ years of development by hundreds of companies still doesn't have a mainstream working compiler. Or memory manager. Or type system. Or generics. It is one step forward and five steps back.

If Freshmeat is going to support "editorials" which are just advertisement's for Sun's corporate agenda, well, I guess times are tough and you take money where you can find it.

If Linux goes down the Java path, they will end up just like everything else written in Java -- at the bottom of the dotcom scrap heap. As far as I know, Java still does not have one resounding success story. Sure, if I have a large programming staff and want to spend time exploring the deep dark secrets of the various VM's and enormously complicated "app servers", I can work with Java. But if I want to develop quality software in a reasonable amount of time, I'd pick something else to work with.

Java is still just hype. There is no good software written in Java. If you want to live in the real world and not the television world of corporate advertising, get away from this cancer. Python, C, C++, PHP, Perl, Ruby, Lisp, Prolog, etc., are all good languages free of the wretched design and corporate control that is Java.

-- Global Trance

Sun, the evil spawn of satan !
This seems to have gone the usual way of a Java pros and cons fight, which I guess we have all seen before.

The tried favourite of 'a big company owns' it therefore we must ostracize a technology is a little boring and frankly idiotic. Sun are actually responsive to developer requests and the language has evolved in the direction of the developer needs.

"its a shit language" (yawn, here we go AGAIN) AT&T on one side.. Motorola or company X on the other. Fundamentally there are far superior / skilled / experinced programmers than you or I around that would disagree. Dont like Java, dont use it. Generic opinions dont sway the argument either way.

"Java is still just hype. There is no good software written in Java.", "Sun, wants to push its mediocre snip.." blatantly wrong and misinformed. Take a look on the bleeding edge of the development, you will see that Java technology within i.e phones etc is now starting to take effect. The technology has matured sufficently, a 'stale oop language ' it definately is not. Yet again, mindless blind following of 'Its a big company and therefore we must hate it' regime.

Cant wait for the .net "its microsoft therefore it sucks (because I saw someone else say that)" where the technology is ignored and the easy rant begins... not.

As for speed, Nadir is right on with the choice of server. JRun 3s well priced and no speed complaints at all with it.

Cobalt
I seem to recall that Sun was intending to move the Cobalts to Solaris. Has that happened yet?

cold coffee anybody?

Java has a number of disadvantages which just haven't been addressed in the 6 years or so that it's been out there... There are two distinct markets for a programming language, big and small applications. Java might be useful for a BIG project ie loads of simultaneous developers who have difficulty communicating :-) However, for most of us who have "small" projects with less than 10 simulaneous developers...

Firstly, its slow... even compared with plain interpreted perl (don't bother comparing it against mod_perl :-)

Secondly, its quite difficult to deploy to a clients' machine... install.sh anybody?

The only thing that java has going for it is marketing... it's fully buzzword compliant... unfortunately for us techies... this is not a good thing :-(

I'm sorry, did we forget about somebody here?
I hate to be a naysayer but did we forget about our little buddy Solaris during this great brainstorm?

Sun doesn't really care about bringing Java support to Linux. In fact, things like Blackdown and JBoss are around it seems inspite of Sun's ignoring Linux almost completely. I guess we should atleast be lucky that Sun is releasing a fully endored JVM for Linux now, but as far as the operating system du jour for Sun it's going to be Solaris, and they show no signs of changing that.

That's because Solaris foremost supports Sun's core business: selling big fat servers. Having Linux around doesn't help their core business. Remember, vendors sorta like vendor lock-in, don't they? It keeps you buying their stuff. If Java is an open platform (as open as Sun is comfortable with, but i'll talk about that in a second) they have to have a way to keep you buying those big E4500s, that's by having the best platform for Java, and catering to the buyers that purchase enterprise class servers: big enterprises.

And don't forget the whole licensing impedance. Sun likes to make Java as "open" and "cross-platform" as they feel comfortable with yet still keep it under strict Sun control. They brow-beat Microsoft for that one. Which is probably what spawned this whole .NET and C# hubbub. "You may have won the battle, Scott McNealy, but you haven't won the war!", shreiks Bill Gates with his fist waving in the air. The problem is that Sun's tendency to control just won't jive with the Linux crowd, which will always strive to be open.

It may seem like a match made in heaven, but I don't think either one is ready to sign the prenup.

http://unixpunx.org

Re: cold coffee anybody?
If I may, I'd like to address these 2 common misconceptions really quickly...

> Firstly, its slow... even compared
> with plain interpreted perl (don't
> bother comparing it against mod_perl
> :-)

Don't tell that to the people who've discovered the Caucho application server, atleast. You want benchmarks, you got em

> Secondly, its quite difficult to
> deploy to a clients' machine...
> install.sh anybody?

The whole point of having platform independent Java is that the code is going to be running in the Java environment, not the native one. Then again, who needs to install when you just package up all your portable byte code into a wonderfully elegant cross platform parcel called a JAR and distrbute that? So even if you are sloppy and your Java runs slower than another compiled language the user will make it up now having to spread out your tar ball and compile it themselves.

I think these comments miss the point
I think all of these comments are missing the point of this editorial. I am a die-hard linux fan. I use python, C, C++, VB (unfortunatly) and ASP (see previous). This isn't about the age old "Sun Sucks" argument. This article is meant for sun as much as for the community. Has sun neglected linux in the past, yes, but I'm saying that it's in everyone's best interest to WORK TOGETHER.

While this editorial was written towords sun, there ARE other's in the Java game, namely IBM and HP. While I would love to see sun open java up, I am not worried. As was mentioned above java is not a religeon.

To respond to attacks on what I've said about the current state of Linux in "Enterprise" functionality, Java is a great integration platform, hence JNI, RMI-IIOP, EJB, XML-RPC, SOAP and anything else you want to integrate with it. The EJB2.0 Specs require RMI-IIOP, thus allowing tight integration with CORBA.

I'm not going to respond to the cry of "OOP sucks". OOP, as everything else, has it's place.

.NET is comming, and anyone who thinks it's going no where fast is kidding themselves. This idea has been re-inforced by the anouncement of dotGNU and Mono. It has also been re-inforced by the GNOME and Bonnobo project.

This is a call to work together, not to fight.

Linux Needs Java ish
It seems to be looking at it from 'vendor (be it sun, hp or anyone else) should look to try and strengthen and develop the Java/Linux relationship'. To that extent, it is hard to comment any other way.

I think, more fundamentally, if the view of working together is adopted then the 'Java/Linux vs .net' attitude must also be dropped. Doing so creates the fruitless arguments over whats better or not and leads to ' x product by x sucks'. Of course .net will do some things right, even better than Java while Java will have its own strengths.

Unfortunately, the hole is that Linux is simply not a commerically appealing option for most vendors (this is open to much argument I know..)
The .net plan most certainly is and, in a time when IT cost cutting is in full swing and likely to be for some time, the companies such as HP will go with the more profitable option and no one can blame them. It is of course, very appealing to the informed customer who wants to save where possible.
Bottom line is, I think your argument is totally right and the relationship should be helped as much as possible to 'rival' but not compete with .net. To what extent this will happen depends totally on the profits avaliable by doing so, .net will, I think, be a spot of light in terms of vendors making money, a low powered but effective Java/Linux platform probably wont be the cash-cow that a .net solution will be to its suppliers so its development perhaps isnt on the forefront of many vendors horizons I fear.

Java, the new concept of programming: IF (), WHEN (), SHOULD (), MAY ()
What really disturbs me are all these variable conditions you draw to make us believe in Javas constant success...

But from top to bottom:

> the Web is constantly changing

Why not and why python and others should not be the ones who do that?

> There is already JBoss

There is already much in many languages...

> available for free and completely community supported

except the language used!

> Java is not an easy language to learn for beginners

Even professionals decide for scripting languages because of effectiveness in specific case

> The first is the myriad of languages that will run on Java,
> the most popular of which is JPython.
> The second is the JNI (Java Native Interface).
> The JNI allows Java to call native code in C/C++, Python, Perl,
> or any other language with a JNI binding.

You could use them directly! Why should one brake out his most loved language for Java. For an example one can serve servlets with PHP. If one really wants Java one will write in Java. If one wants to program in C/C++ one can use their libraries. Maybe you really believe that only Java comes with libraries and classes for the internet. And browsers could plug in not only Java Virtual Machines but just security-proofed bindings to locally installed interpreters. This would be the better way in any case, even for Java. But I don't only need performance, I would like to manipulate HTML objects with languages other than JavaScript. Even Java doesn't give me that!

> Java is not exactly known as the fastest system in existence, but the future...

Yeah, here begins what I mentioned in the Subject!

> This is where Sun should come in...

Will they?

> Sun now owns Cobalt...

And whom next? >

If GCJ can be completed...

Doesn't it work right now? So I will wait and stay with PHP ;-) And if it then is platform dependant... (You know what I mean?)

> Sun can have high-performance...

Nice! How does it help me now?

> There is another reason Sun should work with the community...

You should tell it to them and not to us!

> Finally, there is a reason for Sun to help increase Java's integration with languages
> such as Python and Perl.

Look up one answer!

> All of these pieces fit together to allow a beautiful friendship between Sun and Linux.

If they like (or want) to, they can do. But does this mean, that other friendships are to be laid down?

> Once a company is ready to move from small server appliances to enterprise servers,
> who are they going to call? You guessed it: Sun.

And this is the aim of linux community?

> Sun, Linux, and the Open Source Community could make a great team to help
> extend the causes of everyone involved.

Yes, they COULD but must not. Others could and can do actually. Will Java beat them? Is that an advantage? Under what circumstances? Maybe IF Sun thinks about their licensing model etc. There is much to talk on first.

Really, there is not even one single argument in your essay that sticks out why one MUST follow your opinion. There is nothing on it making Java the most needed language. And there are too many dependencies that are not controlled by you or us.

Come back later...

Re: Sun, the evil spawn of satan !

> This seems to have gone the usual way of
> a Java pros and cons fight, which I
> guess we have all seen before.
>
> The tried favourite of 'a big company
> owns' it therefore we must ostracize a
> technology is a little boring and
> frankly idiotic. Sun are actually
> responsive to developer requests and the
> language has evolved in the direction of
> the developer needs.
>
> "its a shit language" (yawn,
> here we go AGAIN) AT&T on one side..
> Motorola or company X on the other.
> Fundamentally there are far superior /
> skilled / experinced programmers than
> you or I around that would disagree.
> Dont like Java, dont use it. Generic
> opinions dont sway the argument either
> way.

What generic opinions *looks up* you mean the
distinct and, thus far in this thread, uncontested
rundowns of the explicit failures of Java as a
language?

I don't like java, and I don't use it. This would,
however, be a public debate on the merits of Java
as a technology. Therefore you need input from its
most ardent admirers as well as its most adject
despisers. If you just had the admirers talking
here, you could have avoided having an article at
all.

> "Java is still just hype. There
> is no good software written in
> Java.", "Sun, wants to push
> its mediocre snip.." blatantly
> wrong and misinformed. Take a look on
> the bleeding edge of the development,
> you will see that Java technology within
> i.e phones etc is now starting to take
> effect. The technology has matured
> sufficently, a 'stale oop language ' it
> definately is not. Yet again, mindless
> blind following of 'Its a big company
> and therefore we must hate it' regime.

According to me, it is mediocre. Not only that,
but Sun is pushing it. I'm not sure I'd go so far
as to say no good software is written in it. I
will go so far as to say that there is no good
software in Java that could not have been done
better and faster in a different language.

I am not mindlessly doing anything and your
dismissal of a not inconsiderable listing of Javas
faults as mindless does nothing to convince me
that they may be untrue.

> Cant wait for the .net "its
> microsoft therefore it sucks (because I
> saw someone else say that)" where
> the technology is ignored and the easy
> rant begins... not.

Yes, down with rants where the technology is
ignored! oh the irony

> As for speed, Nadir is right on with
> the choice of server. JRun 3s well
> priced and no speed complaints at all
> with it.

Compared with what? is it good enough for the job?
good for you. But if you're in no hurry, riding on
the back of a slug is fast enough. Have you
compared these things with the more versatile
languages available today?

beyond internet application serving
So far the discussion has avoided one of Suns aces in the pack..
The whole wireless entertainment market is estimated at 5 billion in 2005. Sun has a nice solution 'ready to go' with J2ME / CLDC. The key thing is you have the worlds leading manufactures sitting round a table, listening and actually doing it in Java (in the lucrative wireless market with Nokia and friends). Arguably, love / hate a language, its only as useful as the pay check it brings with it and in this pretty much virgin territory Sun has a solution with Java. Anyone know of Python mobile in the making...

Let them conform to us...
Python has the potential to be a cross platform
OO distributed component system. It will still
need an application framework and something that
is 'not' I repeat not tied to gnome or kde but plays nice with X in general and other OS settings.

Another option is bell labs inferno VM which is really their plan 9 OS. It's fast, it's clean, & the source is available. It's also faster than Java and more flexible, my opinion.

We could LEAD with a new component model
that would first encompass all unices, MAC, & Be.
Then the windows commmunity can then say to themselves like it seems the linux community is doing now with .NET. "Gee the water looks warm over there, I think I'll take a dip". It doesn't
have to be an assault against .NET alone. Delphi
is a perfect example of a programming model that
was considered better than VB and still has to
this day obvious advantages like SPEED.

You want to start a fight? First you have to pick
your target.

Peter

He's Right.

Divided we fall.

I've been complaining the same thing as the editor, but there are too many "my way is the only way" people out there to listen.

Sure, Python, Perl, OCmal, Ruby, x86 asm, etc, etc, may be great languages, but they don't have the gigantic push and cash that Sun has been putting into Java.

There are courses offered in Java, certificates that head hunters can look for on resumes, there is Java for small devices, Java for big enterprise-scale solutions. Java has the backing of IBM too, infact the best runtime and compiler for Java right now is not done by Sun, but by IBM. Sun's documentation on their site can take you from knowing nothing about the language to being pretty proficient with it, you can then spend time reading about almost any aspect of Java (from EJB, to JES to the KVM). When you're done you can head over to Inprise and download JBuilder for Linux, or use Sun's Forte, and over to IBM to grab WebSphere, or over to JBoss and setup a system scalable, secure, and powerful enough in you own bedroom to run a national bank.

There's an entire Java-based industry out there.

The Java naysayers; "Java is slow, Sun doesn't support Java on Linux, Sun has absolute control on Java etc.." should also be going on about how "Linux doesn't support 3D, Linux doesn't support SMP, and the GPL is bad", because it's along the same lines.

The editors point is simple. Let's leverage the work being done by Sun and promote Linux as a good OS to run Java on.

Everyone jumped on him, talking about how OCmal is better and how Python is better, etc, etc.. Where's the Python certificates that head hunters can look for, where's the enterprise-scale Python roll-outs? How many web pages use OCmal for their server-side scripting language? The answer is: There aren't any, yet.

And where there are, I'll be behind Python and Ruby, and whatever other languages too. But right now, let's just stop arguing about that, and instead of being used by a big company, use them instead. Sun has cash to throw at Java, so let's use them!

When Microsoft comes out with C# and .NET, there's going to be a flurry of activity and news, and information, and hype, and none of it is going to be on Linux.

United we stand.

Replies to replies
Ok,
One common thread i am seeing is 'linux isn't a solaris' step up!'. No it isn't, but it's already being used in simmilar fashions by IBM and HP, as well as caldera. Linux is being used as a steping stone to the more traditional high end os'. I am simply giving a reason to sun (ibm, hp...) as to why they should help with improving java on linux.

As for language support and licensing from sun, again there's ibm hp and the open clones. THIS IS AS MUCH A PLEA TO SUN, IBM, HP... AS IT IS TO THE COMMUNITY!!!!!!!!!!!!!!!!!!!!!!!!

Re: Java needs to not be proprietary, though

> One of the catches with Java is that Sun
> owns it and
> maintains control of it. This means
> that the future of
> Java on Linux is tied to whatever Sun
> perceives its
> own self-interest to be. Sun can
> choose in the future to
> neglect Linux as a Java platform, for
> assorted other
> possible reasons., which would mean
> that newer
> versions of the official Java would
> not appear on Linux.
> It also means that if Sun dies or
> simply abandons Java
> altogether, no one can pick up where
> Sun left off and
Well, Sun do release the sourcecode...
There is the Blackdown project, www.blackdown.org
IBM is also big in the javaworld as well as the linux
world, and they provide a up-to-date VM for Linux.

Re: cold coffee anybody?

> Java has a number of disadvantages which
> just haven't been addressed in the 6
> years or so that it's been out there...
> There are two distinct markets for a
> programming language, big and small
> applications. Java might be useful for
> a BIG project ie loads of simultaneous
> developers who have difficulty
> communicating :-) However, for most of
> us who have "small" projects with less
> than 10 simulaneous developers...
>
> Firstly, its slow... even compared
> with plain interpreted perl (don't
> bother comparing it against mod_perl
> :-)

Do you have any benchmarks to prove this, or are
you
using a very old VM, with not for of JIT
compiler?
> Secondly, its quite difficult to
> deploy to a clients' machine...
> install.sh anybody?

Ant, jakarta.apache.org
Or InstallShield for java.

>
> The only thing that java has going for
> it is marketing... it's fully buzzword
> compliant... unfortunately for us
> techies... this is not a good thing

The fact that you can write apps in 'notime', and
still
maintain clean code with a good and
extensible design doesnt matter?! (i have yet to see
that in a perl application)

Re: beyond internet application serving

> Python mobile in the making...

It's called pippy...
It's a python interpreter that runs on my Palm m100...
You can look it up.

Re: He's Right.

>
> Divided we fall.
>
> I've been complaining the same thing
> as the editor, but there are too many
> "my way is the only way"
> people out there to listen.
>
> Sure, Python, Perl, OCmal, Ruby, x86
> asm, etc, etc, may be great languages,
> but they don't have the gigantic push
> and cash that Sun has been putting into
> Java.
%

Are you kidding?

You are saying there are too many people
saying my way is the only way and in
the next paragraph you say we should use
Java because Sun has the money to push it...

What smack are you on?

Java isn't write once, run anywhere...

The only thing that Java has to offer for itself
is that it is object based and runs on more
then 1 platform. That's all it has in 6+ years
of development.

Java is useless as a Client side development
language... don't tell me this is not true...
look at Limewire, on my 2x 800MHz CPU box
at home it is slow... It's a pig.

Java doesn't scale, it never has and
it may never... yet to be seen...

> JBoss and setup a system scalable,
> secure, and powerful enough in you own
> bedroom to run a national bank.

Please I beg you... show me one application
written in Java/J2EE that scales to the extent
it could support a national bank...

> The editors point is simple. Let's
> leverage the work being done by Sun and
> promote Linux as a good OS to run Java
> on.

If Sun has the money, they can pimp their
own product, I will use what ever is best
for the job...

You are missing the point of Linux...
Linux is freedom to choose, freedom
not to choose Microsoft, not to choose
Sun, but to choose between MANY
alternatives.

Competition is good, the more competitors
the better.

O'lay!

-Ironstorm

Re: He's Right.

> You are saying there are too many people
> saying my way is the only way and in
> the next paragraph you say we should
> use Java because Sun has the money to push

I'm not saying you should use Java, I'm essentially saying that
you should cut Sun some slack, and possibly add Java to the skill
set you already have. Chances are you program in at least 3-4
languages already, and when your company decides it wants to use
Java, remember to mention Linux and that Linux is a great platform for Java.

> Java isn't write once, run anywhere...

And Linux has no usable desktop. Next.

> The only thing that Java has to offerfor itself
> is that it is object based and runs on more
> then 1 platform. That's all it has in 6+ years
> of development.

That's one way of looking at it, if you ignore everything
else that's been going on in the past 5+ years.

> If Sun has the money, they can pimp
> their own product, I will use what ever is
> best for the job...

Then you can expect to have Linux trampled on in the upcoming fight of Microsoft
vs Sun.

> Linux is freedom to choose, freedom
> not to choose Microsoft, not to
> choose Sun, but to choose between MANY
> alternatives.

You just shot down my choice to use Java on Linux, and flamed me
and Java in the process.

> Competition is good, the more competitors the better.

Too much competition causes over saturation and nobody benefits,
not customers, not companies, not developers, not Linux.

Re: He's Right.

>
> % Competition is good, the more
> competitors the better.
>
> Too much competition causes over
> saturation and nobody benefits,
> not customers, not companies, not
> developers, not Linux.
>

Having lots of competition with the ability to
interoperate via standards of data exchange
(ala XML) negates this argument.

Both consumers and companies can benefit
from the absence of vendor/platform
lock in (except vendors, i.e. Sun) -> the
price point can then approach 0... 0 is
a very good number have in the expense
column for anyone.

The computer hardware industry is an
example of this... the cost of PCs is
continously falling, there are tons of PC
vendors - Dell, Compaq, IBM, mypcshop.com
- building PCs out many different types of
hard drives, video cards, etc... All of
which interoperate.

This good for both corperate customers and
consumers at large.

More competition > Less competition.

Re: cold coffee anybody?

> % Firstly, its slow... even compared
> % with plain interpreted perl (don't
> % bother comparing it against
> mod_perl
> % :-)
>
>
> Don't tell that to the people who've
> discovered the Caucho application
> server, atleast. You want benchmarks,
> you got em
>

Not really trying to pick sides, but I don't think I'd really trust benchmarks that put java on top from a company that is trying to sell java

Re: He's Right.

>
> Please I beg you... show me one
> application
> written in Java/J2EE that scales to
> the extent
> it could support a national bank...
>

American Express, Banc of America Securities, Banco Zaragozano (heh), Bank of America, Bank of China, Bank of Ireland, Bank of New York, Bank of Tokyo-Mitsubishi, Bank One Coporation, Banque de France, Banque Nationale De Paris, Capitol One, Chase Manhattan, Citigroup, Commerzbank, Coutts Bank, Credit Agricole, Credit al Industrie, Credit Lyonnais, Credit Mutuel d'Anjou, Credit Suisse, Deutsche Bank, Fannie Mae, Fidelity Investments, First Union Corporation, First USA, FleetBoston, Freddie Mac, GoldBank, Goldman Sachs, Industrial and Commercial Bank of China, Italian Ministry of Finance, J.P. Morgan, Lehman Brothers, MasterCard, MBNA, Mellon Financial, Royal Bank of Canada, Royal Bank of Scotland, Sumitomo Bank, Union Bank of Switerzland, U.S. Bancorp, Wachovia, and Wells Fargo are just a few examples of large banks or bank-like corporations that have successfully deployed enterprise scale solutions running in Java on BEA's Weblogic.

Here are some customer profiles describing the applications:

http://www.bea.com/solutions/capitalone.shtml
http://www.bea.com/customers/profiles/chase.shtml
http://www.bea.com/solutions/citigroup.shtml
http://www.bea.com/customers/profiles/deutschebank.shtml
http://www.bea.com/customers/profiles/etrade.shtml
http://www.bea.com/customers/profiles/italianminfinance.shtml
http://www.bea.com/customers/profiles/royalbank_scotland.shtml

Hey and I don't even work for/with BEA. I just think their product kicks ass and I've seen Java do amazing things. I hate to see people shooting their mouths off about stuff they don't know, and unfortunately this forum is full of it... not just in the message I'm replying to either, but just about all of them. This is ridiculous.

Re: He's Right.

>
> %
> % Please I beg you... show me one
> % application
> % written in Java/J2EE that scales
> to
> % the extent
> % it could support a national
> bank...
> %
>
> Here are some customer profiles
> describing the applications:
>
%
> http://www.bea.com/solutions/capitalone.shtml
%
> http://www.bea.com/customers/profiles/chase.shtml
%
> http://www.bea.com/solutions/citigroup.shtml
%
> http://www.bea.com/customers/profiles/deutschebank.shtml
%
> http://www.bea.com/customers/profiles/etrade.shtml
%
> http://www.bea.com/customers/profiles/italianminfinance.shtml
%
> http://www.bea.com/customers/profiles/royalbank_scotland.shtml
>

Links are good, except I was looking for
large scale applications developed in J2EE,
not press releases about customers who have
licensed BEA's weblogic app servers...

The only one on that list that has an APPLICATION
to show for there license of any significance is
e*trade's mutal fund program... I haven't used
e*trade so, I can't judge it's sucess... interesting
that there is some evidence to suggest it can be
done though. :)

My company is evaluating Websphere (from IBM),
but we haven't built an enterprise scale
application with it yet... we are not even
sure if it will scale.

-Ironstorm

Re: He's Right.
Well ...

I am using Java on Linux and i am very happy with it building a kinda bigger publishing system. The first generation years ago had been a Perl implementation. To complicated to manage 100.000+ lines of code ..

This version runs some sites already saving lots of money for the users and generating income for me.

So, without Java and Linux the whole process would have been harder at least.

I also use PHP and Perl as well as VB on the Client side. Its always fun to find the latest Bug after a new service pack or DLL comes up. MS-XML? Which version? Which namespace error ... Oh well. But even in this not so perfect environment VB makes a bit of sense here an there.

The author`s point is, that Linux and Java could help each other big time and thats absolutely true.
If Java and Linux are a good match, or a bad one, this won`t hurt a Linux/Perl or any Phytonic solution.

Its not about the only solution to a problem.
If the Sun JDKS sucks, try IBMs ...
If tool A sucks, try tool B.

These religious wars are from last century.

Just do not waste your time in these worthless
fights and just write better products.

My app runs on Windows too .. thanks to Java. But it is slower (overall performance on linux very good ... but an appserver ain't no graphics engine which should use C or whatever..) and anything on Windows is more instable. So i am not endorsing it.

So for me Java and Linux is a good match. Plain easy. If a Perl solution makes sense, well .. i am using mod_perl etc.

And ... the internet market is not dead.

Juergen

Re: He's Right.
Iron Storm,

Your full of SH*T. Doesn't scale? No one uses it? I don't know if they still publish it, but Devx.com's enterprise today is nothing but case studies at the enterprise level in several technologies, including java. How about schwab and home depot (yes i know, not a bank). AND THIS WAS A YEAR AGO!!!!!!

Re: Java needs to not be proprietary, though

>
> % One of the catches with Java is that
> Sun
> % owns it and
> % maintains control of it. This
> means
> % that the future of
> % Java on Linux is tied to whatever
> Sun
> % perceives its
> % own self-interest to be. Sun can
> % choose in the future to
> % neglect Linux as a Java platform,
> for
> % assorted other
> % possible reasons., which would
> mean
> % that newer
> % versions of the official Java
> would
> % not appear on Linux.

> And, to be honest, if JAVA died with
> SUN, the industry would
> always find ways around by chosing
> different technologies.
> Remember, JAVA is only a tool, not a
> (software) religion. I
> always have to remind myself, that we
> can also live without
> JAVA-

True, but if people were to follow the editorial author's
advice and push more for Java on Linux so that Java
became a more entrenched part of the Linux
landscape, then it would be more difficult to recover if
Java were to go away.

Re: Java needs to not be proprietary, though

> Well, Sun do release the
> sourcecode...
> There is the Blackdown project,
> www.blackdown.org
> IBM is also big in the javaworld as
> well as the linux
> world, and they provide a up-to-date
> VM for Linux.
>

Sun has not released the source code of Java to the
general public. Also, both the Blackdown and IBM
JVMs are based on Sun's code, especially the
Blackdown JVM.

Re: He's Right.

> The computer hardware industry is an
> example of this... the cost of PCs is
> continously falling, there are tons of
> PC vendors - Dell, Compaq, IBM,
> mypcshop.com - building PCs out many different
> types of hard drives, video cards, etc... All
> of which interoperate.

Good example. How many different video card vendors
are there out there that do their own hardware? ATI,
Matrox and then all the NVidia licensed cards.. How many
of those do 3D, fast, and well? The NVidia ones.

How many different types of CPUs are there in your average PC? (I'll ignore the PPC guys for now :), There's 1. An x86 compatable one. How many different vendors? Two (or three if IBM still makes Cyrix), AMD and Intel.

Soundcards? There's the cheap crystal sound stuff and there's Creative Labs (yes, there are a lot of others, but those are the main ones).

Hard drives? I don't even know anymore. Quantum was bought by Seagate or something? We'll say there are 3-4.

Java has name recognition among businesses now. Imagine if whenever someone thought "Java" they thought "and of course we'd use Linux to run it on."

It'd be like people saying "Well, we bought an Intel Pentium, of course we'll run Linux on it."

Amusement

I am always amused by these kind of
discussions. A whole bunch of people who
generally have no idea jump up to offer their
"opinion" (aka FUD and disinformation).

Here it seems to be anti-Java so I will address
that side. The java "is too slow" camp or those who
propose solutions like gcj for speed
improvements is amusing. If they knew anything
about the state of the art they would know that this is
in general a fallacy. Some JVMs can also run
faster than natively compiled and optimized
code (The common example being IBMs JVM vs
VisualC++ compiler). At one stage Java was too
slow for many things but now this is not really the
case.

It is true that there is many things suck in java.
Examples include; string manipulation is Unicode
and immutable, memory usage is anywhere between
1.4-2.5 times more than c (this is where solutions
like gcj and jet come in), startup time is long (will be
solved in jdk1.5 via Isolate API) etc. There is plenty
more issues with java - especially when you look at
some of platform implementations (Swing is a memory
hog and when poorly used is slow).

Other things that suck about java include things
like we had to wait till jdk1.4 to get basic
functionality like file mapping into memory,
interaction with native memory, async io etc

Another somewhat amusing stance I noticed was
that java is not enterprise ready. Java however
happens to be the only platform I know of that
already has standard interface for complete suites of
enterprise pieces - ie databases, "enterprise"
objects, messaging, directory services, user
messaging services etc. There are other solutions
out there or will be out there in future (ie CORBA is
cross vendor, SOAP will be cross vendor, LDAP is
cross vendor, ODBC etc) but no complete
programatic or conceptual model. Hence why you
have seen the relatively quick adoption of J2EE
despite some of it's limitations. If anything enterprise
represents one of javas success stories.

There also potential sucess stories brewing with
the smaller end of spectrum (telephony, set top
boxes, gaming consoles etc).

Whether java needs linux or vice versa I am not
sure. If you want linux to become a real enterprise
platform then java becomes essential. If you want it
as a desktop platform then java is not needed (QT is
far easier IMHO).

Anyways I assume it is now time for a troll to
jump up and say things like; "Linux is already doing
X on platform Y" (replace X with some feature you
want and Y with the platform like enterprise or
embedded). Thats nice but without a standardization
process ala JCP and some corporate sponsorship X
will continue to be a minority player.

Re: cold coffee anybody?

> Do you have any benchmarks to prove
> this, or are
> you using a very old VM, with not for
> of JIT
> compiler?

"perl -Tw hello.pl" -v- "java Hello" try it for yourself...

> % Secondly, its quite difficult to
> % deploy to a clients' machine...
> % install.sh anybody?
> Ant, jakarta.apache.org
> Or InstallShield for java.

I've used these... but for ant... you need a working java environment... etc...

> % The only thing that java has going
> for
> % it is marketing... it's fully buzzword
> % compliant... unfortunately for us
> % techies... this is not a good thing
> The fact that you can write apps in
> 'notime', and still maintain clean
> code with a good and
> extensible design doesnt matter?! (i
> have yet to see that in a perl
> application)

quite simply it take more lines of code to do simple things in java... simple one liners in perl take a page in java with imports... regex anybody? $t =~ s/^k.*(\d+)\s+(\d+).*(\d+).*$/$3/s; print $2.$1;

the more lines of code there are the higher the chances of a mistake... and the harder it is to prove a program is correct. Java is just as obscure as the next language... oo with lots of files just makes things harder to keep a track of in your head... C++ is worse in this regard... wonderful for world class experts, but in the hands of normal people it just doesn't cut it...

i've been at this game for 20 years and i've seen more "wonderful" languages than i care to remember... java just doesn't do it... perl ain't it either... but its a hell of a lot closer! oh... and you don't have licencing issues...

Re: Amusement

> I am always amused by these kind of
> discussions. A whole bunch of people
> who
> generally have no idea jump up to
> offer their
> "opinion" (aka FUD and
> disinformation).
%

Which is exactly what you have done.. (your opinions are more valid because.. ? :)

> If they knew
> anything
> about the state of the art they would
> know that this is
> in general a fallacy. Some JVMs can
> also run
> faster than natively compiled and
> optimized
> code (The common example being IBMs
> JVM vs
> VisualC++ compiler). At one stage Java
> was too
> slow for many things but now this is
> not really the
> case.

I think we are all aware of the virtues other JVMs offer. Again right off the discussion topic..

> It is true that there is many things
> suck in java.

Really !? thought it was flawless..

Its very easy to pick a hole in a language and use it as a stick.

> There also potential sucess stories
> brewing with
> the smaller end of spectrum
> (telephony, set top
> boxes, gaming consoles etc).

Yes.. J2ME CLDC / CDC. This was touched upon above. Id question it as 'potential', its a real market evolving and happening right now with standards being agreed with developing companies. It is in one view, already a success. The hurdle of getting the technology there and implemented has been crossed. If it will be used doesnt need answering, the market is lucrative and its not overcrowded yet, nuff said.

> Whether java needs linux or vice versa
> I am not
> sure. If you want linux to become a
> real enterprise
> platform then java becomes essential.
> If you want it
> as a desktop platform then java is not
> needed (QT is
> far easier IMHO).

This wasnt the discussion. The hub was if Sun (and friends) should be pushing it hard on Linux as a competitive answer to .net. Id say definately yes, 'Enterprise' (nice open interpretation buzzword) or not. Whinning and paranoia about Sun because there a big company aside, the two complement themselves very well in a cost effective foundation. The problem IMHO is Sun are, in a way, shooting themselves in the foot over Solaris.

>

Be flexible
Most people tend to say: use Java or don't use Java. I think we should try to be more flexible in choosing our languages, currently I use a lot of Java because it seems practical but I definitely will try switch to something better designed for the next project (here are my 2c why Java is not well designed: http://www.peterbecker.de/texts/javacritique.html -- very early version, focus is on SW-Engineering/OO aspects).

The question is not if Linux needs Java or not, the question is neither if we should use Java or not, the question should be: is Java the appropriate tool for what you want to do. This is hard to figure out, but it is definitely worth some effort -- the programming language is one of the most important tools in your toolbox and you should choose your tools wisely (but without falling into analysis paralysis).

Trying to do everything with one language is like using a Leatherman to fix a car -- it will evolve into a huge and clumsy tool. And the usual argument that the costs of learning a new language are too high is short-sighted and usually ignores the fact that learning a language might give you some new insight -- esp. when it was hard to learn (assuming it is reasonable well designed).

And don't forget about tools like DBMS or Corba -- they allow you to use a number of different languages in one project. This has some drawbacks due to the overhead in build/deployment management and the fact that you need more skills in your team, but I think it is worth the effort and additionally ensures a clean design -- taking shortcuts across such boundaries is far harder than just declaring a private member public ;-)

Just some late thoughts,
PeterB

PS: PhiRatE -- did you get to use all this languages on your job? When can I start? *g

Re: Amusement

>
> % I am always amused by these kind of
>
> % discussions. A whole bunch of
> people
> % who
> % generally have no idea jump up to
> % offer their
> % "opinion" (aka FUD and
> % disinformation).
> %
>
> Which is exactly what you have done..

Don't be stupid. Nothing that I have mentioned is
FUD or disinformation.

> (your opinions are more valid because..
> ? :)

Maybe because I have written my own JVM.
Perhaps because I partipate in JCP (the
organisation that standardizes language specs and
extentions). Perhaps because my work has hit the
barriers in language and also highlighted it's
strengths. Could be because I have used it
essentially since the begining. Any of those answer
will work I guess.

> % If they knew
> % anything
> % about the state of the art they
> would
> % know that this is
> % in general a fallacy. Some JVMs
> can
> % also run
> % faster than natively compiled and
> % optimized
> % code (The common example being
> IBMs
> % JVM vs
> % VisualC++ compiler). At one stage
> Java
> % was too
> % slow for many things but now this
> is
> % not really the
> % case.
>
> I think we are all aware of the
> virtues other JVMs offer. Again right
> off the discussion topic..

Read the responses. There is discussion about this
in previous comments...

> This wasnt the discussion. The hub was
> if Sun (and friends) should be pushing
> it hard on Linux as a competitive answer
> to .net. Id say definately yes,
> 'Enterprise' (nice open interpretation
> buzzword) or not. Whinning and paranoia
> about Sun because there a big company
> aside, the two complement themselves
> very well in a cost effective
> foundation. The problem IMHO is Sun are,
> in a way, shooting themselves in the
> foot over Solaris.

It is not really abour the OS but the hardware. Sun
makes it money from hardware. Trends in linux point
to many light weight boxes to replace the one huge
computer. When the management of clusters
becomes cheaper it will be more cost effective and
reliable to use a farm rather than single computer -
no matter how many CPUs. Pushing Linux could
potentially be very bad for buisness due to the low
cost of x86 based clusters (when they mature).

I'll make it really simple
Give me an IDE equivalent to VB, be it Java or Python and I'll make the switch. The only reason I've ever VB'd was due to the ability to draw my forms and embed events.

The particular semantics of a language are of little concern for me, programming is programming and I can get it done regarless of the syntax as long as my tools are sharp.

Re: I'll make it really simple
For java there's FORTE, JBuilder, NetBeans and a whole bunch of others, check out balckdown.

For python there's boa-constructor :)

Re: Enterprise?
I definitely agree. And I just laughed when I read that point about CGI / PHP applications just not being good enough for Enterprise.

I would never touch Java or JSP for web development. Sun likes to make us think that Java is the do-it-all language, however I feel it's far from it I'm pretty glad this push for pitting JSP against PHP isn't going too far because JSP just isn't the right tool. Someone put it a good way in saying that if the job were to hammer in small nails, PHP is a small hammer and JSP is a sledgehammer with an attachment so it can hit small nails.

Re: cold coffee anybody?

%
[... snip ...]

> quite simply it take more lines of

> code to do simple things in java...

> simple one liners in perl take a page in

> java with imports... regex anybody? $t

> =~ s/^k.*(\d+)\s+(\d+).*(\d+).*$/$3/s;

> print $2.$1;

I tend to not involve in language discussions because it's usually "religious". But this is just a bit too much bullsh.. !

Eric S. Raymond, as undisputed an authority as can you can have and an EX Perl institution, states that he doesn't use Perl anymore for anything with more than a couple lines.

Why ? Because Perl code is very HARD to maintain and very errorprone.

It's also the COMPLEXITY of the lines that counts.

And, of course, Mr. Perl shows off using a regex; as if only Perl had regex and as if regular expressions were the solution to every problem.

Listening to Perl people I see how true that old chinese saying is:

- A simple-minded man doesn't understand the world and knows it.

- A wise man knows that the world is complex and tries to understand it.

- A stupid man doesn't know about the world but thinks he knows it.

> the more lines of code there are the

> higher the chances of a mistake... and

> the harder it is to prove a program is

> correct. Java is just as obscure as the

> next language... oo with lots of files

> just makes things harder to keep a track

> of in your head... C++ is worse in this

> regard... wonderful for world class

> experts, but in the hands of normal

> people it just doesn't cut it...

OO is probably the single most important thing in software development, in particular for non-experts. And experts tried hard to use their mastery of "C" in order to design as object-oriented as possible. Do you need OO in scripting ? Sometimes, maybe often, not. But then, scripting isn't about developing major applications.

Scripting is, in summary, about creating a solution ("quick and dirty") for a problem that isn't worth (or can't afford) a long full-blown development. The solution must not be nice, neither fast; its just needs to get the job done.

Software development, somewhat to the contrary, is about developing a powerful, fast, elegant (add your attributes ...) that is worth and can afford a full-blown deveopment cycle.

The fact that some people use a scripting language, in

particular Perl, for big projects typically has more to do with not mastering other, more appropriate, languages than with Perl being adequate for major software development.

In fact, more and companies have a no-Perl policy for anything but small scripts.

> i've been at this game for 20 years

> and i've seen more "wonderful"

> languages than i care to remember...

> java just doesn't do it... perl ain't it

> either... but its a hell of a lot

> closer! oh... and you don't have

> licencing issues...

Well, I didn't see that many "wonderful" languages. In fact, the only languages that claim quite openly and often to be "THE" language are Java and Perl.

Some other comments (on statements made above:

- Yes, Sun seems to keep Java quite "public" and free. But: If they please, they close the door tomorrow morning.

- No, Java apps isn't fun to install. Often enough there are competing JVMs piling up on my disk and often enough they are downloading more needed classes .

Is there THE language for web-development ? I doubt it.

I personally use - potentially C(++) enhanced - Python. But I wouldn't dare to put that as a bold "exclusive truth" statement.

My Tool is better than your Tool
My Tool is better than your Tool and that makes me a better person.

I can also retort to everything you say - and that means I'm not an insecure bigot who has invested too much of my self worth in some specific little artificial corner of existence.

Geeze..

Re: He's Right.

>
> Please I beg you... show me one
> application
> written in Java/J2EE that scales to
> the extent
> it could support a national bank...

OpenAdaptor (http://www.openadaptor.org/), written by a bank.

Andy

Re: beyond internet application serving

>
> It's a python interpreter that runs on
> my Palm m100...
>

Your Palm m100 is 10 to 20 times better than most cell phones J2ME is riding on - assuming memory and CPU power as two axes producing computer power as an area in a graph.

Re: I'll make it really simple

> Give me an IDE equivalent to VB, be it
> Java or Python and I'll make the switch.
> The only reason I've ever VB'd was due
> to the ability to draw my forms and
> embed events.
>
> The particular semantics of a language
> are of little concern for me,
> programming is programming and I can get
> it done regarless of the syntax as long
> as my tools are sharp.

Begining the design of an app from its GUI is not the best way to reach a solid and flexible OO design. It forces you to use a real-world based analysis. It may seem good at first to do it, but OO designs that only reflect real-world objects usually lack abstraction and generalization, and easily either become inflexible, or grow indefinitely to meet all variations of the original problem - thus becoming hard to maintain.

If you've developed and maintained your own applications for over one year, you know what I'm talking about.

Most nonsense being said here is from people who never wrote a large app or maintained legacy s/w. Always rewriting from scratch is not viable, so software must be written for maintenance.

Anyhow, I'd discourage anyone from using VB to anything but small apps. It's hell to create good designs and maintainable apps.

Cool to see the evolution of this thread.
It all starts, as usually, with some quick to speak guys who piss off others a lot. So much that usually quiet and, I will say, usually much smarter guys have to step in and try to reason... Like alvays. Like the life is. Have a beer! I will...

Denis.

java performance - superstition vs. real numbers
People who make comments about java's performance generally reveal in the process that they've never used java, or at least not in the last few years.

Benchmarks show java as being in the same ballpark as C across multiple types of computation (integer, floating point, etc.). Perl is fast at only one thing: strings -- try doing floating point in perl (really, try it).

See e.g. the java-vs-C benchmark page at www.idiom.com/~zilla/Computer/javaCbenchmark.html,
or search on Tom's hardware for a similar benchmark.

Re: java performance - superstition vs. real numbers

> People who make comments about java's
> performance generally reveal in the
> process that they've never used java, or
> at least not in the last few years.
>
> Benchmarks show java as being in the
> same ballpark as C across multiple types
> of computation (integer, floating point,
> etc.). Perl is fast at only one thing:
> strings -- try doing floating point in
> perl (really, try it).
>
> See e.g. the java-vs-C benchmark page
> at
> www.idiom.com/~zilla/Computer/javaCbenchmark.html,
> or search on Tom's hardware for a
> similar benchmark.
>
>

I belive that your statement is somewhat sweeping. I have programmed in Java in the last four months, using a new compiler. Not only did the compiler take longer than any C/C++ compiler I've used, but to RUN the Java programs took 5 times longer. It didn't seem to matter what kind of task is was either. Number crunching, strings, GUIs or internet connections were all much slower than in C/C++.

I think Java has a long way to go and I would perfere it gave up before it gets there.

Re: java performance - superstition vs. real numbers
% I belive that your statement is
> somewhat sweeping. I have programmed in
> Java in the last four months, using a
> new compiler. Not only did the compiler
> take longer than any C/C++ compiler I've
> used, but to RUN the Java programs took
> 5 times longer. It didn't seem to matter
> what kind of task is was either. Number
> crunching, strings, GUIs or internet
> connections were all much slower than in
> C/C++.
>
> I think Java has a long way to go and
> I would perfere it gave up before it
> gets there.

This is what may be called sweeping statement, IMO. "To RUN Java programs"... does that include the overhead of starting JVM, too? Anyway, what's the code that runs 5 times slower in Java? If it's not too embarrassing to show it... As of compilation time, I guess, as a professional you are using jikes or fastjavac, no?

Denis.

Re: Enterprise?
>Funny. I distinctly remember using remote method
>calls, load balancing, fail-over and transactions
>without java.

Funnier thou, it makes us two :-))

In early 1990s there was no java to write "Enterprise" software, so we used VAX/VMS Pascal instead (-;
funnier still: this code is still in operational use!

Java is just a badly fixed [pronounced: broken] C++. It brings nothing new and takes away a lot.
Marc says "Java is not an easy language to learn for beginners." But:
(* /;-)
beginners should learn pascal: to find out about nested procedures and functions.
No such in pointerless interpreted assemblers like java.
And java looks like an insult to a C++ programmer.
(-;/ *)

On the other hand Perl, PHP, Python are interesting to learn and bring lots of new and powerful
concepts, which also makes them fun! All of them are available on ANY modern unix and MS stuff. Using them gives
you power to choose the right tool to do the job. Using java sledgehammer rarely unscrews ...

I do not want to state java is not usable. It might be. If you want/must run in a browser, if you must use java, ... then java is THE answer.
If you are concerned about performance, security, portability, rapid development, then have a look around.

Instead of smelly hot java, have a cup of Lapsang Souchong, a glass of wine,
a bottle of beer, a can of juice, a shot of schnapps or a hand rolled cigar for that matter...

I do not believe in Sun's benevolence. All they are after is control and/or bigger share of market - not diversity, for that brings choice!
And people want choice, for it enables their free will.

And choice is what life is all about.

As a fine conclusion I'll quote PhiRatE:
>Don't propogate the myth. Linux only needs Java
>because PHBs won't take any notice of anything
>else. It is not a technical need, its a marketting
>need, nothing more.

Re: He's Right.

> Your full of SH*T.

Aren't we all?

> Doesn't scale? No
> one uses it? I don't know if they still
> publish it, but Devx.com's enterprise

Why don't they publish what they are using
it for or let BEA or Sun or IBM publish how
great it is? Maybe they do in the US, and
it just doesn't make it up to Canada....

> bank). AND THIS WAS A YEAR AGO!!!!!!

Must be some of the first fully J2EE compliant
stuff ever done.

MS is ditch Java off of XP... that's cute....
and it's gonna be painful for Java on the client
side...

-Ironstorm

Re: Cool to see the evolution of this thread.
Amen brother!

Re: Enterprise?

> Java is just a badly fixed
> [pronounced: broken] C++. It brings
> nothing new and takes away a lot.
> Marc says "Java is not an easy
> language to learn for beginners." But:
> (* /;-)
> beginners should learn pascal: to find
> out about nested procedures and
> functions.
> No such in pointerless interpreted
> assemblers like java.
> And java looks like an insult to a C++
> programmer.
> (-;/ *)

<br><br>
i'd have to disagree. sure java takes away some of the functionality of C++ like pointer arithmatic and whatnot, but that's a good thing for the beginner (because they're less likely to mess something up). i think for learning object oriented programming, java is a great language because it forces you to think object oriented and not procedural. once you get in this habbit, then you can move onto C++.... and even if java is an "insult" to a c++ programmer, your average "beginner" isn't going to be a hardcore C++ programmer.
<br><br>
as far as speed issues go... if you use a JIT, its no longer interpreted, correct?

Re: Enterprise?
Well, I did not want to *insult* anyone; that's why I enclosed the text in Pascal-style plus wink-n-smile-in/out comments.
What insults *ME* is that (lots, not all!) java programmers think and speak of java as THE language. Java brings nothing *really* new. That does not mean is no good and does NOT mean I do (or will) not use it.
I AGREE with you that the point is to THINK in OO way in order to get 'this habbit'. And getting habbit under the skin is important. Not the language of choice! The point is to learn to program. Not to make religious statements on programming languages: All and of them rule and all of them suck at the same time: depending on how good you are in (the art of) computer programming.

Let's not forget that!

NEED A GOOD PROJECT FOR ENGINEERING
Somebody please suggest me a good topic for project related to Networking in linux,.. or anything in linux..

...
I frankly cannot see the need for all this flaming and ranting... Linux needs to support all the current technologies if it is to succeed in the "Enterprise" league... Java may or may not be the best of languages out there (i don't consider myself an expert to delegate such matters) BUT right now it is generating a lot of hype and Linux is there with it...

True, Sun is "flirting" with it's Linux crowd by allowing bits and pieces of its goodies to be supprted in our favourite OS but then again it chooses to neglect us when it sees fit (JavaComm API anyone?)...
Nevertheless, it should be no cause for a civil war between Linux developers...

My view of the whole Linux experience so far has been that I (as a developer) have the option of using any tool available (be that Java, Perl, PHP, et al.) so I believe it is important that Java (not unlike the rest) is supported in Linux

However, I don't agree with the writers view that Linux NEEDS Java... as I said, Linux needs developers with open minds... Arguing about one's toy being better the others doesn't help things...

...a tool cannot be good or bad, it's what you do with it that counts

Java r0x, but not now
hm... time for my narrow minded thoughts. I'm not a developer of anything in "Enterprise" dimensions. My experiences with java are not the best - especially when it comes to installing this stuff. compared to installing apache with php mysql and postgresql on linux it's a pain to do tomcat, apache, java-sdk, mysql, mysql jdbc driver,... also on linux.

community - the Perl, php, mysql, postgres community is far better (maybe not bigger) than the java community is. with better i mean there's a lot more free sample code out there, plus there's a lot more people helping in different newsgroups or web-boards. may it be the buzz-word java or my incompetence, but there's not that many pages out there having "real life" source code for java applets or servlets. There mostly precompiled classes and a small documentation. Well for the newsgroup part java is bigger but after a closer look there's not much more substance in it than in a PHP group. I think having a huge supporting community is very important when trying developing i.e. medium sized webpages at reasonable costs.

speed of java - hm... on my testing machine JSPages are generated far slower than phpages, but again no Java or PHP hero talking to you here - maybe my fault... but try this one - program a audio tool comparable to cubase VST32 or Samplitude 2496. hehe - lacking in audio device support? Java not fast enough? hehe - i know this wont work for a long time with java, especially for professional audio.
So java is not the perfect tool for ALL stuff out there. Like every programming language it will find it's place somewhere in a special sector. not likely in the professional CAD or Audio world but probably in many business-sectors.

Java language features - well java is crippled C++. But the language is still in development - as it had no support for Operator Overloading in the beginning, it has that feature now! And rethink how long do c++ compilers exist that fully support templates? is it 3 years or 4 ?? The language was developed quite some time before any "real" c++ compilers came out... same goes for java, the big idea exists but java's too young to tell now if it will succeed or fail.

Fazit - using java is not comfortable for normal users - installing is a pain, it's too complex for programming beginners, its compiler makes former c++ programmers wanna smash the screen ;) plus creating intelligent installers for a product concerning all OS out there is quite hard. Java may be a cross-platform but is not a "cross-application type" programming language. with that in mind it is a useful alternative to develop software without paying a cent for the compiler package. same goes for linux and many other programming languages. Therefore i try to stay open to many alternatives. Things in IT world change very fast - COBOL almost died now and so will C++, JAVA, PERL and all others sometime be banned. the problem is that "sometime" will happen within our life as programmers, so keep your eyes and ears open.

my 2 cents

HTD

Re: A Personal Note

> I'm sorry for the lack of articles on
> freshmeat lately. I just bought
> a house, and was busy for several
> weeks, then was off the net for ten
> days. I'm back, and the article
> machine is cranked up again.

Well, that was a nice theory, but then I went out to work in the yard
and got into some poison ivy, and had blisters on my fingers so bad
that I couldn't type for a week. Let's try it again...

More dribble.
This is more FUD. I use the right tool for the right job. The fact is that much of the time Linux is the right tool, or can be in the hands of someone who can use it. Another fact is that the number of people who can use Linux effectively is far less than those who can use windows effectively.

This article mentions that Linux costs less than windows. But the TCO of a windows box is usually far less than a Linux box. Backups are easy(you can pay a MCSE $25/yr to run Backup Exec). Where on Linux everything is usually more complicated.

Not to mention that Sun's business practices are more questionable than Microsoft's. The only reason they aren't attacked like Microsoft is a matter of market cap.

Think twice.

Re: Enterprise?
I work full time writing Enterprise Applications for the likes of banks, financial institutions, and other regular businesses. Guess what two choices for writing those apps are 98% of the time? -- Java or ASP/Dcom/MSgarbage.

Why? Simple. While you are right on the mark with the benefits of perl/python/eiffel/etc, none of those have developed platforms that support:

1 - real transactional behavior -- Does your language of choice support cross DB transactions (meaning, I need a transaction that spans the legacy mainframe RDBMS and the new Oracle DB just installed for new project X), transparently, and without the need to code begins, commits, and rollbacks?

2 - clustering -- Can your language of choice seemlessly execute across 5 different servers, with objects floating between according the app server's balancing scheme, the whole thing remaining fully functional even if you kick the power out of 4 of those machines?

3 - security abstraction - can you define security/authorization roles at deploy time, without relying on a programmer who "thinks" he understands security to develop some secure code for you?

4 - resource pooling -- does the perl interpreter automatically figure out how to efficiently allocate resources such as DB connections without your programmers needing to write some half-witted pooling code themselves?

Granted, you could build all of these features into an app server around ANY language. BUT IT HASN'T BEEN DONE YET, except for EJBs and Microsoft's offering. Sorry. Your languages just don't offer these features yet. As such, they ARE NOT FIT FOR THE ENTERPRISE. Do you think a 45 Billion in asset bank is going to trust a team of 4 programmers to write code that handles million dollar financial transactions without these features?

And sure, "good programmers can write all this stuff themselves." They can code transactions. They can code security. Yes. So? All that stuff adds to development time, is less trusted by management, and therefore costs exponentially more.

Linux needs Java for the Enterprise. Unless you want .NET.

There are no other alternatives currently. Of course, I'd love to see some nice Perl or Python alternatives arise. But they ain't there yet...

Java's way of preventing bugs (and features)
Java tries to prevent bugs, but I think it does it poorly...

instead of preventing bugs by good design, making it easier to write good code,
Java prevents bugs by removing features, making it harder to write bad code (and good code if you know what you're doing)

in essence, Java prevents bugs from unexperienced programmers, who shouldn't have anything to do in a production environment ;-), but prevents smart optimisations from experienced ones...

These "removed" (now i'm thinking of Java as a development of C++) features (pointers, templates etc.) are features that give C++ its great power and flexibility.

So, Java is good at preventing bad programmers from making bugs, but thus also prevents smart programmers from making smart solutions...

These are my opinions.. Complain if you think i'm wrong...

// Olsner

Re: Enterprise?

> Well, I did not want to *insult* anyone;
> that's why I enclosed the text in
> Pascal-style plus wink-n-smile-in/out
> comments.

Oh, i thought they were little guru-with-turban
emoticons :)

Re: Java's way of preventing extra costs
I agree.

***Think "Extensibility, Maintainability, Stability" when choosing programming language***

Java is simplified version of C++, and that pissed me off when I started coding in Java, but it has some advantages. First, it is EASIER (read- CHEAPER) to code in Java than in C++ (I don't know about perl or PHP, I don't think anyone has done any studies on the subject). Managers used to say that it takes 1-2 years to become a Java programmer, and it takes 3-6 years of practice to code in C++ well. And so code written in Java is 2-3 times cheaper than that written in C++. Also, Java code is much more easy to read than perl or C++, that means it is good for large projects that must be maintained a long period of time.

Second- speed. Java might be slower than C++ (maybe slower than PHP or perl in some cases), but usually the slow part is the JVM startup. Also- Swing is quite slow and uses a lot of memory. But making web applications with Java where it does JDBC and some String crunching is fast enough for most situations. But the biggest problem is not speed anymore- it is about how fast you can CODE it and.

Third- OOP & OOD. The reason I dumped perl and PHP (I use it for some quick & dirty scripts, but not more) is because of poor support for object oriented programming. Try implementing a few Design Patterns and you will see what I mean. And I have already given up writing bigger projects without OOD. Java is BASED on objects, and it encourages strong OO Design even if you sometimes don't want it.

Fourth- class libraries. Java has strong API. Well, perl has good support here too. C++ is different matter. Most of libraries are C, and quite a lot of C++ libraries are broken or offer not sufficient features. Java libraries are usually enough to do the job. And writing web applications in C++ ://// Tell me if you know any frameworks for this.

Well, I think that should be enough.

Re: Java's way of preventing extra costs
Couldn't agree more!

I have also come to the conclusion that development time and maintainability is often more important than running performance.

Development speed and maintainability are Java's great strengths. This (I think) outweights Java being so slow.

I would like to know what you think:
Is Java so much better developing in than it is slower?

Also, do you have any experience with how gcc-java executables performs compared to running in a JVM?

Re: Java's way of preventing extra costs
% Also, do you have any experience with
> how gcc-java executables performs
> compared to running in a JVM?

I don't have any experience with gcc-java. I planned to test it as soon as I find time, now when gcc-3.0 was added to debian/unstable.

So far speed of java language was good enough for me, so I wasn't looking for improving JVM. Usually some optimizations in application improve speed more. I am writing web applications, and the biggest speed problem here is Apache-Tomcat. F.e. resin is much faster, but it costs money...

Re: Amusement

> There also potential sucess stories
> brewing with
> the smaller end of spectrum

Right... That's exactly what you all have been saying for 6 years. I'm still waiting.