OSNews

You mean 1.6, ne?

Actually no. As of "Tiger" Sun switched from the 1.x system to the x system. Hence we've had Java 1, Java 1.1, Java 1.2, Java 1.3, Java 1.4, Java 5 and will soon have Java 6.

They've adopted the same scheme for J2EE as well, the next release is going to be called J2EE 5.

Apparently spelling isn't a valid reason for modding your comment down, so I won't do that. Notice how I wrote "your" comment.

Just to make this more clear, observe the following example:
"Nobody is going to take your comments seriously if you're too lazy to learn elementary school level spelling."

Now onto the meat of my post:
Calling Perl, Python and PHP upstarts is ridiculous.

Apparently spelling isn't a valid reason for modding your comment down

Yeah, strange that. I always thought the one thing OSNews was lacking was a solid community of spelling and grammar Nazis.

For example, I'm incensed that I can't mod you down for following a colon with a capital letter. Just to make this more clear, observe the following example:

"The meat of my post is this: no-one is going to take your spelling rants seriously if you can't even manage basic grammar".

Sheesh....

There's always a "product/company x is dying" article out there, it's the kind of fiction that attracts readers who haven't yet gotten tired of that cliche headline. For some people the headline still induces some form of panic, denial and shock (or hope and glee) until they've seen almost identical headlines come up a few more hundred times.

A very miniscule number of these articles are sometimes correct, other's might present plausible arguments but forget that companies are run by people with brains (whatever the size and capacity) that are capable of changing the company in such a way as to make it profitable again.

Excatly, corporate IT has lots of fat to burn. They can use the hardest tools.

Books sales are UP becaue nobody was buying them before?

Ruby book sales are up 1500% and have surpassed Python sales...

You have got to be kidding... citing book sales is now a measure of a technologies success? WTF??

Citing? The GP Cited no reference to this proclaimed 1500% increase. BTW 80% of statistics are made up on the spot.

Right! I have to concur. If 'sales' had anything to do with a technologies success, then Windows would be a truly good product.

I believe book sales are up because of the whole recent "Ruby trend", and every trendwhore and his brother/sister is wanting a piece of the 'Ruby brothel'; that's why the surge in book sales.

--Jed

> Ruby on Rails is the one gaining the most
> momentum... Ruby book sales are up 1500% and
> have surpassed Python sales...

I don't buy either claim.

Ruby on Rails: All smoke and no fire. Show me one good case study of someone who has build a complex three-tier web app in RoR? I have yet to find one.

Ruby book sales: I don't buy this one either. Please provide a source for that claim... There are like what, 3 books on the market on Ruby?

http://radar.oreilly.com/archives/2005/12/ruby_book_sales_surpass_p...

I just said Ruby is gaining the most momentum. Java is probably 1000x more widely adopted. Ruby is clearly getting a ton of attention and is growing rapidly though.

You can definitely write code more quickly and more tersely in Ruby than Java. And I think terseness matters. Why write 10 lines of code when the idea can get conveyed clearly in 2? Terseness means easier maintenance and it's easier to follow/read code as you don't fill your screen with a bunch of worthless BS like getters/setters.

Also, the APIs in Java are very overarchitected. Try sending an HTML email using the Java Mail API. Struts and Hibernate are complicated diasters. Look at Rails and Active Record for the beautiful Ruby equivalents. No XML files!

All this being said, I personally wouldn't use Ruby for anything other than a web app. *Yet* It's biggest weakness IMO is no support for native threads (green threads only).

Ruby may be good for new webapps, but it just wasn't designed for legacy databases and it lacks features such as support for multiple primary keys and database table remappings. If you mention this to Rails. developers, they'll mostly agree with you, but state that that's not the focus of Rails (at least not yet). The focus of Rails is "green database" apps and it wants to be the best out there for these apps.


Java isn't the best choice for "green database" apps, but it is great for legacy databases which is where it shines.

> Java isn't the best choice for "green database"
> apps, but it is great for legacy databases which
> is where it shines.

I disagree. Java is great for new Webapps. And virtually no one works directly with SQL anymore in Java. It has some of the most powerful object persistance frameworks available. For example, Hibernate.

When it comes to any kind of database application, Java, with it's powerful object persistance frameworks, is far more powerful than Ruby.

wait for 2.0 then.
one thing that concerns me is the performance speed of ruby (or ruby on rail).

it doesn't seem that fast?

> one thing that concerns me is the performance speed
> of ruby (or ruby on rail).

Yep. It's pretty slow. In fact, it is the slowest of the so called scripting languages. Because it is purely interpreted.

There is some work being done on a bytecode compiler for Ruby. But I have no idea when it will be suitable for production use.

May be slow, but premature optimization, like choosing a language that's fast, is for young people, who still have lots of time for mistakes in their lives.

"May be slow, but premature optimization, like choosing a language that's fast, is for young people, who still have lots of time for mistakes in their lives."

Sorry. You can't apply the premature optimization argument here. It doesn't work whem comparing a language like Ruby (which has many other drawbacks besides speed that make it unsuitable for large and complex web apps) and Java.

Besides calling you on your performance disregard for Ruby, maybe you want static typing? Good. Now you have a fast language with static typing. That's enterprise enough, indeed. That's Java true calling for existance, because it loses in programability, though it has EJBs which rarely are used, just like Enterprise Messaging. Are you using WebLogic already? Maybe IBM's or Sun's Enterprise Servers? You rock.

I don't want such Enterprise crap which you don't even use most of the time. You use basic ORM with Tomcat, probably. Maybe JBoss with its shortcuts. Or JoNAS. Right?

That's "Enterprise". That's why Bruce Tate who wrote 3 books on Java chose Ruby for his startup. That's why the author of Ant chose Ruby for his startup. Those guys don't need your Enterprise fluffy.

"Besides calling you on your performance disregard for Ruby, maybe you want static typing? Good. Now you have a fast language with static typing."

Static typing is important for three reasons:

- It catches assignment problems that might otherwise go uncaught and cause subtle logic errors to creep into applications at runtime.

- It makes it clear in a project involving multiple developers, exactly what a method is going to return, or what it accepts as a parameter.

- It guarantees type safety, and guarantees that casts are safe, which is important for reliability in a mission critical app. It ensures an illegal cast cannot be made on an object for example.

This has nothing to do with premature optimization. But it has everything to do with making code self documenting and maintainable. Static typing is important for large and / or multideveloper projects.

"though it has EJBs which rarely are used"

EJB is regarded as overly complex by many developers. And there are better solutions. Like Spring, which is what I typically use. Spring's IOC and dependancy injection make many things a lot easier. And Spring 2 greatly reduces the amount of configuration that must be done.

"I don't want such Enterprise crap which you don't even use most of the time. You use basic ORM with Tomcat, probably. Maybe JBoss with its shortcuts. Or JoNAS. Right?"

No. typically I use Resin combined with Spring. Sometimes I use Struts with Spring, but more recently, have been going all Spring.

And actually, I do use a lot of that "Enterprise crap". For example, I use Spring's validation framework. And I use Webflow. And that so called "Enterprise crap" actually could make many people's lives a lot easier, even outside of the enterprise, if they would just take the time to learn it. How many times have you have written a bunch a boilerplate code to handle session tracking and such between web apps that require multiple pages, and multiple forms? And to validate data and prevent the user from moving to the next page if it is invalid, while at the same time not screwing up the overall flow? That's the kind of thing that so called "enterprise crap" like Webflow takes care of, and reduces a lot of the boilerplate code that needs to be written.

"That's why Bruce Tate who wrote 3 books on Java chose Ruby for his startup."

Bruce Tate changes his opinion about languages faster than the wind changes direction. First he liked Java, than was touting the praises of Python (which was not that long ago) and now it is Ruby. He drifts in whatever direction the most popular FAD currently is. But long after Ruby has failed to gain any kind of marketshare, and still remains a niche player that has failed to break into the Web app market, Java will still be there. And still be going strong. And then Bruce Tate will probably follow the wind direction back to singing the praises of Java again.

Edited 2005-12-15 01:23

Um, bullshit.

Rails works fine with legacy databases. It just doesn't provide automated model generation, you'll have to fill in some details to make it work.

It's surprising how many misconceptions there are about Rails ....

> It's surprising how many misconceptions there are
> about Rails ....

Not really, given that Rails people have decided that the best way to promote Rails is to make outragious claims for it, throw up a lot of smoke, and make FUD claims about all other Web app languages to try to trash them and make those considering Rails believe it is the only real option they have since all other solutions are seriously flawed. It's no wonder most of the Rails hype and claims are taken with huge lumps of salt.

Oh please.
Every technology gets a bit overhyped. Rails is no different.

I'd like to see some examples of this FUD and smoke that you're moaning about.

> I'd like to see some examples of this FUD and
> smoke that you're moaning about.

Go read the mailing lists and such for Rails.

I have several Java books. I have a couple of Python books. I don't have any Ruby books.

I'm not likely to buy any more Java books since I've got pretty much what I need and I'm comfortable with it. Ditto Python. I might buy a Ruby book or two to learn that language to.

The point is, however, people don't buy the same book over and over again, and a sale indicates interest, not use (I've never actually used Python for anything... yet). Looking at book sales figures is probably a worthless measure (particularly of Java popularity -- the market is saturated and mature there).

And WebObjects also went Java a few years ago.

Seriously, Perl, Python, ruby and etc are all great and have their place. But the most important thing is to steer clear of ".net", because, "mono" notwithstanding, MSFT's offerings are not cross-platform, and, thus, can only be deployed internally in closed, all-Windows. shops.

"and, thus, can only be deployed internally in closed, all-Windows. shops."

You've apparently never worked in an enterprise level shop then. Windows can talk to non-Windows machines quite easily these days, and any application can communicate w/ other apps via a number of different mechanisms. There is no reason whatsoever to steer clear of .Net; almost all shops have a blend of MS and non-MS machines.

> There is no reason whatsoever to steer clear of .Net;

Sure there is. you lock yourself into deploying on the Windows platform. And don't tell me mono. Because complex three-tier Web apps written for .NET likely aren't going to run in mono. Not nearly all of the .NET API has been ported to mono (and most likely, some of it never will be).

With Java, on the other hand, it is *extremely* common to develop and test on Windows, but deploy on Linux for example.

The applications themselves may be confined to Windows, but if designed properly can very easily communicate with applications on other platforms. I won't even begin to get into Mono, b/c as a serious .Net developer I think it's rubbish...they are basically attempting to reinvent the wheel with stuff that's already tried and true (and very mature) on *nix, which can be done with those tools, and again (if designed properly) can communicate with Windows/.Net applications down the line if need be.

The point is, why limit yourself to one platform? There are plenty of things that Windows (and .Net) do well, and vice versa for other platforms. I've said it several times in this thread...any decent size IT shop will always have a mixture of languages/platforms/OS's...they all have their niches, and can all play nice together if need be. There's no need to limit yourself to one platform anymore. Closed or not, MS has a huge presence in IT...and going in with that sort of attitude (that MS is closed or whatever) isn't a good mindset to have. In the end, the stakeholders in the project don't care what platform it was developed on/deployed to...as long as it's done within budget, and most of all, "just works" when all is said and done.

Edited 2005-12-14 06:36

Well, some of us have probably missed the big news of Javas death. They are just trying to make us see the light, which is C# or Ruby or Python or Perl or something.

The fact that Java applications exist and are being developed even as I write this, is a fact that must not disturb those who know better.

I wonder if I'm heathen for having Eclipse on my machine?

"dying since the 90's" is a good thing; look at what it did to Apple, they've been dying more every day!

Since you already are voted up to 5 points, and I think they ought to change the scale for such a brilliant comment, I just wanna say, I'm happy to see you can read between the lines and analyze the situation.

Java isn't going anywhere soon... it's just that this journalist tries to compare enterprise applications which are mission critical, with some obscure community site which can be taken offline for maintenance once a week...

It's an article from BusinessWeek. What else can you expect from them ?

For all these facts, could you provide some supporting documentation? Even if only for the ones that I call BS on :-) :

"Fact: Java is still the most in demand programming skill in the developer job market."

"Fact: J2ME is now installed on over 2 billion mobile devices. That's right. 2 billion."

Etc...

"For all these facts, could you provide some supporting documentation? Even if only for the ones that I call BS on :-)"

Certainly.

Java and job demand:

As of today, these are what the dice.com numbers look like as far as jobs wanting various skills in order from most in demand to least in demand.

Java: 12181
.NET: 10091
C: 9022
C++: 7113
Perl: 4035
PHP: 786
Python: 509
Ruby: 72

Java is the clear leader. And Ruby, the overhyped Java killer, only has a whopping 72 job postings. PHP, which some also like to proclaim will be the Java killer, has only 786.

J2ME and 2 billion mobile devices:

This was a little off. Java is in about 2.5 billion embedded devices. About 708 million of those are mobile phones. Here is an interesting thread on the topic:

http://answers.google.com/answers/threadview?id=571139

On linux, you don't have a proper Java implementation.

Are you assuming that you like Windows just like most Java developers? :-)

I repeat: I don't care. I only care when you want to say that Java is good. It's not. Java is bad for server-side and client-side. But as long aas you are getting your paycheck, you won't have a problem with it. Java employs maybe 50% of the corporate programmers. It's a pain to watch RedHat adding support for Java. But I understand it. Though I would prefer that Havoc Pennington were working on Gnome and Metacity instead of on Java server technologies for RedHat.

Are you sure? Last time I checked, at least Sun, IBM and BEA have their Java implementations running on Linux at the very same level than their Windows versions...

Amazingly, that's almost completely the opposite of the truth. All the Java development I have ever done, whether small, medium or enterprise scale, and whether corporate or personal, has ended up running on a unix box.

My pet peeve is with people using Windows for Java development. For example, when they prefer to use Eclipse on Windows rather than on Linux.

My pet peeve is with people using Windows for Java development. For example, when they prefer to use Eclipse on Windows rather than on Linux.

So? What's wrong with that?

On Linux you have the best tools. On Windows, you have the Java developers

On Windows you have just as good tools -- in fact, you have the same tools. I don't see any reason to drag operating systems into this holy war.

Amazingly, that's almost completely the opposite of the truth. All the Java development I have ever done, whether small, medium or enterprise scale, and whether corporate or personal, has ended up running on a unix box.

Agreed. I think people interested in Windows and not necessarily other platforms are going to be more into .NET than Java.

Let's ask then.

How many of you Java programmers like to use Windows? :-) Nobody wanting to raise the hand? :-) Come on, it's not such a shame.

Actually, most of the Java developers I know do all of their development on Windows machines, regardless of what platform the binaries will ultimately be deployed to...once they finish their work, it's up to the operations/deployment teams to figure the rest out (as far as ironing out the plaftorm specific kinks). At least that's how it's been on the larger projects I've worked on. In reality, most Java guys have nothing against Windows or .Net...they chose to be Java devs, just like I chose .Net. As it relates to corporate IT, the platforms are more complimentary than supplementary, and work great together if done right.

Thanks for sharing it. Now I can relax. I think you are right in using DotNet, BTW. It's a tool to respect.

That is crap. I develop java on my mac, which has a proper java implementation.

AJAX still needs software on the server to work :-)

But that is not "Enterprise" stuff. That's academic stuff.

ORDBMS are great when the life-time of you data is about the same as the life-time of your webapp and that one or two webapps are the key users of the database.

ORDBMS break down when the life-time of your data vastly outlives the life-time of your webapp or when you have several webapps accessing the data with completely different visions on what the object model is (i.e. enterprise software).

This is where RDBMS shine. If you partition your data into 4th or 5th normal form, you can slice and dice your data in every useful combination that each webapp needs and if you have a good RDBMS and properly designed queries you'll have great performance because you'll only have to send back exactly the information you want. If you're having latency issues, then I suggest that you look at your queries and your indecies and see if you're doing something wrong (like doing an O(N^2) operation when an O(N) operation will do).

Please don't think that because you have a good hammer that all your problems are nails.

I do not think the relational model is ever going to be a relic, sql is a relic but the relational model is very usefull. I ask you what is so intresting about xml? the relational model has some intesting theoretical results. and relations and sets are also very fundemental to math. for xml or objects that is clearly not the case.

sorry to reply to my self.
but i think i have a nice analogie . relational model is is to a database guy as lineare algebra is to a graphics person or formal languages are to compiler writers

"All those PHPs, ASP.NETs MySQLs etc. are based on relic relational data model from 70s."

PHP and ASP.NET has nothing to do with the relation data model. They are *programming* languages that can use pretty much any data model.

Please don't confuse the 'relational model' with 'current database products that use SQL'. While the ignorant or unintelligent may agree with you, those with knowledge will correctly deduce that you are strictly bush league.

"or let companies port Java to the "little" platforms (that are currently unsupported by Java) at no price."

They already do. This is how Java on FreeBSD works. Sun does not make an official FreeBSD port of Java. But they do allow the FreeBSD team to create the port themselves.

"or let companies port Java to the "little" platforms (that are currently unsupported by Java) at no price."

They already do. This is how Java on FreeBSD works. Sun does not make an official FreeBSD port of Java. But they do allow the FreeBSD team to create the port themselves.

Not at no price[/i]. FreeBSD has to pay through the nose for permission and Sun takes excessively long times to certify it, which means that FreeBSD's Java binaries lag far behind mainstream. I think they're still at 1.3.1. Sure, FreeBSD's patches allow you to compile Java 1.5/5.0 yourself, but neither they nor you can legally distribute the generated binaries.