A First Look At Red Hat Developer Studio

juanignaciosl writes "The first beta of Red Hat Developer Studio was published yesterday. RHDS seems promising. This IDE is a bunch of Eclipse plugins that comes from the fusion of JBoss IDE and Exadel Studio. The main advantages it offers are: JSF development improved, in particular integrating RichFaces and Ajax4JSF libraries; Seam (next J2EE middleware standard?) integration; and plugins for JBoss, Hibernate... Here are my first impressions."

Comments

But does it support JCV

[DelitaTheFridge]

I need an integrated IDE solution to support the latest JBC and WAJAX 2.1.2 standards, along with full SDJ support. Can this do that? Seriously, do these flavor of the month java libraries mean anything to anyone?

Re:But does it support JCV

[Azarael]

I'd settle for an IDE that would actually build and deploy one of their *example* J2EE projects.. apparently this is the hard part.

Re:But does it support JCV

[bigtangringo]

RAM, NIC, HDD, XHTML, CSS, XML, PNG, JPEG, BSD, et cetera ad nauseum.

Seriously, do these acronyms mean anything to anyone?

In case you may have never considered it, people actually involved with these technologies may actually know what they mean and do.

Re:But does it support JCV

[WED Fan]

Whole new bunch of acronyms are coming down the pipe. I don't think all are web, and may not be associated with a particular OS or file format. Microsoft just trade marked, patented and are looking for fast track ISO approval of the following:

• SUP
• ERCA
• LIFR
• AGILIST
• ICEX
• PIA
• LIDOC
• IOUS

This according to an attorney in their IP department, M. Poppins.

The sound of it is quite atrocious.

Does anyone have insight on these?

Re:But does it support JCV

[Anonymous Coward]

> Seriously, do these flavor of the month java libraries mean anything to anyone?

Yes, they make them specifically for no-talent hacks on slashdot can sneer at them and say how they could write a distributed transaction backend with reliable multicast messaging using PHP and MySQL in a week.

Re:But does it support JCV

[andphi]

Have you tried a spoonful of sugar?

Re:But does it support JCV

[kalirion]

Hell, I could do that in an hour, if I can define the following terms:

distributed
transaction
reliable
multicast messaging

Re:But does it support JCV

[Anonymous Coward]

You should settle for a noose. It's quicker, and less painful.

Re:But does it support JCV

[FatherOfONe]

Well put.

I have to say that the place I use to work standardized in Java and Linux on the servers and everything worked well. Then I left and they hired a new kid out of college that couldn't believe how "long" it took to code stuff using JSF/Java. He pushed and pushed to do a project in PHP. He quoted around one fifth the time to do a project that another developer had quoted to do in Java, saying that the time using PHP would more than make up the difference. In short, the project took around 2X as long as the original Java quote (~10X as long as he had quoted) and thus they are back to working with Java again. I am not about to say that any language is bad, but when you focus 70+% of your effort on business logic, (most of our work), then it is a little hard to believe someone when they say that language X is 5X faster than language Y.

My question about this new IDE from RedHat is this:
Can I do visual JSF development in a true WYSIWYG environment like Netbeans?

Can I do Swing development in a WYSIWYG environemnt like Netbeans?

Can I easily choose not to use the custom components that you include? I would assume so, but my fear is that RedHat focused on this product working with JBOSS and getting it to work with other application servers may be a pain.

I like Eclipse, but I have found Netbeans 5.5 to be better for what I do so migrating back to Eclipse would take some great features, and would be interested to see how far this has come. Oh yeah, and one last, but very important thing. You still don't hack Eclipse on Linux to run under the GCJ crap do you? If you did that then I can only imagine all the problems I had using your product before would be back again. I hope now that Java is under the GPL that you don't mess with that abomination (GCJ) and have included the real JVM with your Linux and more specifically don't have any of your tools reliant on the GCJ.

Re:But does it support JCV

[Anonymous Coward]

Or perhaps flying a kite will take your mind off of it.

Re:But does it support JCV

[LDoggg_]

Why the hate towards GCJ?

Sure it wasn't as fast or complete as the Sun JVM, but it was being developed as an open source project during the years that Sun had kept still their JVM and libraries proprietary.
Heck, gcj and gnu classpath were good enough to run tomcat a couple of years ago, and it's not like anyone was forced to use it. While Sun's jvm was proprietary, Redhat could only bundle eclipse with gcj on their own distro. But it never took away from anyone wanting to run eclipse on the sun jvm or adding Sun's jvm to to the jvm target list in the bundled eclipse.

And for the record, there was never a time that you HAD to hack eclipse to run under gcj to get it work on Linux. The Sun jvm for linux was released before eclipse was, and eclipse ran fine on it well before it was able to run on top of gcj.

Re:But does it support JCV

[FatherOfONe]

I guess I do hate GCJ now that you mention it. The countless bugs I had to fight just to realize that it didn't work well has left scars on me that won't go away quickly.

I also know some people at RedHat (not well, I will admit) and their hate for Java makes view of GCJ appear to be love. They flat out HATE HATE HATE Java. Now they kept saying that they "couldn't ship Suns JVM", but the reality is that they don't want to ship one or include one easily like SuSE or Ubuntu.

If I implied that Eclipse had to be hacked to run on the JVM then I am sorry. I meant to say that RedHat hacked Eclipse to run under the GCJ, and thus lots of things didn't work well. The developer was left to think that Java + Eclipse was installed and would work great, yet it was this hacked version of Eclipse and GCJ that it was running on. The thing had a ton of bugs and would technically run, but not well.

Now it looks like RedHat/Fedora is looking at including the GPL version of Java, once they do that and then use this JVM as the main/core system for all Java programs then I will definitely change my tune on Redhat. Until then I will be skeptical.

Lastly, I want to make it clear that I am actually a Redhat fan and wish them well, but I ask you to talk to some of their engineers and ask them their opinion of Java. Now that they have bought Jboss it might be different, but a few years ago, I would equate it to walking in to Microsoft and asking them what they think of Linux if Linux just took 40% of their business.

Their mantra use to be "You get to Redmond through Palo Alto".

Re:But does it support JCV

[aevans]

What's the alternative? Some other part of your body being smart?

HuH?

[russ1337]

"This IDE is a bunch of Eclipse plugins that comes from the fusion of JBoss IDE and Exadel Studio. The main advantages it offers are: JSF development improved, in particular integrating RichFaces and Ajax4JSF libraries; Seam (next J2EE middleware standard?) integration; and plugins for JBoss, Hibernate.."

Now I know that is in English, but I have no idea what half of those words mean.

Re:HuH?

[140Mandak262Jamuna]

You think he knows more about the words than you do? Those who know, do. Those who don't talk.

Re:HuH?

[Frymaster]

You think he knows more about the words than you do? Those who know, do. Those who don't talk.

those who don't and think they do, post on slashdot.

Re:HuH?

[Anonymous Coward]

I think I know what you mean.

Re:HuH?

[Bluesman]

That's not the submitter's problem. You need to bone up on some acronyms, or you'll never make a goodJavaProgrammer. Here's a quick lesson in what you should do:

Write a thousand different programs using acronyms that start with J that do nothing except fuck up the data as it's being transmitted between the database and your application. Then, you have to write automated tools that also are acronyms that start with J and contain the word "Bean" in there somewhere, and those exist to generate parts of those previously mentioned thousand programs.

Then, write some Swing components that have nothing to do with all of this, and call those by almost exactly the same names, so that people get confused and can't do a proper Google search for documentation. Name an IDE after the Swing components, too.

Finally, call it all "middleware," give it yet ANOTHER name and bundle it all together, making sure that everything breaks if you don't include fifty different XML configuration files in the proper directory hierarchy that changes with each version.

Then when all of this doesn't work for more than one project because it's hopelessly complex, do it all over again and call it the next greatest revolution in Java middleware.

Re:HuH?

[russ1337]

LOL, you don't sound bitter or anything.....

You made my day.

Re:HuH?

[Eddi3]

If you don't know what was said there, you probably don't need to know.

Re:HuH?

[Bluesman]

Actually, I was being a bit facetious.

People actually DO get J2EE apps to work. Here is a very informative instructional video by some Japanese researchers who show how it's done:

J2EE Example

Re:HuH?

[mattgreen]

Hahaha, this is so dead-on.

Too many programmers out there that think they can write middleware, when oftentimes the best thing to do is just to Finish The Damn Project.

Re:HuH?

[TheLink]

I find it funny/scary that the XML config files are often larger than fairly complex programs written in other languages ;).

Re:HuH?

[Reverend528]

Then, write some Swing components that have nothing to do with all of this, and call those by almost exactly the same names, so that people get confused and can't do a proper Google search for documentation.

You know, you can avoid these kinds of naming problems if you just add a few more subdomains to your website.

Re:HuH?

[thelima]

C'mon! Haven't You heard about VS.NET (aka Orcas) and shiny new C# (formerly known as NGWS) with LINQ and DLINQ support, together with ASP.NET, VB.NET and WCF (aka Indigo), WPF (aka Avalon), WCS (aka Infocard), WCS (aka Workflow), running on top of CLR and MSIL (which is MS implementation of ACME CIL) which together is known as CLI, of course running on top of Vista (aka Longhorn). Just to scrape the surface. Ya 'now, it's the digital era...

Re:HuH?

[Reverend528]

I believe the appropriate tag is 'yahooseriousfestival'.

Re:HuH?

[morgan_greywolf]

Then, take the 'middleware', bundle it with hacked-over versions of Tomcat, Log4J, J2EE, etc., into a 'suite' and call it some bicaptialized name like Web and some shape like Cube or Sphere or something, with each product named some two-letter acronym like AS for the application server, and MQ for the 'middleware', etc.

Finally, get sued by IBM for violating their patents and trademarks. ;)

Re:HuH?

[Bluesman]

Me too. I think that's what happens when you try to make a all-in-one solution without realizing there's a tradeoff between power and flexibility. You either make something that's too specific to be useful so that users have to constantly fight against design decisions you made, or something that requires almost as much work to make it do what you want than writing it from scratch would.

The ridiculously complex configuration files are a symptom of moving as many design decisions as possible out to the last possible moment. Complexity isn't reduced, it's just in a different place.

Which, ironically, makes the whole thing that much more complex, since now you have multiple places things can go wrong.

I tried using this stuff years ago, and found it wasn't close to worth the hassle, especially for a single developer. I just did a search for "J2EE success story," and the vast majority of hits were about a small team of Python programmers replacing large J2EE teams that failed to produce a working product.

But maybe I'm wrong, and the people who know much more than I do about this can list a hundred different projects where J2EE saved the company. It just seems like it's overhyped and people are really much more concerned about the scaffolding they're using than the work that they're supposed to be doing with it.

Re:HuH?

[gatkinso]

And thank god you don't need to!

Re:HuH?

[LiquidCoooled]

Wasn't the .net CLI going to be called gonad or something?

Re:HuH?

[OneSmartFellow]

This is the typical problem that developers in large corporations run into. Do I hard-code these values or put them in a fragile configuration datastore.

There seems to be some unique requirements in corporates where configuration is an issue.

I have struggled with this basic problem for years, and have yet to come up with any hard and fast rules about what should be hard-coded, and what should be configuration.

This is where experience comes into play. Stuff like server names, database names, comms ports, and locale make sense to put into configuration (even though by rights they should all be 'discoverable'). If your config entails more than about a dozen or so options, you're gonna start hating life.

When your job as a developer includes third-line support, and you spend hours debugging only to discover *it's a configuration issue* , you'll come to the same conclusion.

Re:HuH?

[bberens]

I've successfully used JMS in a high availability application (security system monitoring software) before and I have to say that you're correct. J2EE stuff is nice but it can also be very complex. In our situation JMS worked out exceptionally well and saved a ton of development time. However, it's just a tool. 99.9% of the time you don't need J2EE stuff to accomplish what you're doing. I think of J2EE in many the same ways I think about AJAX. It's a nice toy/tool but a little bit can go a long way and it should be used sparingly. Unfortunately in the corporate environment we tend to get buzzword happy. Not much you can do about it.

Re:HuH?

[Anonymous Coward]

Now that is the best introduction to J2EE that I have seen... and I've been developing in Java since 1999. I've stayed away from J2EE exactly because of this, and stayed on the client side where theres less bull shit. Used Swing and Eclipse RCP, but had more fun with client side stuff with Thinlet (thinlet.com) initially written by one guy.

Re:HuH?

[kaffiene]

...just like most of the morons on /. who claim that Java is crap yet have no idea how it is actually used.

Re:HuH?

[Gazzonyx]

You think he knows more about the words than you do? Those who know, do. Those who don't talk.

those who don't and think they do, post on slashdot.

When did management find out about slashdot?! Quick, distract them with a whitepaper, I have to change my bio!

Re:HuH?

[josath]

you need to de-spam your blog. or are all those comments about "gay wrestling" really legitimate?

Re:HuH?

[Shados]

What do you mean?! We know how its used! Have you seen JSP?! Its exactly like PHP! You stick all your logic and your tags in one big file with a bunch of conditionals, and you're set! Of course we know how java is used. Oh, and servlets? Same thing as PERL/CGI! Wha? Whats a design pattern? MVC? Strut? Bah, buzzwords!

Re:HuH?

[filterban]

We use J2EE (Hibernate, JSF, Tomcat/Weblogic) at work, and overall, it's a great solution -- once you're familiar with it. You're right - it may not make sense for a single developer, especially one that doesn't have an understanding of how it all should tie together.

Due to our use of J2EE/JSF, we have large numbers of reusable components that make our application development (generally speaking) very quick and straightforward. Our apps generally speaking run well and efficiently. Done right, JSF/J2EE/Hibernate works great. You just don't get the newbie-friendly environment that .NET gives you.

Honestly, the best part of JSF is that it is so testable. You can easily integrate with JUnit and have it run unit tests so much easier than with any other framework that I've seen. Additionally, JSF's multi-language support is ridiculously cool. We had to migrate an existing app that was English only to English/Spanish, and it quite literally was a matter of adding a new properties file defining the new text/image links and basically - the best part - one line of code.

That's not to say that Python isn't an effective solution, either. It is. But J2EE has its place too.

You could write your entire web app using Perl, PHP, Python, C, Ruby, J2EE, or any of a number of frameworks. Generally speaking, they are all effective tools, and you just have to pick the right one for your situation.

Re:HuH?

[computational super]

I just did a search for "J2EE success story," and the vast majority of hits were about a small team of Python programmers replacing large J2EE teams that failed to produce a working product.

I'm not convinced that this is entirely Java/J2EE's fault as much as it is that big corporations love Java and hate Python (and Perl) and have "big-corporatized" Java. Java was actually pretty cool when it first came out... if big industry embraced Python today, by tomorrow you'd see an explosion of PBPEL, P2EE, PDBC, PFaces, PSF and PMS applications and by the next day you'd be assigned to a committee to evaluate the product vendors to find the scalable enterprise solution that was the best fit for your business integration challenges. After a month of vendor selections, you'd narrow the candidates down to the top two contenders, and they'd pitch their expertise in the field, and then your company would sign a multi-million dollar contract with one of them. Then you'd start the process of trying to figure out how to get around the inherent limitations of the "solution" and do plain-old Python programming inside the "container" without it being too obvious to the higher-ups that that's what you're doing...

Re:HuH?

[Rycross]

Where I work, this kind of thing is caused by me going to business and saying "The spec says this, but that can mean either A or B. Which do we really want?" and the business, not wanting to make a commitment that reflects poorly on them, says "Make it configurable!"

Re:HuH?

[El_Oscuro]

This works for all languages and methods.. The only time you don't want to do this is if you have to maintain YOUR own code.

Re:HuH?

[TheLink]

Actually I disagree "slightly". I see power and flexibility as related.

You can have power, flexibility and even ease of use.

The trick is to pick the right defaults. It's definitely not always easy for the designer.

What put me off from Java when it first came out was stuff like:

Method to find the number of rows returned from an SQL query - go to the last row, get row num. Lots of stuff like that.

Basically a lot of the libraries felt like they were written by someone who was told by someone else to write it because it was in the requirements list. Not written by someone who was going to use it to make his life easier.

My colleagues then who had to use java were regularly encountering things like that. I was doing perl at that time, and I was quite surprised at how hard it was for my colleagues to do simple stuff in java. Maybe they sucked, but we searched for better ways (online etc), there often were no better ways.

Well maybe it makes the java programmers feel like they've done a lot of great work at the end of the day, and that they've earned their salary ;).

Re:HuH?

[Anonymous Coward]

Fuck. I'm with you. And I do Java coding for living (been doing it since '99).

Re:HuH?

[iwan-nl]

PMS applications? *shudders*

Re:You mean...

[LinuxGeek]

Yup, they are still around despite the best efforts of SCO. And RH is making useful and relevant (cross-distro too) tools unlike the Me crap from SCO. I am running this on ubuntu and can even deploy on different platforms! Sounds like RedHat is confident in the quality of their products, unlike that sue-happy company.

Re:You mean...

[jellomizer]

Your confusing Red Hat with SCO.

I for one...

[markov_chain]

...welcome in advance our glossary-posting overlords!

Sounds promising..

[jshriverWVU]

for Java developers. What I'd really like to see is something like Project Builder/Xcode or Visual Studio for Linux. (Not that I like Visual Studio) but if you need to pound out a GUI it's pretty easy. There's a relatively large learning curve for developers wanting to do work for X. There is a myriad of libraries with their own widgets to choose from, languages, IDE's. It get's a bit confusing when you just want to sit down and code something, or add a GUI to an existing cli app.

Trolltech's suite so far has been the best one I've seen yet but has licensing issues. I've tried KDevelop and it's not that bad, but still not great. The ones I've seen for gnome have been even harder.

A good IDE for developing GUI applications, should help the developer a bit more with the GUI stuff and not make it mandatory that you know every call to every function of every widget for whatever library that package supports. If you knew that, might as well stick with Emacs/vi/nano and code it. Which it seems is how most development is done. (which isn't bad) but makes it harder for someone else starting out and wanting to give it a try.

Re:Sounds promising..

[ianare]

There are several IDEs for linux, my favorites: boa-constructor, glade.

Re:Sounds promising..

[Anonymous Coward]

try netbeans (www.netbeans.org). it's fairly straight forward, and has a nice gui builder for swing called matisse. it's drag and drop from a palette.

personally, I use eclipse with swtbuilder and other stuff, but netbeans out-of-the-box experience is much better.

Re:Sounds promising..

[jshriverWVU]

Can you use Netbeans/Eclipse/etc with languages other than Java?

I downloaded and checked them out briefly before, but it seemed like all of the good Gui IDE's for Linux were all geared toward Java. Would like something for C/C++.

Re:Sounds promising..

[TheAwfulTruth]

GUI builders are not IDEs they are one of MANY components to a real IDE.

An actual IDE takes care of GUI design, code editing, debugging, project management, documentation, source control and on and on. And they are scripted environments with plug-in interfaces, compilers and debuggers source control front entds etc. so that you can choose what programs you want to use for what tasks.

That is an Integrated Devlopment Environment. Admitedly IDEs are not for everyone but as this is the subject...

Visual Studio is hardly "perfect" but it is BY FAR in advace of /anything/ I've used on Linux that calls itself an "IDE".

Re:Sounds promising..

[Jeek Elemental]

eclipse can switch to c/c++ among many others, sniff around http://www.eclipse.org/

Ive just started using it, probably overkill in the extreme for me but its very nice and works out of the box for gcc.

Re:Sounds promising..

[Reapman]

My problem is integrating a library with em... I'm by no means a pro at programming, but I don't think it should be nearly as complex as it is to add libraries such as SDL to Eclipse, KDevelop, or Ajunta (or w/e it's called) Or importing existing code into a project. As much as I hate em, I still haven't found an IDE as polished as MS's Visual Studio (yeah yeah all you really need is gcc and vi, but I like my Full on IDE damnit :D)

Re:Sounds promising..

[davek]

Have you tried Lazarus and Free Pascal?

http://www.lazarus.freepascal.org/index.php

It looks very promising for people familiar with the Delphi brand of RAD (Rapid Application Devel., I don't even know if that acronym is still hip or not). I myself haven't written a project in it yet, but the few tests I've thrown at it seem to work OK. Its the most functional IDE I've seen for Linux/X11 yet.

Re:Sounds promising..

[kazade84]

What I'd really like to see is something like Project Builder/Xcode or Visual Studio for Linux.

Two words... Code Blocks

Re:Sounds promising..

[thelima]

De gustibus not dispu^H^H^H^Hblah, blah, blah. But You know - I could hardly consider development environment something without support of unit-testing built-in (don't tell me about VSTS). I could sacrifice all graphical wizards for single one feature of Eclipse IDE which is not available in VS.NET (without third party plugins) - live compilation together with quick fixes. To give You feeling, when I have a bug it is immediately highlighted (no save no magic keystrokes). Not only this but also I can see *ALL* dependent resources as "invalid". I sut, the, simply press CTRL+1 I usually get useful quick fixes offered by IDE. Such as: surround with try catch, initialize plus dozen other. See Quick Fixes in Eclipse. Of course there is much more juicy details unavailable in VS. Light years ahead refactorisation, far, far better auto-complete (in VS.NET is just a toy, nothing comparable to eclipse), excellent support for team-programming - just two words - Mylar and Team support - show me something comparable to structural diffs in VS.NET. Should I write more? VS.NET is just below average IDE, with lot's of useless Wizards You will never use in real life, but lacking some *crucial* thing others have for Years. I could bet, I can write comparable java webapp without fancy wizards faster than You wit those wizards...

Re:Sounds promising..

[Anonymous Coward]

You should use IntelliJ or Eclipse then. VisualStudio is five to ten years behind the state of the art IDEs.

Re:Sounds promising..

[LeafOnTheWind]

For serious Java and J2EE developers (I gather you're only using the C/C++ or Java Standard edition), it is the only reasonably tool. NetBeans is ok, but for real enterprise development Eclipse is unparalleled. GUI building utilities are decent, but they're irrelevant in J2EE applications and usually a waste when designing large applications. Instead, Eclipse provides things like Ant, JUnit, JSP, multiple server integration, EclEmma, Subclipse, and literally dozens of other utilities. I know it sounds like overkill if you're just designing a small app, but for large projects each and every tool is vital to productivity. For example, using JUnit, you can separate your tests, but at the same time keep them synced with each and every class and method in your program. In addition, on failure JUnit provides extensive exception/debugging handling utilities and stack traces. By consolidating all your cases together, it makes regression testing as easy as pressing a button. Along with the JBoss and other enterprise development plugins, Eclipse becomes your entire programming platform, including your web browser.

Re:Sounds promising..

[Doctor Memory]

I still haven't found an IDE as polished as MS's Visual Studio Yeah, when you have complete control of the widgets, libraries, languages, run-time environments and the OS they run on, you can crank out some pretty slick stuff. I get a kick out of Windows asking me if I want to debug stuff like Acrobat Reader or Lotus Notes because they've gotten an access violation or some such. Visual Studio isn't the best IDE I've used, but it's definitely the best integrated. And the fastest — if you single-step through some of the .NET stuff and have some non-trivial variables in the watch window, it's fairly speedy. The amazing thing is, if you drill down into some of the collection classes or system objects, they've got literally hundreds of attributes. And VS manages to keep the window updated as I step, no matter how deep I've gone.

At least, it seems to...I'm not sure I'd be able to tell if it didn't...

Re:Sounds promising..

[Anonymous Coward]

eclipse works better than Visual studio for me.
and yeah, same applications can be written in java or c, unless u are a language fanatic.

Re:Sounds promising..

[istartedi]

I could hardly consider development environment something without support of unit-testing built-in

OK. Enough already. How much effort does it really take to add an extra target to your makefile, run the unit test and diff the expected output against the actual output? Every "unit testing framework" I've ever looked at has been thousands of lines of cruft just so you can do something like "assert(output < 2);".

I've got an open mind though. Convince me. Please don't cite Cxxtest as an example. That's exactly the kind of junk I'm talking about. Pulling in Perl or Python and munging a header into a test app, just for simple assertions? Pulleeze.

Re:Sounds promising..

[0kComputer]

The quick fix thing you speak of looks like somewhat of a crutch for sloppy developers who don't know how to write quality code in the first place. I'm not sure what "Wizards" you're speaking of w/ VS.Net. The wizards thing was very prevelant in the C++/VB 6.0 stuff, but it's basically non-existant now.

Unit testing is included in team system, as well as a very powerfull merge/diff tool. Any GUI development (ASP, Winforms) is orders of magnitude easier in VS.net than its java/eclipse counterparts. Hell, If i wanted, I could even write an XNA app and play it immediately on my X-Box. VS.net may not be perfect but its the best thing out there, but that's common knowledge.

Re:Sounds promising..

[thelima]

OK, You definitelly don't understand unittesting. Sorry. But answering Your question - sure, it's not a problem add few lines to Makefle, but, You know, most up-to-date IDE have something more than just "run my test suite". Like - find if given code (cless, function) is covered by some test (if yes - find those tests). Run only one test from with the whole suite (ex. just this function, I'm currently editing). Auto-generate test fixture for given class. Rerun only recently failed test. etc. Just try how does it works in eclipse. Feel the difference.

Re:Sounds promising..

[thelima]

The quick fix thing you speak of looks like somewhat of a crutch for sloppy developers who don't know how to write quality code in the first place. Nope, it doesn't definitely. It's one of the best tools out here raising developer productivity. Consider TDD. You write test first. Without single line of real code. Of course at the beginning You've lots of errors - like cannot resolve existing types, cannot find given function etc. Just press CTRL-1 and You have this fixed in one second. Just try it.

Any GUI development (ASP, Winforms) is orders of magnitude easier in VS.net than its java/eclipse counterparts. HTML development is NOT a DEVELOPER task - it's HTML designer task. I don't care about it. If i have to write webapp, I write really simple HTML, develop MVC to make it working and give the rest to HTML designer. I don't need fancy HTML tags editors - those are just slower.

Unit testing is included in team system, as well as a very powerfull merge/diff tool. How much does it cost in Microsoft case to get this *basic* functionality? Does it powerful "merge" tool works with CVS or SVN?

Re:Sounds promising..

[Vintermann]

> Visual Studio is hardly "perfect" but it is BY FAR in advace of /anything/ I've used on Linux that calls itself an "IDE".

What about eclipse?

Re:Sounds promising..

[Vintermann]

Netbeans no, I don't think so, Eclipse, definitively. There are extreme amounts of eclipse plugins out there, the problem is getting them to work, setting them up and understanding them.

Re:Sounds promising..

[Xiaran]

Eclipse provides things like Ant, JUnit, JSP, multiple server integration, EclEmma, Subclipse

OK, Im a J2EE developer that uses mainly Eclipse(and emacs) in day to day use. And *all* of the things you just mentioned are available with netBeans(I mean Ant? Old crappy netBEans had Ant support). Ive been using netBeans recently(I used to loath it with a passion years ago) to do some GWT stuff and I found myself pleasently surprised.

Re:Sounds promising..

[Xiaran]

http://www.netbeans.org/products/cplusplus/

http://www.ibm.com/developerworks/opensource/libra ry/os-ecc/

Re:Sounds promising..

[Xiaran]

Oh and netBeans seems to have a lot of ruby support these days... tho I havent really used it.

Re:Sounds promising..

[LeafOnTheWind]

Dude, emacs?? But you're right, NetBeans can add most of that functionality with plugins, but eclipse has little things that just make you smile - like holding control to follow a link to a function or SWT vs Swing. Also, I find that Eclipse has better integration with WebSphere. Granted, I haven't used NetBeans in forever, but I found that Eclipse provided a better platform for rich client applications - if I ever need to write a python script, all I need to do in Eclipse is create a new Python project. Also, I've heard that Aspect Java fell through with NetBeans, is that true? Anyway, the points I was making weren't to disparage NetBeans, but to point out that the tools in Eclipse are actually used by people and aren't just fluff. Both are certainly better than vi, javac, and a web browser. :)

Re:Sounds promising..

[Xiaran]

Actually Im more of a eclipse guy... Ive just been using netbeans for a private GWT project... I havent used it really in years and Im quite impressed on how much better its got.

Oh and emacs... Ive been using it for years. I can use it over text only interfaces where I usually have a nice ant build process for everything. WOrks a treat for remote working on development machines

Re:Sounds promising..

[KlomDark]

Ha ha, TDD. What a dumb fad. You've been sucked into the Cult of the Fat Dork. TDD guys are usually the worst fuckers to have to code with. It's a big crutch to allow companies to use lesser-skilled developers while making sure they don't make a mess of things.

Next you'll be telling me how Agile programming can change your life.

It's all a stupid cult for bad programmers. The one article I always avoid in my DDJ is the Scott Ambler blabbering of this months new cultist pep talk. Complete drivel.

You write really simple HTML.

Why would the Team Server merge tool work with Subversion? TFS is also a source management system, far superior to CVS, about equal with Subversion. Oh, and yes, it actually does. You can even tweak TFS to use Subversion as it's source management system. (In a pure .NET environment that would be stupid, but in the case of integrating projects with Unix/Linux teams, it's a really nice thing.)

Re:Sounds promising..

[Anonymous Coward]

Linux(or rather Unix) is the IDE. My GUI builder is one component of my Linux workstation. I don't think this environment is perfect, but it is far better than any version of Visual Studio I've used.

Re:Sounds promising..

[thelima]

Ha ha, TDD. What a dumb fad. You've been sucked into the Cult of the Fat Dork. TDD guys are usually the worst fuckers to have to code with. It's a big crutch to allow companies to use lesser-skilled developers while making sure they don't make a mess of things. Next you'll be telling me how Agile programming can change your life. It's all a stupid cult for bad programmers. The one article I always avoid in my DDJ is the Scott Ambler blabbering of this months new cultist pep talk. Complete drivel. You know, if You don't understand something, it doesn't automatically mean it's "dumb". Stating such a bold statements about something You definitely didn't use nor seen in real life is - well - definitely "dumb". That's my comment to above.

Why would the Team Server merge tool work with Subversion? TFS is also a source management system, far superior to CVS, about equal with Subversion. Why? See, world is sometimes more complicated than Microsoft describes to You, my Friend. Whatever VSTS is nice or not - it has few eliminating drawbacks:

• it's exclusively tied to Microsoft technology,
• supported almost exclusively by microsoft technology
• virtually not supported by anything else
• it's extremely expensive comparing to to value it brings (it's probably why You didn't answered how much doest it cos from my prev. post)

And world nowadays is little bit more complicated. Unless You are all-Microsoft house, and most entities nowadays are not, VSTS in not usually a choice. And I don't tell You only about Java.

Oh, and yes, it actually does. You can even tweak TFS to use Subversion as it's source management system. (In a pure .NET environment that would be stupid, but in the case of integrating projects with Unix/Linux teams, it's a really nice thing.) And, could You answer please - how does it cost per seat to have such a rudimentary functionality like diff/merge and unit testing built-in, in the Microsoft case? Just to have fair comparison of free (in the Eclipse case) and non-free products.

Re:Sounds promising..

[KlomDark]

"And I don't tell You only about Java."

What you mean there?

Re:Sounds promising..

[thelima]

It simply means that virtually nothing, except VS, has built-in support for TFS. I mean useful tools like: SQL tools (say TOAD, TORA etc), Enterprise software like UML modelers, Project management software (including MS Project right know), documentation systems, BPMS modelers and lots of other tools, which usually have excellent CVS and SVN support built-in.

Re:Sounds promising..

[dodobh]

(x)emacs.

Re:You mean...

[Reverend528]

Java is too apparently.

TextMate

[Paulrothrock]

I have yet to see an IDE that has a text editor that compares to TextMate. The fonts are ugly, color/theme management is poor, integration with the PC is poor or non-existent, and macros and custom code are much more difficult than TextMate.

These may be good when you need to manage massive projects, but I can't stand to use them for actually writing code. If there was only some way to replace the text editor in these IDEs with TextMate but keep all the trappings that make compiling and deploying these apps easy.

Re:TextMate

[Alan Shutko]

Actually, the IDEs (or at least Eclipse) seems to fall down in our department when dealing with larger applications. It's great for toy apps, but a lot of the plugins have problems scaling. The core ones are usually adequate, but third party ones are dreck.

Re:Promising and Eclipse in the same sentence?

[nuzak]

There are some truly nifty plugins for Eclipse: Mylar comes immediately to mind. Exadel Studio has some decent functionality, but it's grotesquely slow and not terribly stable either. And JBoss IDE always struck me as a few miscellaneous plugins that didn't really accomplish anything. The Hibernate stuff is all right if you're using the xml files, but I use JPA, and it's useless there.

All I really want are some clicky wizard dialogs for the functionality in seam-gen, and a decent stable IDE for Drools/JBossRules (which I don't think RHDS even includes).

Netbeans would be a more ideal IDE except for how it makes you really regret it if you subvert the bureacracy of the "add new file" wizard in a project. With eclipse, I can just drag new files in and hit refresh. That and Netbeans still has no TestNG plugin (the existing one always sucked, and doesn't even work in recent versions). I suppose I could pay for IDEA if I was still doing Java full time.

Re:Promising and Eclipse in the same sentence?

[Anonymous Coward]

I love Aptana for my Rails programming. It's pretty damn awesome.

Well that was brief

[Anonymous Coward]

I have to say this a pretty scant entry for the front page of /. Amazing it got through really. Still this does look like quite a nice tool - I hope the port it to one of the better IDE's eventually. Eclipse always seems to me to be the Windows of Java IDE's. Hugely, unaccountably successful. It does everything but it doesn't do anything well. I actually prefer Netbeans of the free/open source tools though neither are a patch on IDEA.

"next J2EE middleware standard?"
Probably. Parts of Jboss Seam and Google Guice amongst others are being formalised into the JEE6 spec -
See http://jcp.org/en/jsr/detail?id=299. Oh and its JEE now, not J2EE. Has been for quite a while).

Re:Promising and Eclipse in the same sentence?

[thelima]

Could You name one that "Company"? I worked for World biggest Healthcare company, few Financial Institutions (including Banking) whose standardized on Eclipse with great success. You know, real programmers, don't necessarily need fancy GUI wizards for composing HTML pages, drag-and-drop class/build-file wizards, programming etc. - simply because developers doesn't project HTML, unless You work for one-person, swiss-knife DYI "company". You know, usually companies have HTML developers who use their own tools (and those are not Eclipse nor VS, from what I can see). So without generalization please. Artur

Some info about our project

[_marshall]

Hey everyone.. I work on the JBossTools and RHDS Team and just wanted to give some community-level info about our project.

Red Hat Developer Studio is our commercial offering of the JBossTools open source project (formerly known as JBossIDE), which has a vibrant community of users and contributors. You can check out our project(s) at the following URLs:

JBossTools main page: http://jboss.org/tools
JBossTools blog: http://jbosstools.blogspot.com/
JBossTools 2.0.0.beta3: http://sourceforge.net/project/showfiles.php?group _id=22866&package_id=242269&release_id=531957
RHDS 1.0.0.beta1 (based on JBossTools 2.0.0.beta3): http://www.redhat.com/developers/rhds/index.html

Feel free to drop by #jbosstools on freenode, we'd love to hear from you!

Right idea, wrong reasons!

[Gazzonyx]

That's not the submitter's problem. You need to bone up on some acronyms, or you'll never make a goodJavaProgrammer. Here's a quick lesson in what you should do:

Write a thousand different programs using acronyms that start with J that do nothing except fuck up the data as it's being transmitted between the database and your application. Then, you have to write automated tools that also are acronyms that start with J and contain the word "Bean" in there somewhere, and those exist to generate parts of those previously mentioned thousand programs.
Huh? Are we using JBuilder? Bean... ah, we're talking J2EE. Something to build...something to build... if only we had 'factories' or something.

Then, write some Swing components that have nothing to do with all of this, and call those by almost exactly the same names, so that people get confused and can't do a proper Google search for documentation. Name an IDE after the Swing components, too.

How about that 60 meg folder called 'Docs' that comes with the JDK? It's even got pictures! You can drill down to the 'swing' section (think about the naming and that 'J' thing again while viewing this vs. the SWT ;) )

Finally, call it all "middleware," give it yet ANOTHER name and bundle it all together, making sure that everything breaks if you don't include fifty different XML configuration files in the proper directory hierarchy that changes with each version.

I guess it could be called 'glueware'. Try starting with '.', the hierarchy descends from there. Regardless of what your manager read in 'Buzzword of the Moment Daily', you don't have to use any XML.

Then when all of this doesn't work for more than one project because it's hopelessly complex, do it all over again and call it the next greatest revolution in Java middleware.

Trying using SCM, separation of concerns, encapsulation, and polymorphism.

All that being said, I don't much like java; I just think you're flaming it for the wrong reasons. How about if statements, operator overloading (or lack, thereof), switch statements, and the fact that it FINALLY just got autoboxing? Oh, yeah, and could they make the object names any longer? I can usually almost instantiate an object on a single 80 column line if my variable name is less than 4 letters. And why the crap am I always having to manually repaint the screen?! Just my $.02, YMMV. Fire in the hole!

Re:Right idea, wrong reasons!

[dkf]

Huh? Are we using JBuilder? Bean... ah, we're talking J2EE. Something to build...something to build... if only we had 'factories' or something. You do realize that 'factories' are old hat? You need a factory factory factory these days...

Re:Promising and Eclipse in the same sentence?

[Anonymous Coward]

I know the parent is just flamebait, but for the record I'm blown away anyone would have this opinion. After using visual studio for several years, switching to Eclipse at a new job was a breath of fresh air. Sure you don't have the drag-and-drop UI design, but that aside the out-of-the-box features blow VS away and that's not even starting with the plugins, many of which are free and open source like the IDE itself.

Re:13 years of hype

[linuxIsLife]

it is still slow really ?

Re:Promising and Eclipse in the same sentence?

[Anonymous Coward]

You folks haven't done non-Java/Ruby (aka, non web app) development in Eclipse it seems. The CDT is a waste of bits - other than nice syntax highlighting, most of the eclipse functionality doesn't work.

Re:13 years of hype

[porkThreeWays]

I think the most frustrating thing with Java is how many things they did wrong in early versions that we have to deal with today. It's an extremely confusing language to learn from scratch because for everything that doesn't make sense, someone has to give you a history on why it's done that way. I agree that half of Java has been marketing (but hell, half of .net is marketing too!). It really is an amazing language with a very rich set of API's provided by Sun, don't get me wrong. I just think Java has such a bad history it may never recover from all the preconceived notions it is currently entangled in.

Jesus Christ, will someone please rip off ASP.NET?

[melted]

Jesus Christ, will someone please rip off ASP.NET? I've looked at all the crazy proliferation of Java web frameworks and they all pretty much suck. You have to maintain a bunch of XML files for things that ASP.NET just figures out on its own, docs suck, architecture is bizarre. It's all just a giant, productivity draining mess. Why can't I just have transparent interaction between the page and code? Why do I have to "register" crap (through XML file) that should just be available transparently from page code? Why do I have to create "navigation rules"? Why do I have to "declare beans"?

No wonder turds like Ruby on Rails are so popular. I'd rather shoot myself than use Java for web development.

Ahhh Kylix

[Anonymous Coward]

The linux community had this, it was called Kylix. It was bought to u by Borland. Borland priced it too high and the free versions weren't adopted by the OSS community. But tools like Visual Studio don't come free, but you really can't fault Borland for wanting to charge for it products.

Did you try Netbeans

[Lobais]

I believe the GUI editor of the nice free Netbeans IDE is working something like the one of Visual Studio.
Not that I ever tried Visual Studio, but Netbeans Matisse GUI editor is really some of the most impressive UI stuff I've seen.

If you want a GREAT development environment...

[FlyingGuy]

Start getting everyone you know to start leaning on the folks at CodeWorks to get Delphi & C++ Builder ported over to Linux. Say what you will about Borland, but imagine those 3 tools being completely X-Platform between Win32, Linux ( Gnome & KDE ) and OS-X. That alone would put a very LARGE dent in Visual Studio.

Re:If you want a GREAT development environment...

[codepunk]

Why so they can end of life the product again and leave you holding the bag?

Fully cross platform open source object pascal IDE

http://lazarus.freepascal.org/

Re:If you want a GREAT development environment...

[El_Oscuro]

I think the sweetest IDE I have ever used was Borland C++ Builder. It combined the ease of VB GUI development with the power of C++. You could also compile direct to a single .EXE. No DLLs, JRE, Frameworks, or other nonsense. The entire Visual Component Library (VCL) that made up the GUI widgets was based on the ANSI string class, so you didn't have to muck around with string memory management yourself (no buffer overflows). You could also use all of the VCL string handling routines independently of the GUI. Borland even made it cross platform with Kylix for Linux, which included both Delphi and C++ Builder.

The problem is Borland basically abandoned it. The power of the VCL was very poorly documented, with few examples (at least it was there, no asking for a CD whenever you pressed F1). Then, after version 6, they stopped updating and dropped Kylix support altogether. Borland kicked Visual Studios butt in Windows development in every way possible until .NET, and didn't require a framework for programs to run. But Borland just let it wither on the vine...

Borland already made it cross platform with Kylix. I'm sure it would not be hard to make an OS-X version too. The C++ Builder 6/Kylix 3 (released over 5 years ago) are competitive with .NET now. All Borland/Codegear has to do is release them again with better documentation and support for important new technologies like XML/web services.

Re:If you want a GREAT development environment...

[Vintermann]

I'm stuck with an old version of Builder at work due to lock-in and porting issues. I would not touch that proprietary piece of evil with a ten foot pole.

Re:If you want a GREAT development environment...

[FlyingGuy]

Correction to my last.. CodeGEAR.com I think I am getting old.

Re:Jesus Christ, will someone please rip off ASP.N

[wzzzzrd]

try wicket. no xml, no navigation rules, not a single piece of code in your markup files (it's simply not possible), ALL logic is in the java files. no stupid bean mapping to forms, a component concept (oh, there i can download a tabbed panel component, let's do this) that actually works. it really is what i think MVC should be like.

and a very good api design, KISS, no overhead and all that core servlet stuff is hidden from you.

Re:13 years of hype

[Billly Gates]

Google maps on my cell phone written in 100% in java loads in 3 seconds.

Doesn't sound slow to me.

Troll?

[despisethesun]

How in the fuck was the parent post modded troll? It was a perfectly reasonable response to the GP.

Re:You mean...

[Anonymous Coward]

You are confusing "your" with "you're".

"Your" means you have possession of something. "You're" is a contraction for "you are". Note the apostrophe, and notice how the addition of the apostrophe allows you to lose a letter (the "a"). That's how contractions work: you substitute the apostrophe for the missing letter(s) and combine the two words.

Rather than assume you weren't paying attention in first grade I'm going to assume you don't have English as a first language and give you a pass this time. But it's actually an important thing to remember- contractions are easy and simple, and using them incorrectly makes you come across as an ignorant dumbfuck.

Re:Promising and Eclipse in the same sentence?

[Trillan]

(Hardly flamebait. An opinion outside what the majority thinks, apparently, but still not intentional flamebait. Go ahead and look through my history. I'll wait here.)

Eclipse is truly, truly terrible. It is hopelessly slow and buggy, and lacking in such basic features as to be useless. I'm absolutely shocked anyone could tolerate it, let alone call it "a breath of fresh air."

Open source is nice, but software needs to be judged on its own merits, rather than just the merits of its license. If the best defense you can come up with for software is that it's open source, you have completely failed to defend it.

Re:Promising and Eclipse in the same sentence?

[Trillan]

I didn't like the syntax highlighting, either. It worked great for my code, but seemed to completely fail to parse system headers. Maybe they've fixed that by now, though.

Re:Promising and Eclipse in the same sentence?

[Trillan]

Sure! Palm development environment is constructed with Eclipse. In a word, it sucks. But that isn't fair to sucking. I have never, ever seen a less capable development environment. Sure, the compilers suck - that isn't really Eclipse's fault, though. That's a plugin, right? The debugger also sucks. Oh, but that's a plugin too, right? The editor sucks, too. I think that's actually built in functionality, but I could be wrong. Qt's Eclipse-based tools? Also suck harshly. Symbian's? Suck harshly. Eclipse's own C Development Kit? Sucks harshly. Go ahead - I'm waiting. Tell me an Eclipsed-based product I can download that doesn't suck. I've just named at least three that you can download, one that's even included in the initial download.

Re:Promising and Eclipse in the same sentence?

[Trillan]

One last thing: If the only IDE you can compare Eclipse to is VIsual Studio, you've led a sad, sad life. No wonder Eclipse seems almost decent.

Re:Jesus Christ, will someone please rip off ASP.N

[jma05]

It's missing an important piece to be compared to ASP.NET. Proper IDE support. After all that is what this discussion is about (No, the Eclipse, Netbeans plugins don't hold a candle).

Re:13 years of hype

[kaffiene]

You are a moron. Java is not slow.

Do you SERIOUSLY think that Java has survived over a decade and become the most popular development language in the world because of _hype_? Do you think that eBay and Google use Java because of hype?

This kind of brain-dead me-too critique of Java wasn't accurate five years ago and it even less accurate now. If you want hype, go see Ruby. Java is continuing on regardless because it gets the job done.

Re:13 years of hype

[kaffiene]

Yeah, being the most popular development language in the world is a shit of a position to be in. How will they ever cope?

Re:You mean...

[Pojut]

"Your" means you have possession of something. "You're" is a contraction for "you are". Note the apostrophe, and notice how the addition of the apostrophe allows you to lose a letter (the "a"). That's how contractions work: you substitute the apostrophe for the missing letter(s) and combine the two words.

Notice how you put the period OUTSIDE of the parenthesis.

But it's actually an important thing to remember- contractions are easy and simple, and using them incorrectly makes you come across as an ignorant dumbfuck.

Being a grammar nazi when you yourself make such an elementary mistake makes you come accross as an ignorant dumbfuck.

Re:You mean...

[Pojut]

And apparently, I'm an ignorant dumbfuck because the entire sentence wasn't inside a parenthesis. /slaps self /backs away

Re:Jesus Christ, will someone please rip off ASP.N

[killjoe]

What are you talking about. Both eclipse and netbeans (and intellij IDEA) are vastly superior to Visual Studio.

Re:13 years of hype

[superpulpsicle]

You should take back your comment. Cause java IS slow once you start dealing with graphic library loading. If you use it to write ascii or computational programs then it is as fast as any other language.

Re:13 years of hype

[porkThreeWays]

It's disappointing that I got modded troll because it's acknowledged by most in the Java community that there were many big design mistakes in the first few versions of Java. Java 5 (aka 1.5) really was the first great version of Java.

The original GUI toolkit was admittedly thrown together for the sake of having a GUI toolkit. Swing is leaps and bounds better, but it's very confusing to beginners learning two GUI toolkits at the same time. If you didn't know the history of why there are two it's very confusing.

The original garbage collector sucked hardcore and was slow. The current garbage collector is actually pretty good, but for many they equate Java with being slow because of old versions.

Containers are leaps and bounds better and much more type-safe, but again it's confusing to beginners why there are so many redundant ways to use containers. There are numerous optimizations at the compiler level. The biggest being the ability compile code adaptively instead of the whole program on startup. I/O is confusing to learn and imo overly complex. Again, this is because of Java's subpar original I/O subsystem.

Java has really grown up and gotten leaps and bounds better over the years. Java today is what it should have been in the first place and what was originally advertised. That's where the marketing came in. Java honestly wasn't very impressive to me when it first gained attention. Today I'm very impressed by it. But most people don't understand how much Java has grown up and in their minds they have Java of 1999 stuck in their heads.

Re:You mean...

[Anonymous Coward]

The important question now is do they support Oracle back-ends and middle-ware. Some of us have to run with the adults you know.

Simplifying the Process (was Re:HuH?)

[jklappenbach]

While I agree that complex configuration files are a bane to development, I disagree with the assertion that J2EE requires them. ORM technologies like JPA are utilizing Java 5 annotations to declare configuration inline with code instead of XML. Frameworks like Wicket and GWT are providing developers with Java solutions to UI that are devoid of XML configuration, JSP, and markup-heavy implementations. IMHO, Wicket deserves to be called a breath of fresh air.

I do think it's a mistake for J2EE to include a particular view framework in its specification. JSF, while an innovation in the 90's, is simply a pig wearing lipstick compared to some of the new frameworks out there. Frameworks that, for example, are built on AJAX instead of including it as an afterthought.

I suppose the bewildering set of choices may be the root of the problem here. But if you make the effort to do your research, you'll find that many of your assumptions are incorrect.

I have one for you.

[Qbertino]

Emacs, maybe?
Honestly, if all that a purely Mac OS X oriented editor is gonna come up with is some obscure, yet-another-programm-specific configuration and automation language that I have to learn to automate and speed up my everyday tasks, then I might aswell use the CLI version of Emacs right away. Ok, the 20 basic commands are really bizar (Crtl+V == Page down; Alt+V == Page up, Ctrl+x (for 'eXecute') Ctrl+s == Save, etc... ) but when I then go on to learn automation via Lisp, at least then I know my programm is free and runs on anything that uses electricity. And it's not more difficult to use than TestMate (apart from the first 20 I mentioned). And the OS X Terminal Fonts look just as good as everything else on the Mac.

Frankly the best editor in existance (OSS or not) is jEdit. The only problem I have with it is that it can be a performance hog and it bogs down my 1Ghz iBook with a mere 512MB (I know, I should've gotten 1GB) to much when I'm running other stuff in parallel. Aside from that, jEdit is way beyond any other editor out there (features and ease-of-use), including TextMate. However, if I ever should get so far as to start automating my programming and editor functions I'm not gonna use BeanShell (jEdits Script) or some strange proprietary TextMate Script PL. I'd rather use lisp for that. And be sure it runs on every enviroment I'll ever encounter for the rest of my life.

Why people think that a proprietary 2006 Emacs clone is the cream of Editors just because it uses the neat looking Aqua tray is beyond me. Use Emacs if you like that sort of stuff. Its plattform independant and it's so insanely fast on our modern computers that you'll finally know how much crunching power those 2Ghz+ really have.

Re:I have one for you.

[Vintermann]

The ultimate difference between a text editor and and IDE as far as I'm concerned, is that a text editor doesn't really parse your code like a compiler does. Oh, it does a little parsing for syntax highlighting, and if you strap in some extra applications you can get some cross-referencing, documentation access and even a little autocompletion... but an IDE knows what type that variable is.

I tried a very-beta haskell plugin for Visual Studio. It was a real shock (I've just used editors for haskell before). When you've got a type system as powerful as Haskell's, and the IDE instantly tells you when you make a type mistake... In Haskell just about _all_ errors in your thinking will show up as type errors, so it's like having a guru by your side patiently explaining things to you.

That's when I switched from a small, fast editor (Nedit) to a slow, plugin-heavy one for regular editing (jEdit), and started making serious efforts towards groking Eclipse. I understand now what an IDE can potentially do for you.

Re:I have one for you.

[Anonymous Coward]

You are confused. The Visual Studio editor is working just like Emacs. It is the compiler that is really telling you these things. The real distinction you are seeing as editor vs. IDE is reallyi in the tools(compilers, debuggers, etc.) that are available and the amount of effort put in to integrating them with the editors. Visual Haskell was developed in part by one of the GHC developers, and part of its development included modifications to GHC. This is not a coincidence. haskell-mode for Emacs doesn't do as much as Visual Haskell does at this point, but there is no reason that it can't be made so.

Re:Jesus Christ, will someone please rip off ASP.N

[nuzak]

Sounds like you're looking for Seam. All the xml stuff is optional in Seam, so you'll actually come to like the xml configs for its good parts (like complex navigation rules) when you're not forced to use it for everything. No declaration needed, you don't even need "backing beans" or even a class for your page (take that, ASP.NET).

If you really want tight integration between page and code, there's wicket (which I frankly find awful) or Tapestry5 (which looks really nice, but it's not finished yet). Or if you prefer to just code guis straight to the web without all the multitier nonsense, there's no shortage of frameworks to choose from there, including GWT, Echo2, and Thinwire. There's even a few that will publish swing apps straight to the web, so you can design your page in Matisse. I don't personally recommend that approach, but it's there if you want it.

Re:Jesus Christ, will someone please rip off ASP.N

[jma05]

I was talking about the Wicket integration into these IDEs. ASP.NET is well integrated into VS. Wicket integration through plugins is relatively thin. Don't get me wrong. I have not used VS in 3-4 years although I have a copy and have been using Netbeans and Eclipse.

Re:13 years of hype

[jma05]

Let me put it this way. Java IDEs are the ONLY programs that make me wish I have a newer computer.

Re:13 years of hype

[jma05]

Java is now good because of Metcalfe's law. It has reached critical mass. Like you, I hated Java when it first came. It was the wrong solution for most things it was initially used for. Now things are beginning to look better. One thing that still bothers me is the mindset of the libraries and frameworks available. Many seem to fail to curb complexity. Either that or a person like me is not what they had in mind when they designed these.

It is often far simpler to use a dynamic language to cook something up very agile on the server side than resort to bloated app servers, xml hell, design pattern hell etc. Java, after 13 years still makes designing a client UI with some DB components more complicated and far less productive than it was with Delphi 1 or VB 3. And where are the third party components? I am pretty sure there are far far fewer Swing components now than Delphi with its modest user base did in 2000 (about 4000 OSS and 4000 commercial components in 2000 if I recall right).

Java will eventually get it all right but why is taking so long?

Re:Promising and Eclipse in the same sentence?

[mjsottile77]

I'm not a big fan of eclipse (although I've tried it over, and over, and over, hoping one day it would make me "see the light" with some new, nice version). I wouldn't say it's popularity is due to a license though. Eclipse has momentum due to it being hyped as being representative of "industry best practices and standards". Some people are quite happy to use that as a metric for deciding to adopt it, regardless of whether those "best practices" apply to their particular programming problem or industry, or if the tool is actually technically superior to others available for the problem domain. This is a pervasive attitude in computing - there is an unusual and hard to explain aversion to having a diversity of ways to solve a problem. It's very much a "all you have is a hammer, so everything looks like a nail" mentality to me. I say use the right tool for the job, be it a language, framework, programming pattern, or editor. Choosing a "standard" because it's all you know or are comfortable with, and contorting it to do something it wasn't designed for is one of the contributing factors to all of the bad software that exists in the world.

Re:Jesus Christ, will someone please rip off ASP.N

[killjoe]

I don't know about wicket per se but I bet it has an eclipse plug in. Every framework known to mankind has an eclipse plug in.

Re:13 years of hype

[kaffiene]

"Graphic library loading"????

Do you have two whole brain cells to rub together?

I can tell you that OpenGL and hardware accelerated 2d graphics work quite happily within Java.

Re:Promising and Eclipse in the same sentence?

[juanignaciosl]

Speaking about Visual Studio was not a matter of quality but of integration. Until Netbeans 5.5 I had never seen before a "all in one" open source package which let me develop a new project from scratch and deploy it in an application server. You use to have to install many different apps and libraries until it works. Nevertheless I wanted to avoid "direct confrontation" between NB and RHDS, since it's what people was doing for sure (just read comments here). Mentioning VS was neither comparison nor quality but an example of integration.

Re:Promising and Eclipse in the same sentence?

[Trillan]

In theory, it should be possible to integrate Visual Studio with nearly anything. In practice, though, it's another one of those things that always seems to suck. Having never tried to write a plugin for it, I can't tell you if it is due to the complexity and/or flakiness of Visual Studio or just incompetence on the plugin developers. I suspect it's mostly Microsoft's fault...

Meh. I can't even think of a good IDE right now. I like a lot of the way Xcode works, but not all of it. Still, I thought Eclipse as a more-weak-than-average entry overall.

Re:Promising and Eclipse in the same sentence?

[juanignaciosl]

It's not the _connectivity_ what I whas talking about, but "packaging". VS is the best (and almost only) option if you use MS-only technologies, and it allows you to begin with them with little effort. That's what I think's worthy with RHDS: JBoss + RichFaces + Seam + JPA made easy.

IMHO, of course

Re:Promising and Eclipse in the same sentence?

[Trillan]

Oh, certainly. I think VS is the best tool for Windows development, precisely because it's made by Microsoft. I honestly wouldn't use anything else. I hate the UI and a lot of the features, but there's little argument that it gets the job done in a way that only the OS developer could probably pull off.

Apple's Xcode is similar on the Mac. There's a few things I like more than Visual Studio, and a few things that I like less. Overall, I think it's just about the same, with maybe a slight edge on design (which you'd expect when comparing an Apple product to a Microsoft one). It isn't as strong as it should be, though.

And yes, I see the same potential in RHDS. I know it's going to be well integrated. I just wish it was also good interface design, if you follow me. :)

And all of them have their strengths as IDEs, but it bothers me that they're all so weak. It shouldn't be hard to build an IDE that doesn't suck. I don't know what it says that programmers can't design good UIs for programmer tools.

Re:Promising and Eclipse in the same sentence?

[juanignaciosl]

Well, I don't agree "I don't know what it says that programmers can't design good UIs for programmer tools.". Eclipse can be better, but it's not bad IMHO.

Re:Jesus Christ, will someone please rip off ASP.N

[jma05]

Read my post again. I am saying that Wicket has plugins (for all 3 major IDEs in fact) and that they are nowhere as integrated as ASP.NET natively is with VS. Having a plugin says nothing. A plugin may do little more than add a couple of config files and add a build task to the project. That cannot equate with polished products.

Re:Jesus Christ, will someone please rip off ASP.N

[killjoe]

>I am saying that Wicket has plugins (for all 3 major IDEs in fact) and that they are nowhere as integrated as ASP.NET natively is with VS.

What the hell does that mean?

What specifically is your gripe about the wicket plug in?

Re:13 years of hype

[amightywind]

Wow just 3 seconds! On a processor running at 600MHz. You Java types have been so trained to low expectations you don't know what speed is.

RH Dev Studio

[Anonymous Coward]

This cheapens the concept of "development". I guess Redhat is doomed only to be a webapp server and nothing more. Its time we get over java and move on.

Re:13 years of hype

[Anonymous Coward]

Yes, really slow.

Re:Jesus Christ, will someone please rip off ASP.N

[chochos]

I checked the HelloWorld example, it look very similar to Tapestry (another Apache project). Two very similar projects under the same roof... no wonder web development is such a mess. Granted, Tapestry is more complex than wicket (at first sight; I'd have to look deeper). Looks like in wicket they got rid of the intermediate .page of .jwc files (XML files with the component definitions for the stuff in the HTML files; the page/jwc files is where you define the connections between UI components and the control classes). Worth checking out, this wicket. If it's as powerful as Tapestry and is really simpler, I think we might switch. We have some medium to large size apps in Tapestry.

Re:Jesus Christ, will someone please rip off ASP.N

[jma05]

> What specifically is your gripe about the wicket plug in?

What gripe? I don't have any. My point is that ASP.NET is not comparable to Wicket. It's like comparing Delphi to Eclipse with a C++ plugin. Both will do the job. The development experience is nowhere similar.

1. Can you drag and drop widgets into a WYSIWYG page designer and set properties in a Wicket plugin?
2. Can you find (free or commercial) the same spectrum of third party components that will also show up in your workbench?
3. Can you visually compose compound components?
4. Does wicket support data binding like ASP.NET as well as have visual tools for the same?

Re:Jesus Christ, will someone please rip off ASP.N

[pkphilip]

Try click.sourceforge.net. Seriously. No nonsense, just a couple of jar files, no messing around with a hundred xml files, no stupid theories about IoC, Dependency Injection etc.. All pretty simple.