Slashdot Mirror

Eclipse and NetBeans are both written in Java, albeit with different GUI toolkits. Eclipse uses SWT, while NetBeans uses Swing.

You'd think that Eclipse would be faster due to SWT interfacing with the native widget set, but I guess this shows that GUI speed isn't everything. As a side note, Azureus also uses SWT.

Don't get me wrong - I like the idea, and think it could work really well for the environments that will support Firefox across its (potentially unlimited) lifetime, but it seems to greatly overlap with javascript/flash. Each is practically limited to either specialist uses, or else work as least-common-denominator products for the potential Firefox environments. That means a lot of sprite games with simple interactions, graph and UI effects in popularized widgets, web portal software, and even the occasional spyware exploit finding a way to mark a user's trail. There will be ports of simple software from other environments, but limited interaction with the outside environment (by design), being chained to a time-limited browser session, and lack of the easy ability to really exploit the running environment will severely limit what toys and tools can really be created.

That's why I've taken a liking to Eclipse recently - it takes a nice set of the fast-development architecture of java development, and allows them to be used by C/C++, Python, and others cross-platform. Has anyone started working on a really nice integration of Eclipse into a Firefox plugin?

Ryan Fenton

I heard Oxygen is also popular among breathers.

Well, the reason why they want to wait for approval is because some big names are participating such as IBM and Novell. I really don't think they're too interested in a patent battle with Microsoft when Microsoft seems to have the upper hand. IBM, for one, does not like to lose in court.

What's the difference between the Higgins project and OpenID?

So what happened to OurEclipse? Oh wait, I guess it's still alive and well :)

http://www.eclipse.org/ Eclipse is an Open Source development framework. Or are you commenting on the format of the link?

I agree that there are serious issues with the whole AWT-Swing debacle. But I don't personally believe that is what killed off desktop Java. I think the original GPL-incompatible Java license killed off Linux desktop Java development and Microsoft (probably correctly from maintaining a monopoly perspective) killed off Windows desktop Java by dropping it and making .NET.

There was hardly much point in Sun fixing all the issues with desktop application development given that desktop Java was hardly deployed anywhere.

Also it suffers from the fact that the Operating System vendors always want to make the latest greatest proprietary OS features available to developers and Java's guaranteed cross-platform-ness prevents them from doing this through Java in a well integrated way.

With regard to ActiveX in desktop Java, have you tried SWT rather than Swing? http://www.eclipse.org/articles/Article-ActivexSup portInSwt/index.html

Java technology and library development may have been steered towards web-oriented selling points, but the language itself isn't inherently oriented towards helping web developers and the like. Specialist scripting languages can always be developed to make specialist tasks faster - and Java (the language)is far too purely object oriented to be as specialist-efficient as some of the less object-oriented languages, without really stretching things.

In fact, my favorite uses of Java (the language) aren't web-apps at all, they're applications like Azureus, and Eclipse. That's perhaps what Java (the language) is really best at so far from my perspective - cross-platform development of portable frameworks. It's because of that, that Java (the language) has a stronger future than Java (the technology), as a strongly object-oriented language developed to be portable.

Ryan Fenton

How could I forget to mention http://www.openoffice.org/ which is a great office productivity suite and http://www.eclipse.org/ which is great developer IDE suite? All OSS, of course.

As I replied for the previous Netscape RSS DTD article http://slashdot.org/comments.pl?sid=216818&cid=176 03480, caching DTDs from the network is not the answer if there is the possibility they will not be there in the future:

The proper thing to do is for your application to use an XML catalog for resolving entities/URIs and bundle the DTD files with the application. There is a good article at http://xml.apache.org/commons/components/resolver/ resolver-article.html that helped me out. In addition, if you are using Eclipse with the web tools platform, you can customize the catalog so it resolves DTDs and entities locally. See http://wiki.eclipse.org/index.php/Using_the_XML_Ca talog.

You are right. I wish I would have seen this article earlier so that I could have posted sooner -- and others to get to see the "solution"!

Ever since I started developing on a laptop during my commute, I discovered that XML-based programs like J2EE servers would simply stop working. I experienced the same thing at work where, by default, your desktop applications (namely Eclipse) do not have access to the internet, and the servers will never have access to the "Internet".

The proper thing to do is for your application to use an XML catalog for resolving entities/URIs. There is a good article at http://xml.apache.org/commons/components/resolver/ resolver-article.html that helped me out. In addition, if you are using Eclipse with the web tools platform, you can customize the catalog so it resolves DTDs and entities locally. See http://wiki.eclipse.org/index.php/Using_the_XML_Ca talog.

I've found that Mylar makes me extremely productive. I've tried task lists and post-it notes before, both real and virtual. With Mylar embedded in Eclipse, I have my tasks in the same application as my code, so they don't get lost or require a context switch to view. I'm using it in conjunction with Jira (it works with Bugzilla and Trac just as well or better), and I keep my bigger tasks organized by future release in Jira. I can see them in Mylar if I want to, but they're usually folded up and out of the way, except for the release I'm currently working on.

When I start to work on a large Jira feature, I create a bunch of small local tasks in Mylar that will only take an hour or two to complete, and I schedule them. It only takes a couple of seconds. If I get into the middle of a task or refactoring, and I find something else that needs to be done, I just create another local task if it's small or Jira task if it's big.

This has nearly eliminated all my procrastinating. A lot of why I procrastinated is because my releases are large, and I could never create enough post-it notes or tasks in any other system to make the releases manageable. Breaking the tasks down also makes them more fun because they turn into small annoyances that I want to get rid of. I come back from lunch thinking that I can get two or three done before I go home instead of wondering where I left off on a three month release.

There are a couple great side effects of this system. It's really easy to create release notes from the Jira tasks, and it's easy to create status reports from the local tasks in Mylar. I never forget anything. I also find it easier to go home at the end of the day, since I can create a task that will let me know right where I'm leaving off, and I make sure to activate the task, clear out all the other editors, and leave the task open. I write down everything I'm thinking at the end of the day, and I'm less likely to lie awake at night thinking about what I still have to do.

I've found that Mylar makes me extremely productive. I've tried task lists and post-it notes before, both real and virtual. With Mylar embedded in Eclipse, I have my tasks in the same application as my code, so they don't get lost or require a context switch to view. I'm using it in conjunction with Jira (it works with Bugzilla and Trac just as well or better), and I keep my bigger tasks organized by future release in Jira. I can see them in Mylar if I want to, but they're usually folded up and out of the way, except for the release I'm currently working on.

When I start to work on a large Jira feature, I create a bunch of small local tasks in Mylar that will only take an hour or two to complete, and I schedule them. It only takes a couple of seconds. If I get into the middle of a task or refactoring, and I find something else that needs to be done, I just create another local task if it's small or Jira task if it's big.

This has nearly eliminated all my procrastinating. A lot of why I procrastinated is because my releases are large, and I could never create enough post-it notes or tasks in any other system to make the releases manageable. Breaking the tasks down also makes them more fun because they turn into small annoyances that I want to get rid of. I come back from lunch thinking that I can get two or three done before I go home instead of wondering where I left off on a three month release.

There are a couple great side effects of this system. It's really easy to create release notes from the Jira tasks, and it's easy to create status reports from the local tasks in Mylar. I never forget anything. I also find it easier to go home at the end of the day, since I can create a task that will let me know right where I'm leaving off, and I make sure to activate the task, clear out all the other editors, and leave the task open. I write down everything I'm thinking at the end of the day, and I'm less likely to lie awake at night thinking about what I still have to do.

The first Linux distro I started with was Slackware 9.0. It took me 3 weeks to get it up & running, with the help of a lot of people from LinuxQuestions.org; But as they say: "Once a slacker, always a slacker"

I tried other distros, but none give me the freedom to manipulate my system like Slackware.

In short: If you like (not mind, but like) reading manuals and changelogs before using a new utility and program, then you'll be able to pull it off with Slackware. Otherwise, go with Kubuntu (It's Ubuntu but with KDE as its Desktop Environment instead of GNOME).

The Fedora Core project has a life cycle of 6 months and is a test ground for RedHat; What ever is new, is thrown there for tests; Once stable, it's moved to the Enterprise & Advanced server editions (not free).
And it's really sluggish and you'll face some weird conflicts with some applications.

I recommend KUbuntu, as a free distro. If you want a commercial one where support is ready for you whenever you want, go with Suze, Mandrake or Linspire.

If you'd like to work with both Linux & Windows at the same time, without rebooting each time, consider using Windows as a virtual machine (VMware).

===========

For your all-in-one needs for programming & web-design, I strongly suggest Eclipse. You can use it for Java, PHP, C++, HTML, CSS, ...etc.

Good luck.

The other advantage is that Java and C++ can be had for free with nice helpful IDEs. The Eclipse CDT (http://www.eclipse.org/ ) with cygwin (http://www.cygwin.com/) provides a fairly full IDE (and you can tailor it's key interface to EMACS :-) ). Eclipse also looks to be a fair Java IDE and there is also netbeans (http://www.netbeans.org/).

As a book I would suggest one I have enjoyed using in its online form, Thinking in C++ by Bruce Eckel (http://www.mindview.net/Books/TICPP/ThinkingInCPP 2e.html). It has the advantage that there is also a second volume that takes you into using more advanced features (fancy templating, STL, and some more complicated patterns) that they can look into or you could use as a follow on.

So, you're saying the user experience is similar to Eclipse, but lots more people will get to see it! Woo hoo!

(I grabbed NeoOffice last week when the zero-day Office exploit was reported, but haven't needed to use a word processor et al since then. My iBook is going to sweat when I load it up, but I've used Eclipse on it for years...)

The Java IDE used to develop Java applications? Sorry, it doesn't count; it's just a Java tool to make more Java. It doesn't make Java useful.

Eclipse projects provide tools and frameworks that span the entire software development lifecycle, including modeling, development, deployment tools, reporting, data manipulation, testing and profiling. The tools and frameworks are primarily focused on building JEE, web services and web applications. Eclipse also provides support for other languages, such as C/C++, PHP, and others.

First a disclaimer that I work for IBM. Similarly, my comments don't necessarily reflect that of my employer, etc, etc, etc...

http://www-306.ibm.com/software/awdtools/developer /application/

I use it for all the coding work I do. It has a GUI building tool that only generates Java code. Its much better than the old visual age line of products that required separate files to manage the GUI construction. You can even edit most of the generated code manually without fear of no longer being able to use the GUI builder. It uses (Eclipse) as its base platform so other eclipse plugins can be added to the tool. It includes extensive tools for other web related development as well (DB definitions, EJB work, XML, etc...)

Granted, it's listed as ~$2k for 1 license, so its not for everyone, but you'd be hard pressed to find a java/web development function that it doesn't help you with.

Dan

There's Netbeans which has a GUI builder. (Which, IMHO, sucks. Or at least did the last time I used it.)

Eclipse also has plugins (the Visual Editor) to allow creating GUIs graphically. (Which is, IMHO, worse than Netbeans's.)

So, there are GUI building tools. They just suck. (IMHO, of course.)

Personally I'd like to see something like XUL where the GUI is specified using a markup language. Which is why I started writing one and then never released it.

You could do a lot worse than to check out SWT / JFace / Eclipse RCP Framework.

If you haven't heard of these things:
* SWT is a relatively low-level cross-platform widget toolkit. It is at around the same level as Qt, GTK+, and the GUI bits of MFC.
* JFace adds a layer of abstraction on top of and is built to work with SWT...helping you effectively split your widgets and the data feeding them in nice ways. It also provides a whole lot of convenience classes for doing common things more easily.
* Eclipse RCP (Rich Client Platform) ties all the above together in a huge framework for building any sort of GUI application. This is where you will learn the most about design pattern applications to GUI projects. Eclipse itself is built using this framework.

The learning curve can be a bit steep if just starting, but I haven't come across anything so well put together before. It's full of lots of good ideas.

Even if you choose not to use it (or can't use for good reasons) you will learn a great deal about how to put together a GUI application and how to apply certain patterns and techniques effectively. A few well known "gurus" have been involved in the design and you can be sure that a lot of effort has been put into the architecture and deciding what works and what doesn't in terms of patterns and techniques.

This book is a really great way to get into it all:

http://www.amazon.com/Eclipse-Rich-Client-Platform -Applications/dp/0321334612/sr=1-1/qid=1164595827/ ref=pd_bbs_1/103-2317093-7394243?ie=UTF8&s=books

And for more info ...

http://wiki.eclipse.org/index.php/Rich_Client_Plat form

So... people should be given a choice when it comes to the question of how much choice they should get.

I looked at a couple of the popular ones, installed Awffull and played with it for a bit. But it wasn't immediately
obvious to me that any of the common ones supported aggregating stats across domains / hosts. Eg, I have 10 virtual servers on this
Apache box, give me a sorted list of hits per domain/host. Probably one or more of the popular open-source stats packages
*does* do this, but I didn't feel like spending hours examining different ones and installing them. Since my needs were very basic
I just wrote something of my own.

Since all my domains are ultimately served by a Java webapp running on JBoss (I redirect from Apache to JBoss with mod_jk) I just wrote a servlet filter to write hits to a postgresql database. That's it,one table with the hostname, date-time, user-agent, and a handful of other things I care about. Now, getting the info I need is a simple as a quick sql query with pgadmin III. Although I'm looking at using the Eclipse BIRT stuff for looking at the data, as my next project.

I looked at a couple of the popular ones, installed Awffull and played with it for a bit. But it wasn't immediately
obvious to me that any of the common ones supported aggregating stats across domains / hosts. Eg, I have 10 virtual servers on this
Apache box, give me a sorted list of hits per domain/host. Probably one or more of the popular open-source stats packages
*does* do this, but I didn't feel like spending hours examining different ones and installing them. Since my needs were very basic
I just wrote something of my own.

Since all my domains are ultimately served by a Java webapp running on JBoss (I redirect from Apache to JBoss with mod_jk) I just wrote a servlet filter to write hits to a postgresql database. That's it,one table with the hostname, date-time, user-agent, and a handful of other things I care about. Now, getting the info I need is a simple as a quick sql query with pgadmin III. Although I'm looking at using the Eclipse BIRT stuff for looking at the data, as my next project.

One word:
Eclipse

Eclipse: http://eclipse.org/

Crystal Reports

I think this is good news for Linux desktop. Java isn't that awful language for desktop stuff any more, especially thanks to things like SWT. Heck, even Swing appears to have become really good over time. SWT apps are generally indistinguishable from C++ apps in terms of responsiveness, the only problem seems to be the slightly ridiculous memory use. But hey, Eclipse is the new Emacs - Eight Megabytes that are Constantly Swapping is nothing these days, so we just need something else to joke about. =)

I also think this will be great in light of the great fears about Mono's licencing and Microsoft's suspected foul play. If Sun dumps us a time-tested GPL'd cross-platform application development kit that basically does everything Mono does (with slightly more verbosity), heck, I'm all for it. And Sun backs GNOME too - if Novell and Mono stuff go to oblivion, we at least have a stable, time-tested similar platform to look to.

write gorgeous apps in C# that everyone wants, encourage competing distros to integrate those apps, then laugh as Microsoft takes out their competition in court. And coincidentally distract developer and user attention from the steadily emerging Free Java which is swimming along nicely with the addition of SWT on to the base of GCJ with the recently merged GNU Classpath libraries. And coincidentally helps to put extra pressure on Red Hat after their JBOSS acquisition. I think we're looking at a convergence of multiple interests antagonistic to Red Hat here and Microsoft is going to eat up the Linux distributors one at a time. Right now Red Hat is the biggest threat and so Oracle and Microsoft want to acquire or destroy. Either will do fine.

There is an excellent IDE for PHP. It's called PHP Eclipse, which is a plug in for Eclipse.

There is a very good reason to install Firefox via your distribution's package management system. If another packages relies on the Gecko libraries to embed a web browser (examples: Eclipse, Listen), and you installed Firefox manually, you'll wind up with two copies installed.

Actually, on Ubuntu you'll probably wind up with both Mozilla and Firefox installed (grr), but you see my point.

the Linux platform needs ... A set of standardized GUI libraries that are free for both commercial and open-source applications.

You mispelled "proprietary" there. It's cool though, lots of people do it.

Also I'm not really sure why you think the Linux kernel needs widget toolkits.

Anyway, yeah there's totally no toolkits that allow proprietary apps.

Wish I'd looked at the Eclipse documentation more closely before posting... there's some fun gotchas buried in the EclipseCon 2006 CDT PowerPoint Presentation available on the Wiki. Here's some delightful reminders of what's wrong with the CDT:

"Turn off Build Automatically" -- In Eclipse/Java, you'd never need to tell someone to do this, even in the largest of projects, because the build runs quickly and incrementally (using the built-in Eclipse Java compiler). But in CDT, the only way to build is to run your entire toolchain using a Makefile. (So instead of fixing this, they provide features to auto-generate the Makefile!)

"The CDT full indexer is very expensive on large C++ projects (Recommendation: Don't use it on such projects)" Gee, thanks! That's the thing that makes Eclipse (in Java) so Eclipse-y, you know? So make sure you turn that off on large projects.

Oh, and there's my personal favorite FAQ: Can I debug Java and C++ at the same time? Answer? "If you can get this to work, please let the cdt-dev mailing list know!"

The Eclipse CDT is a joke. Even Visual Studio can handle reference searches on large projects.

Wish I'd looked at the Eclipse documentation more closely before posting... there's some fun gotchas buried in the EclipseCon 2006 CDT PowerPoint Presentation available on the Wiki. Here's some delightful reminders of what's wrong with the CDT:

"Turn off Build Automatically" -- In Eclipse/Java, you'd never need to tell someone to do this, even in the largest of projects, because the build runs quickly and incrementally (using the built-in Eclipse Java compiler). But in CDT, the only way to build is to run your entire toolchain using a Makefile. (So instead of fixing this, they provide features to auto-generate the Makefile!)

"The CDT full indexer is very expensive on large C++ projects (Recommendation: Don't use it on such projects)" Gee, thanks! That's the thing that makes Eclipse (in Java) so Eclipse-y, you know? So make sure you turn that off on large projects.

Oh, and there's my personal favorite FAQ: Can I debug Java and C++ at the same time? Answer? "If you can get this to work, please let the cdt-dev mailing list know!"

The Eclipse CDT is a joke. Even Visual Studio can handle reference searches on large projects.

A personal fave is that the debugger integration in eclipse is second to none.

|Oh, really?

Both Eclipse and Derby are the result of previous shopping sprees by IBM.

Eclipse was developed by the IBM Ottawa Software Lab. This lab started life as OTI, a company which developed Smalltalk technology, that IBM bought in 1996.

Derby is the open-source version of the Cloudscape DB. Cloudscape was a Java DB company which was acquired by Informix in 1999, which was in turn acquired by IBM in 2005.

And seeing what the products look like when IBM builds them in-house, we are probably better off that way.

Oh, really?

Don't get me wrong. I love the Borland of yester-year, but are Borland tools relevant anymore?

With all the free goodies available, development on most platforms can be done without spending a dime.
Just off the top of my head, things like GCC, Xcode, and Eclipse come to mind.

Would you like to give some modern day examples, or is this just the "Java - write once, test everywhere"

Most likely, the companies that are developing cell phones (and not using Linux as their underlying OS) are ones that have been in the market for many years and were doing so back when Linux was not available as a platform for this purpose. That means they bought into an expensive toolchain (operating system, compiler, debugging tools, source code editor, repository if any, system libraries, etc.) and have sunk costs as well as developer inertia to moving to a new platform. As long as they figure in the royalties (if any) as part of that cost they can decide whether staying with that platform makes sense. They may also feel that the requirement to include source code (which might give away what they consider proprietary information about the internals of their system which might give competitors an advantage) may not be appropriate and thus prefer a licensing system that does not require them to do so. This is why some companies pay for licenses from open-source developers in order to get the "you must relase too" requirement of the GPL to be waived.

In switching to another platform (like linux) they also have to figure in training costs (whether or not they are willing to spend money to train people which they then figure will run to another company) and the costs in lost productivity as programmers learn how to use the new toolchain and platform, and/or new APIs for programming the new operating system. Availability of ancillary tools can also be important (is there a PC-based emulator for their phone (or embedded device) available for Linux? There may be one available, included or previously developed with the toolchain they are using.) Any time you change technologies there is a learning curve unless the new and old systems are identical, and chances are they are not. If the new system is considerably better, then the learning curve could be small and there would be an increase in productivity. If it is not better, the learning curve could be steep and there can be a significant permanent decrease or even loss of productivity.

And to put it bluntly, until Eclipse came along, the toolchain for Linux basically sucked with the possible exception of Borland's Kylix, but since most software for Linux is C/C++ and not Pascal, (and quite potentially for very small embedded devices, assembly) that doesn't help much. Having been a programmer for over 20 years, and seeing the difference with tools like Visual Basic and Turbo Pascal for Windows/Delphi, let me tell you there is a big improvement in usability and in productivity over writing Fortran using punch cards on an IBM 370 equivalent mainframe (which tells you how far the technology has changed, at least during the period I've used computers, and the changes (and improvements) are coming even faster). Using text editors alone to develop software (unless you're developing for a text-only environment) to be used in a graphical interface environment is a big pain. Especially once you've experienced the difference. In fact, even if you are developing for a text-only environment, a number of the features of these Rapid Application Development systems can b

There is an upcoming fully supported php ide coming to the eclipse platform.

There is also a eclipse plugin for php development, but i haven't tested it.

I want a fast editor with basic project management, API integration and code completion.

Also, here's how to manually create a Mac OS X application bundle from a Java program. The example is for an SWT program, but there would be no changes needed for a Swing app. In the shell script, just make sure that you add all necessary Mac-specific command line options.

http://www.eclipse.org/swt/macosx/

I can't believe that the article didn't recommend Eclipse. Even Apple's Developer Connection recommends this wonderful program.

Of course, Eclipse is a good tool because it is multiplatform and highly extensible. I find it great for java and python (through pydev) work on my Mac and other boxen.

First, they have X11 listed incorrectly as a PPC app. If you have an Intel based Mac then the installation CDs includes a pkg of X11 as a Universal app.

Once you have X11 installed then you can also get ...

1) Inkscape (Universal) ... for all of your SVG editing needs.

2) Gimp (Universal) ... for all of your raster image editing needs.

... and I am surprised that they left Eclipse off of the list. While Xcode is preferred for doing OS X development, it is still worth a mention :-)

I'm sure there are others, but these are the ones I use most.

There's no real world, used by millions people, example, of a successful *big* app written in Java/C#/Python/Ruby or any other hyped thing.

There certainly are; here's one I use daily.

But judging by your last sentence, you must be trolling.

+1 for subversive

Another important point - Polarion (the authors) recently submitted a proposal to make subversive a standard project under eclipse.org

http://www.eclipse.org/proposals/subversive/

My manager wants us to adopt a Windows only Eclipse IDE

Eclipse is multi-platform, with versions available for Windows, Linux and OS X.

The new version ain't native, though. From the summary, they mention:

It's going to be based mainly on the Eclipse framework

Eclipse is IBM's Java IDE. Java means you need a JVM. Eclipse requires Sun's JRE, which means x86 Linux only.

Plus, because it's "based on Eclipse" that means it'll use GTK+'s controls, but most likely with Windows' HCI. This means things like the default dialog button being on the wrong side (it's supposed to be on the right, Windows incorrectly places it on the left so Eclipse does too) and various other GUI issues as SWT, the GUI toolkit Eclipse uses, is optimized for Windows. It also means the Mac OS X "port" is going to be terrible, as SWT support for OS X is really lousy.

Java is notorius for being "write once, test everywhere" so this new "Linux" version can be expected to have some interesting Windows-style quirks. And be slllloooooow - first you have the Java speed issues, compounded with Eclipse's natural slowness. Java may be 90% as fast as native code, but Eclipse brings that down to maybe 50% as fast.

From what I hear about Notes, though, this may be about normal for Notes performance...

Eclipse's java requires sun's jvm which conflicts with gjc. Open Office requires gjc in linux for 100% functionality, sun's jvm won't cut it.

I've used the Eclipse Test & Performance Tools Platform with pretty good results:

        http://www.eclipse.org/articles/Article-TPTP-Profi ling-Tool/tptpProfilingArticle.html