Discussions

News: Opinion: Is Eclipse Poised to Move into Quiet Ubiquity?

  1. On November 7th, various cities around the globe (including my hometown of Dallas, hosted by MyEclipse) threw birthday parties for the Eclipse platform, which celebrated the 5th anniversary since the code was made available by IBM. On holidays, it tends to be tradition to do a little self examining; reflection upon the past and pondering of the future. Eclipse is no different, and while most of us know how far Eclipse has come, where is it going? What will the next five years hold? What will the 10th birthday of Eclipse look like? With that in mind, let's explore what I feel are the highlights in the thus far successful life of Eclipse, and where Eclipse may be going: Inception to 3 Years Old As a new and innovative player, Eclipse must prove itself. Everyone wanted their tools to be free, but developed a need for support of those free tools. This spawned the early commercial offerings (including MyEclipse). It also promoted a lot of firefights between those who saw the potential of Eclipse ecosystem and the naysayers who said it wouldn't last and that Eclipse-based commercial offerings were doomed. The downside of free tools at mass adoption begins to rear its ugly head: consumers begin to resort to free, do-it-yourself strategies and end up spending time capital and hard cash to cobble together their own solutions. This fuels growth of low-cost, supported solutions. Eclipse experiences 300% growth. Four Years Old Eclipse has expanded to commoditize every JEE spec. With the advent of WebTools, many other IDEs throw in the towel and base themselves off of Eclipse (Microsoft, IntelliJ and Sun's NetBeans remain the only significant bastions). Enterprise adoption of Eclipse grows rapidly as Eclipse-based tooling becomes a standard, not an exception. The enterprise is now looking for a value-based proposition, not necessarily a free solution, though many continue to utilize the base Eclipse JDK. Eclipse as an RCP emerges, with projects beginning in multiple verticals. Eclipse experiences 100% growth. 5 Years Old (Present Day) The advent of Callisto and Eclipse project release alignment begins to further commoditize the market, resulting in fewer full-featured IDE players in the market. Cross-platform functionality emerges, as Genuitec incorporates NetBeans features into the Eclipse environment. RCP emerges as the standard for rich development. Sun scales back Java developer tools and begins cannibalizing features into the NetBeans platform. Sun also begins to open source portions of Java. Eclipse experiences 100% growth. Years 6-10 – Looking Ahead Eclipse continues to grow at a phenomenal rate per year to the point of being a given in the tech sector rather than a newsworthy item. Blended solutions of open standards and commercial IP take hold and relegate the do-it-yourself enthusiasts to a hobbyist level. RCP has grown into a standard base for rich applications. By year 10, Eclipse has quietly plunged into ubiquity, with a saturated market share that dominates the tools market. Other IDEs become small but vocal niches. So what do you think the 10th birthday of Eclipse will look like? It depends... If growth continues at this rate, Eclipse could become a household name by year 10. The catchphrase (borrowed from our friends at Intel) "Eclipse Inside" may make Eclipse notorious in brand, but simultaneously unexciting due to broad adoption in multiple verticals. We'll have to see, but no matter the exact positioning, the growth will be exciting.

    Threaded Messages (58)

  2. Good luck to Eclipse, but this article seems to me to present a rather arrogant attitude (of course, that does not mean it is wrong!)
    (Microsoft, IntelliJ and Sun's NetBeans remain the only significant bastions). Enterprise adoption of Eclipse grows rapidly as Eclipse-based tooling becomes a standard, not an exception.
    There is a broad range between 'exception' and 'standard', and to claim that Microsoft, IntelliJ and NetBeans remain 'the only significant bastions' is a rather skewed view of things. For example, NetBeans shows no sign of decline, indeed it is gathering more support. I see a real problem with one IDE becoming truly dominant (a status Eclipse has not yet reached), in that it gives too much control over how Java is used to the IDE. I have seen this happen in a group that used Eclipse - they held back on adoption of Java 5 because Eclipse did not support it. (Fortunately, I had the choice of using NetBeans, which had supported Java 5 for some time at that point). Imagine that kind of decision-making industry-wide. Eclipse has led the way in terms of some IDE features (refactoring, for example), but in other ways it has been far behind, relying on plugins for what many of us think should be core features (JEE support, GUI design). What we need is choice. To get that choice we want more than one IDE; instead, there should be better mechanisms for transferring projects and plug-ins between IDEs - we really need JSR 198. My wish for the 10th birthday of Eclipse is that it remains a widepread and highly popular IDE, but others remain to give developers choice, and moving between IDEs and is far easier than now, with an increasing number of portable plug-ins. On a side issue, I will make another prediction about Eclipse on that birthday, just to add a bit more controversy: It will use Swing.
  3. Good luck to Eclipse, but this article seems to me to present a rather arrogant attitude (of course, that does not mean it is wrong!) Well, that's only to be expected as the author represents a vendor (MyEclipse) that has a vested interest in seeing Eclipse completely dominate the market. NetBeans looks exciting but I suspect large scale adoption of NetBeans will depend to a certain extent on how quickly major app server vendors release JEE compliant versions and how receptive customers will be in embracing JEE. Not sure how users will warm up to the so-called SOA development features in NB (BPEL, JBI, OpenESB etc.) as the major SOA infrastructure vendors will most likely base their tooling around Eclipse plug-ins and for their customers there won't be much of a choice.
  4. I hate to fan the flames, but the original article is a troll anyway. Regardless:
    Eclipse has led the way in terms of some IDE features (refactoring, for example), but in other ways it has been far behind,
    I hate to burst your bubble, but Eclipse's support for refactoring cannot hold a candle to what is available in IntelliJ. IntelliJ has always been the leader in this space. If the future world of java is one where one MUST use an IDE to do the job AND one MUST use one and only one IDE then it's way past time to move on to Ruby... A real value would be if people writing plugins could target a generic API so that we could all experience the great refactoring abilities of IntelliJ along side the great maven integration of NetBeans and great modelling and DB tools from Eclipse.
  5. I hate to burst your bubble, but Eclipse's support for refactoring cannot hold a candle to what is available in IntelliJ.

    IntelliJ has always been the leader in this space.
    I have no bubble to burst here; I have no experience of IntelliJ, so perhaps I should have qualified things. However, it will be interesting to see how IntelliJ fares in an increasingly FOSS Java world.
  6. I agree that Eclipse was good so far. I have been using Eclipse since 3 years. But now that i started using Netbeans, i found Netbeans is much better than Eclipse. Netbeans is more than an IDE. I liked JSF, BPEL, Webservices Support for quickly developing and running them. Regards Sreenath.V
  7. intellij[ Go to top ]

    <blockquoteI have no bubble to burst here; I have no experience of IntelliJ, so perhaps I should have qualified things. However, it will be interesting to see how IntelliJ fares in an increasingly FOSS Java world.</blockquote> Considering that a $500/license tool has been able to stand up against a number of free tools for at least 5 years, I'd say it's faring quite well :)
  8. However, it will be interesting to see how IntelliJ fares in an increasingly FOSS Java world.
    Agreed. As of right now, I see Jetbrains lack of support for maven (it had been slated for version 6, but the dropped it when they decided to build their own build environment) as a sign that they are losing their way a bit in this regard. Hands down they have the best core IDE. However, if they fight open source rather than embrace it, they will become more and more of a niche product. Albeit, a very nice one at the core.
  9. No need to be afraid...[ Go to top ]

    No need to be afraid, Eclipse is not going to ... eclipse all the other players, at least as long as NetBeans is out there. Open source communities don't vanish, although some companies might want that to happen ;) Btw, I was playing with NetBeans during weekend and created this little flash demo... can you show me how to create such an application in few minutes with Eclipse (and only with free tools)? At the Tech days in Prague my keynote demo was about databinding support for Matisse... yes we have a solution today (it just needs to be QA'ed and finished, but it works). Time to ask Eclipse - where is the release-quality support for Java EE 5? (is it 7 months to go to the Europa release?). Where is the ease of use in Visual Editor? Where do I get good Mobility tools on the Eclipse platform? What is Eclipse's/SWT's answer to the improvements in Swing (which btw influences the platform competition)? And so on... We've learned our lessons and worked hard to add features our community was asking for (yes better editor is the last large missing piece, check out latest daily builds, better editor is on the way)... now it's time for Eclipse to get some important work done.
  10. Re: No need to be afraid...[ Go to top ]

    I really want to switch to Netbeans, but I can't, because Netbeans is not doing things well and right! Seems to me, they just want to add more and more features. Matisse is a dictator GUI designer! For any kind of modifications, you have to use the GUI builder, it doesn't allow you to modify the generated code the way you want! To me this means putting limitations on developer's creativity. Based on my experience, to get a job 100% done, always you have to do some manual modifications to the generated code! From code completion and code suggestion perspective, Netbeans is really lags behind Eclipse. The good news is that, according to Roman, they are working on this (isn't it too late?). To create an Enterprise Application in Netbeans, you have to have the application server setup first, whereas in Eclipse (WTP), this is not a must, you can start and continue coding using "J2EE Runtime Library". Also, Eclipse does support more application servers by default. As a java developer, I really don't care so much about visual web design! What I'm looking for is a strong integration and good support for all major application servers. Below is a list of application servers supported by WTP vs NetBeans: WTP =============================== Apache Tomcat v3.2 Apache Tomcat v4.0 Apache Tomcat v4.1 Apache Tomcat v4.0 Apache Tomcat v5.5 Generic BEA WebLogic Server v8.1 Generic BEA WebLogic Server v9.0 Generic BEA WebLogic Server v9.2 IBM WebSphere v6.0 J2EE Runtime Library JBoss v3.2.3 JBoss v4.0 JOnAS v4 Oracle OC4J Standalone 10.1.3 Geronimo v1.0 Geronimo v1.1 WebSphere Application server Community Edition v1.0 Pramati 4.1.x NetBeans ================================= BEA Weblogic Server 9 JBoss Application server 4 Sun java system Application Server Tomcat 5.0 Tomcat 5.5
  11. Re: No need to be afraid...[ Go to top ]

    Hi Mohammed, As for Matisse you can modify the generated code. Choose a component, go to the properties window and choose the Code tab. Now you can override code Matisse generates and replace the code by your own. Two-way editing is really hard to be done correctly, so this is how we solve the problem. Yes Eclipse is indeed superior in coding and this will be fixed in NetBeans 6.0. It's never too late to add features people ask for :) As for integration with application servers, the old versions of Tomcat are in my opinion obsolete. NetBeans also supports WebSphere and JOnAS using plug-ins you can download separately. I also don't think the versions are strict - e.g. Weblogic plug-in supports both 9.0 and 9.2. So the lists are not that much different, if you include plug-ins you can get from the web. You are right though, that there are more plug-ins for Eclipse than for NetBeans and we are working on supporting the community to write additional plug-ins. We have launched a new partner program which will improve it and we're hiring a community manager to work with the community in this regard. And as our community grows, more and more plug-ins appear these days (unlike 2 years ago when not many people cared about NetBeans).
  12. No fear...[ Go to top ]

    Eclipse has always been intended for more than just Java development. From the beginning, the entire point of Eclipse projects has been to create frameworks and platforms along with exemplary products. These frameworks and platforms have been leveraged by many companies to make some great products. This focus has evolved of late and projects are starting to put more work into polish. A great example of this is BIRT. In fact, there's a demo on the BIRT site that shows off some pretty cool report generation: http://download.eclipse.org/birt/downloads/examples/misc/BIRT2.1Demo/EclipseDemo.html It's a little on the long side (11 minutes), but it does have a lot of material to cover. The point is that Eclipse is being used for Java development to be sure. It is also a great platform for building rich client applications. But there is so much more that you can do with Eclipse.
  13. Since you asked...[ Go to top ]

    At the Tech days in Prague my keynote demo was about databinding support for Matisse... yes we have a solution today (it just needs to be QA'ed and finished, but it works).
    Golly. Eclipse has this also. It was part of Eclipse 3.2. At the time, it was marked as provisional since the designers weren't totally happy with it. It's becoming official API with Eclipse 3.3. You can use it today (many people already are). I guess that you could say that it's in the later stages of QA, is basically finished and works pretty darned well. http://wiki.eclipse.org/index.php/JFace_Data_Binding
    Where is the release-quality support for Java EE 5? (is it 7 months to go to the Europa release?) We can play the you-don't-have-this-feature game all week. However, it is unproductive. The Eclipse Web Tools team is working hard to provide Java EE 5 support as part of the Europa simultaneous release (http://www.eclipse.org/projects/europa.php). Their project plan is posted so you can keep up with the functionality they're working on and even get involved with the development by testing the various milestones that will be released leading up to the big event. In the meantime, the Eclipse eco-system provides (http://www.eclipseplugincentral.com/Web_Links-index-req-viewcatlink-cid-9.html).
    Where is the ease of use in Visual Editor?
    It's coming. The Visual Editor continues to evolve. Personally, I like VE and do use it often. It doesn't support the GroupLayout yet: they decided to actually support existing layout managers and the whole round-trip engineering thing. If VE isn't quite your cup of tea, the eco-system provides (http://www.eclipseplugincentral.com/Web_Links-index-req-viewcatlink-cid-18.html). If you want to see a really good GUI builder, check out WindowBuilder (which does support GroupLayout and round-tripping). There are others, like Jigloo. And then there are a bunch of specialty GUI builders. There's a tonne of options on EPIC.
    Where do I get good Mobility tools on the Eclipse platform?
    http://www.eclipse.org/home/categories/embedded_device.php
    What is Eclipse's/SWT's answer to the improvements in Swing (which btw influences the platform competition)?
    This is just FUD. What improvements? Be specific. Where--in your opinion--is SWT not rising to the challenge? SWT continues to evolve to meet the needs of Eclipse developers. There's a lot of development teams out there pushing the limits of what can be done with SWT; the SWT team has been responding.
  14. Re: Since you asked...[ Go to top ]

    Hi Beaton, good to see a reply from you :)
    At the Tech days in Prague my keynote demo was about databinding support for Matisse... yes we have a solution today (it just needs to be QA'ed and finished, but it works).

    > Golly. Eclipse has this also. You can use it today (many people already are).

    Good! Which visual designers support it?

    Where is the release-quality support for Java EE 5? (is it 7 months to go to the Europa release?)
    > We can play the you-don't-have-this-feature game all week. However, it is unproductive. The Eclipse Web Tools team is working hard to provide Java EE 5 support as part of the Europa simultaneous release (http://www.eclipse.org/projects/europa.php). Their project plan is posted so you can keep up with the functionality they're working on and even get involved with the development by testing the various milestones that will be released leading up to the big event.

    In the meantime, the Eclipse eco-system provides (http://www.eclipseplugincentral.com/Web_Links-index-req-viewcatlink-cid-9.html).

    I know it's coming, I am just saying that if developers want to use opensource final-release quality tools for Java EE 5 today, they can use NetBeans. Europa is 7 months away, right? :)
    Where is the ease of use in Visual Editor?

    > It's coming. The Visual Editor continues to evolve. Personally, I like VE and do use it often. It doesn't support the GroupLayout yet: they decided to actually support existing layout managers and the whole round-trip engineering thing. If VE isn't quite your cup of tea, the eco-system provides (http://www.eclipseplugincentral.com/Web_Links-index-req-viewcatlink-cid-18.html).

    If you want to see a really good GUI builder, check out WindowBuilder (which does support GroupLayout and round-tripping). There are others, like Jigloo. And then there are a bunch of specialty GUI builders. There's a tonne of options on EPIC.

    Yes, but what I heard from developers is that Matisse is easier to use and more powerful than any of these GUI builders (which is the reason why MyEclipse includes it). So I think the VE guys should realize that - but in the interview with them I heard that VE as a platform is more important than good usability of VE. I don't think that's what the users are asking for but I might be wrong.
    Where do I get good Mobility tools on the Eclipse platform?

    > http://www.eclipse.org/home/categories/embedded_device.php FYI - in recent version of NetBeans Mobility Pack we support SVG on mobile phones, Swing on CDC mobile phones, there's complete support for CLDC, tools for device fragmentation, end-to-end solutions to access Java EE 5 backends from mobile devices, etc. So I would like to see similar tools on the Eclipse platform, to compare.
    What is Eclipse's/SWT's answer to the improvements in Swing (which btw influences the platform competition)?

    > This is just FUD. What improvements? Be specific. Where--in your opinion--is SWT not rising to the challenge? SWT continues to evolve to meet the needs of Eclipse developers. There's a lot of development teams out there pushing the limits of what can be done with SWT; the SWT team has been responding.
    I'll give an example, show me an application similar to Aerith, written in SWT.
  15. Re: Since you asked...[ Go to top ]

    Roman; Since MyEclipse has become part of the conversation, I'd like to point out that yes, we incorporated Matisse. We did so for many reasons, most of which can be found here in Wayne Parrott's blog. But truly, I think that the fact that we were able to incorporate Matisse into Eclipse illustrates what my original post proposed. Eclipse has approaching 150 member companies, and hundreds if not thousands of contributing developers. If they collectively wanted to get together and focus on creating one killer feature, it would be a matter of time. Instead, they choose to leave the platform concept intact and allow the various member companies (like us) to innovate and create. The very ability to port NetBeans features into Eclipse, as we have shown, proves that the technology behind Eclipse is what makes it successful. By incorporating Matisse, Genuitec showed that external technology (yes, even Swing-based technology) can be brought into the Eclipse ecosystem and innovated from the inside-out. This is a strength of the Eclipse ecosystem. Not a weakness.
  16. Re: Since you asked...[ Go to top ]

    Hi Beaton, good to see a reply from you :)
    Actually, most people call me Wayne. Though, I have been Beaton all my life...
    Good! Which visual designers support it?
    That's next. VE is working on support. I'm not sure what the commercial folks are planning, but I'd bet good money that Instantiations is going to support it.
    I know it's coming, I am just saying that if developers want to use opensource final-release quality tools for Java EE 5 today, they can use NetBeans. Europa is 7 months away, right? :)
    Very deft use of the word "final". You're right, Eclipse isn't fully there yet with a "final" release. However, we have software. The Dali project (http://www.eclipse.org/dali) provides support for EJB 3.0 today. And while it is not a "final release", Eclipse projects do have a long history of high-quality milestone releases. And still, there is the eco-system. As Jens points out, Eclipse has traditionally been focused on providing frameworks and platforms that others can use to provide great functionality. EPIC is packed with plug-ins for Eclipse that provide Jave EE 5 Functionality. If you don't want to hunt for plug-ins, there are more refined offerings available from folks like Genuitect, BEA, and IBM. You can find these on eclipseplugincentral.com
    Yes, but what I heard from developers is that Matisse is easier to use and more powerful than any of these GUI builders (which is the reason why MyEclipse includes it). So I think the VE guys should realize that - but in the interview with them I heard that VE as a platform is more important than good usability of VE. I don't think that's what the users are asking for but I might be wrong.
    You're probably wrong. Be honest: you spend about as much time as I do talking with the converted. So, of course the developers you talk to think that Matisse is easier. Most of the folks that I talk to have heard about or have looked at Matisse and think it looks pretty neat but still don't use it. I played with it for a while and (a) managed to break it in about 10 minutes; (b) got pissed off that I couldn't edit existing windows; and (c) got worried that I wouldn't be able to recover my work if I lost that silly "forms" file. Before you answer "we're working on it", remember your response to my answer to the Java EE 5 question. At the end of the day, I personally think that Matisse looks pretty cool, but frankly I don't care (and nor do most the developers I talk to). When you're building Eclipse RCP applications, you tend to build views that are individually relatively simple. VE does a fine job on these. And again, those developers who have more sophisticated GUI building requirements can turn to the eco-system can get mind-bendingly great tools (that the developers I talk to say blow the doors off Matisse) GUI building tools from companies like Instantiations (who, incidently, have a great RCP builder application now).
    FYI - in recent version of NetBeans Mobility Pack we support SVG on mobile phones, Swing on CDC mobile phones, there's complete support for CLDC, tools for device fragmentation, end-to-end solutions to access Java EE 5 backends from mobile devices, etc. So I would like to see similar tools on the Eclipse platform, to compare.
    I guess that you didn't follow the link I gave. Eclipse has a tonne of stuff in the mobile area, from eRCP and eSWT, through to device debugging and target management. Sure, it doesn't have the polish and shine that NetBeans does, but our eco-system is running with this stuff and is putting together some pretty slick products based on Eclipse technology. At Eclipse, we don't just hire a bunch of developers and set them off, great companies like Nokia, WindRiver, IBM, QNX, Freescale, TI, PalmSource, MentorGraphics/ATI, and more come to us and through us collaborate on projects that deliver the frameworks, platforms, and infrastructure they need to then go and build superior products. In the process, they build exemplary tools. These people all know quite a bit about building software for the various devices and know what they need to get the job done. There's already a lot of great stuff there. Do we have a "Mobility Pack" today? no. But real people are getting real work done and that's what really counts, right?
    I'll give an example, show me an application similar to Aerith, written in SWT.
    Ah, you've changed the focus (more FUD Roman, shame on you). But I can play this game to. Show me a great GIS platform built using the NetBeans Rich Client Platform (http://udig.refractions.net/). Or a great reporting tool (http://www.eclipse.org/birt) tightly integrated into NetBeans. See, this is a fun game. I guess that since nobody has built a reporting tool for NetBeans, that NetBeans can't do reporting, right?
  17. Re: Since you asked...[ Go to top ]

    Ah, you've changed the focus (more FUD Roman, shame on you). But I can play this game to. Show me a great GIS platform built using the NetBeans Rich Client Platform (http://udig.refractions.net/). Or a great reporting tool (http://www.eclipse.org/birt) tightly integrated into NetBeans. See, this is a fun game. I guess that since nobody has built a reporting tool for NetBeans, that NetBeans can't do reporting, right?
    Well, to me sounds like this is a cheating answer - or in other words, _you_ are changing the focus. :-) The question about Aerith is not concerning what kind of functionality Aerith is about (it manages photos, but for the sake of this discussion it could be whatever else, an iTunes clone etc...). The point about Aerith is that it looks cool and is very rich of visual effects. Now, let me restate the original question: does exist some SWT application which is so cool in visual effects as Aerith? From your answer pointing at GIS and whatever, I'd say it doesn't exist. :-) But of course I can be wrong.
  18. Re: Since you asked...[ Go to top ]

    Hi Beaton, good to see a reply from you :)
    >Actually, most people call me Wayne. Though, I have been Beaton all my life...
    Oops, sorry, my mistake!
    Very deft use of the word "final". You're right, Eclipse isn't fully there yet with a "final" release. However, we have software. The Dali project (http://www.eclipse.org/dali) provides support for EJB 3.0 today. And while it is not a "final release", Eclipse projects do have a long history of high-quality milestone releases...
    I don't think we'll agree on much here, but looks like this thread cought some interest of the WTP folks.
  19. Re: No need to be afraid...[ Go to top ]

    What is Eclipse's/SWT's answer to the improvements in Swing (which btw influences the platform competition)? Comedy gold Jerry. Swing is the toolkit that needs to show the answers for the disaster it's been over the years. You can talk in 5 years or so when Java 6 has some decent penetration numbers.
  20. Re: No need to be afraid...[ Go to top ]

    Swing is the toolkit that needs to show the answers for the disaster it's been over the years. You can talk in 5 years or so when Java 6 has some decent penetration numbers.
    Old debate, Frank. Been through all this before, including showing you how Swing has been used even with current versions of Java for high-rated applications - even shrink-wrapped ones. Last year a very widely used application - Maple - switched their interface to Swing. Reviews included phrases like "spiffy new interface" and "Still, even such power users will find the new interface a pleasure to use". Swing certainly was something of a disaster, but that was a long time ago. Time to put this old argument in the trash.
  21. Re: No need to be afraid...[ Go to top ]

    Swing certainly was something of a disaster, but that was a long time ago. Time to put this old argument in the trash.
    No dice Steve. Java 6 isn't out yet, and most of have good vision. The Swing disaster will be with for a long time to come.
  22. Re: No need to be afraid...[ Go to top ]

    Swing certainly was something of a disaster, but that was a long time ago. Time to put this old argument in the trash.


    No dice Steve. Java 6 isn't out yet, and most of have good vision. The Swing disaster will be with for a long time to come.
    OK then - here is a challenge - find me a review of the new Maple Swing interface that in any way backs up your view. The fact is, that Swing can, and has, been used to develop first-class user interfaces. If it was a disaster, then there would have been no commercial advantage in MyEclipse incorporating Mattise.
  23. Re: No need to be afraid...[ Go to top ]

    OK then - here is a challenge - find me a review of the new Maple Swing interface that in any way backs up your view.
    Are you kidding me? Here, I'll find you tens of thousands of websites that say Swing sucks, because you know, people have eyes. Stop grasping at straws.
    The fact is, that Swing can, and has, been used to develop first-class user interfaces. If it was a disaster, then there would have been no commercial advantage in MyEclipse incorporating Mattise.
    The fact is that Swing has been a disaster, and just because swing-developer Steve Zara tries to convince himself otherwise, doesn't change anything. Once again Steve, your solipsism is predictable, but boring. But I guess once you've decided to adopt something, it makes sense to delude yourself into thinking it doesn't suck. Hell, even the open source desktop guys don't touch that Swing crap with a 10-foot pole. It looks like crap on every desktop. If only Sun hadn't gone the emulation route.
  24. Re: No need to be afraid...[ Go to top ]

    Are you kidding me? Here, I'll find you tens of thousands of websites that say Swing sucks, because you know, people have eyes. Stop grasping at straws.
    Sure, Swing has had a long history of sucking. But that is in the past. Some of us have moved on. I don't consider following what MyEclipse are doing to be grasping at straws...
    Hell, even the open source desktop guys don't touch that Swing crap with a 10-foot pole. It looks like crap on every desktop. If only Sun hadn't gone the emulation route.
    There are plenty of open source Swing projects, and open source JRE developers have been working to get Swing support going in their runtimes. The Kaffe guys have been doing it, and so have GNU classpath. I find this exciting, not boring. They seem to have abandoned the 10-foot pole. Not only that, this year the Harmony managed to run JEdit, and they were "proud" to demonstrate it. http://www.theinquirer.net/default.aspx?article=32116 You may say Swing looks like crap, but surely what matters is the users. More reviews of Maple: "Many things can be done faster, more intuitively, more productively and more enjoyably…" (Scientific Computing World) "Maple 10's most obvious strengths are in the friendly developments to its general interface... " (also Scientific Computing World) "The new Maple interface is the most exciting development I’ve seen in highend mathematics packages in several years." (Desktop Engineering) Emulation is a great way to do things. Even SWT does it, by emulating components that aren't present on some systems.
  25. Re: No need to be afraid...[ Go to top ]

    ure, Swing has had a long history of sucking. But that is in the past. Some of us have moved on. I don't consider following what MyEclipse are doing to be grasping at straws...
    Once again Steve, it's not about you, and you deciding "that is in the past". I don't know what *you* have move on to, but Swing's problems haven't really "moved on" until Java 6 - and even then, there is always going to be problems with emulation.
    There are plenty of open source Swing projects, and open source JRE developers have been working to get Swing support going in their runtimes. The Kaffe guys have been doing it, and so have GNU classpath. I find this exciting, not boring. They seem to have abandoned the 10-foot pole. Not only that, this year the Harmony managed to run JEdit, and they were "proud" to demonstrate it. http://www.theinquirer.net/default.aspx?article=32116
    Of course Kaffe and Classpath are doing Swing. They're recreating Java SE for Christ's sake! But if you're going to bring up that irrelevance, you should also note that Classpath is using native peers, like OSX.
    You may say Swing looks like crap, but surely what matters is the users. More reviews of Maple:
    Maple is completely irrelevant. I never said that NOBODY had ever used Swing.
    Emulation is a great way to do things. Even SWT does it, by emulating components that aren't present on some systems.
    I think you're finally making some progress Steve. SWT is emulating components that aren't present on some systems. And now in Java 6 we have Swing using more and more of the native APIs. Hmm, see any pattern here Steve?
  26. Re: No need to be afraid...[ Go to top ]

    They're recreating Java SE for Christ's sake! But if you're going to bring up that irrelevance, you should also note that Classpath is using native peers, like OSX.
    Sorry, but you can't have it both ways; first say that the open source guys aren't touching Swing with a 10-foot pole, but then explain how they are implementing it.
    Maple is completely irrelevant. I never said that NOBODY had ever used Swing.
    No, what you said was that "It looks like crap on every desktop." Maple is highly relevant, as it is a very widely used and popular application present on a range of platforms, and praised for the high standard of its Swing interface. It shows that your generalisation is wrong.
    Emulation is a great way to do things. Even SWT does it, by emulating components that aren't present on some systems.


    I think you're finally making some progress Steve. SWT is emulating components that aren't present on some systems.

    And now in Java 6 we have Swing using more and more of the native APIs. Hmm, see any pattern here Steve?
    Sure. It means Swing is improving even more, and will have more fidelity with 'native systems'. I never said Swing could not improve... The point I have demonstrated beyond doubt is that developers have already been using Swing to produce high-quality and well-reviewed user interfaces for years. One of the reasons I picked Maple (apart from the fact it so widely used) is that the developers achieved this with some time ago, with Java 1.4, so any idea that you will have to wait for Java 6 is plainly wrong. Why am I arguing here? Because I am curious as to why Java is still being judged by the way it was 5 or more years ago. I find it strange that in an industry where there is so much change, that ideas in some areas are so fixed.
  27. Re: No need to be afraid...[ Go to top ]

    Sorry, but you can't have it both ways; first say that the open source guys aren't touching Swing with a 10-foot pole, but then explain how they are implementing it.
    Steve I thought this was such a simple concept, that it wouldn't have to be explained to you. I'm not talking about people that are implementing Java SE. Obviously they would be implementing Swing, with native peers (like it should have been done by Sun). It's about application developers whose applications get put in a repository in the open source desktop world.
    o, what you said was that "It looks like crap on every desktop." Maple is highly relevant, as it is a very widely used and popular application present on a range of platforms, and praised for the high standard of its Swing interface. It shows that your generalisation is wrong.
    No, Swing does look like crap on every desktop. The people who made Maple decided that Swing looking like crap was outweighed by whatever perceived benefits they came up with for programming with the Swing API.
    Sure. It means Swing is improving even more, and will have more fidelity with 'native systems'. I never said Swing could not improve...
    Except Sun blundered by starting from the wrong end. They should have been emulating what needed to be emulated and using native peers for what they can.
    The point I have demonstrated beyond doubt is that developers have already been using Swing to produce high-quality and well-reviewed user interfaces for years. One of the reasons I picked Maple (apart from the fact it so widely used) is that the developers achieved this with some time ago, with Java 1.4, so any idea that you will have to wait for Java 6 is plainly wrong.
    Steve, I could produce a usable interface using Ncurses. That doesn't change the fact that it's going to be quite unappealing on a modern desktop. Yeah, I can produce a usable interface using Swing. That doesn't change the fact that pre-Java 6 Swing looks like crap.
  28. Re: No need to be afraid...[ Go to top ]

    Sorry, but you can't have it both ways; first say that the open source guys aren't touching Swing with a 10-foot pole, but then explain how they are implementing it.


    Steve I thought this was such a simple concept, that it wouldn't have to be explained to you. I'm not talking about people that are implementing Java SE. Obviously they would be implementing Swing, with native peers (like it should have been done by Sun). It's about application developers whose applications get put in a repository in the open source desktop world.
    You mean like Sourceforce? Searching for Java Swing gives 13,284 results. Try it for yourself, and see the results. Your statement that open source developers aren't willing to touch Swing could not be more clearly shown to be wrong.
    No, what you said was that "It looks like crap on every desktop." Maple is highly relevant, as it is a very widely used and popular application present on a range of platforms, and praised for the high standard of its Swing interface. It shows that your generalisation is wrong.


    No, Swing does look like crap on every desktop. The people who made Maple decided that Swing looking like crap was outweighed by whatever perceived benefits they came up with for programming with the Swing API.
    No, sorry, you can't get away with this. Maple has been specifically praised for the look and quality of its interface. You can't twist this into some kind of ease of development motivation for developers. Maple is a product that works graphically - the quality of the interface is important.
    Except Sun blundered by starting from the wrong end. They should have been emulating what needed to be emulated and using native peers for what they can.
    Which is irrelevant. Yes, I agree that they blundered with Swing at the start. Not because of using native peers (anyway that term is meaningless on many platforms), but because they lifted the design wholesale from a Smalltalk system, which was far more highly tuned to the Swing approach. It has taken years to overcome the performance issues that resulted from that. But we aren't talking about then. We are talking about the past few years.
    The point I have demonstrated beyond doubt is that developers have already been using Swing to produce high-quality and well-reviewed user interfaces for years. One of the reasons I picked Maple (apart from the fact it so widely used) is that the developers achieved this with some time ago, with Java 1.4, so any idea that you will have to wait for Java 6 is plainly wrong.


    Steve, I could produce a usable interface using Ncurses. That doesn't change the fact that it's going to be quite unappealing on a modern desktop. Yeah, I can produce a usable interface using Swing. That doesn't change the fact that pre-Java 6 Swing looks like crap.
    Again, it is not about a usable interface, in Swing or ncurses. It is about an interface that has been specifically reviewed and praised as being attractive and appealing. Here is another review: "The most obvious addition is a sparkling new graphical user interface...." http://www.spectrum.ieee.org/feb06/2786
  29. Re: No need to be afraid...[ Go to top ]

    You mean like Sourceforce? Searching for Java Swing gives 13,284 results. Try it for yourself, and see the results. Your statement that open source developers aren't willing to touch Swing could not be more clearly shown to be wrong.
    No, I'm talking about real applications that people use and are in distro repositories. I'm not talking about random crap that nobody cares about on sourceforge. Of course the open source fundies didn't consider Java "free", and that's a consideration, but when the toolkit doesn't fit in with KDE or Gnome, it's another strike against it.
    No, sorry, you can't get away with this. Maple has been specifically praised for the look and quality of its interface. You can't twist this into some kind of ease of development motivation for developers. Maple is a product that works graphically - the quality of the interface is important.
    What, by blind people?
    Which is irrelevant. Yes, I agree that they blundered with Swing at the start. Not because of using native peers (anyway that term is meaningless on many platforms), but because they lifted the design wholesale from a Smalltalk system, which was far more highly tuned to the Swing approach. It has taken years to overcome the performance issues that resulted from that. But we aren't talking about then. We are talking about the past few years.
    Sun still uses emulation, so it's not just the past few years.
    Again, it is not about a usable interface, in Swing or ncurses. It is about an interface that has been specifically reviewed and praised as being attractive and appealing.
    No, I'm not interested in random reviewers that are apparently blind, or just don't care if the output of a toolkit looks like ass. The bottom line Steve is that you seem to think that Swing looking like ass went away a few years ago. I disagree. Java 6 looks like it'll be pretty acceptable and will be pretty native looking, but still years away from having significant penetration
  30. Re: No need to be afraid...[ Go to top ]

    The bottom line Steve is that you seem to think that Swing looking like ass went away a few years ago.
    Not me, Frank, users and developers. Thousands of them. Come on... First you say that "even the open source desktop guys don't touch that Swing crap with a 10-foot pole." Apart, of course, from the open source Java implementors. Then you say that those don't count, so I show you over 13,000 projects on Sourceforce in Swing. Then you say that those don't count, because you are labelling 13,000 projects "random crap that nobody cares about" and you need to see real applications that people use and are in distro repositories. Well, that is going to be hard to find, isn't it? Because almost all distro repositories contain pure FOSS, and Swing support on FOSS Java implementations is still being worked on! Well done: you have kept moving the goalposts until they are just close enough together to exclude Swing apps. There are, of course, plenty of Swing apps that run on Linux. At least 13,000 of them! (And, of course, Maple).
    What, by blind people?
    And, of course, any reviewer or user who disagrees with you must be blind.
    Java 6 looks like it'll be pretty acceptable and will be pretty native looking, but still years away from having significant penetration
    You are out of date here as well. Swing is showing more market penetration than even GUI systems designed specifically for Windows: from indeed.com today: WinForms jobs 1,300. Java Swing jobs: 3,000.
  31. Re: No need to be afraid...[ Go to top ]

    There are two things missing in Swing in JDK 1.5 that are very important to me: font anti-aliasing (which I've had in Eclipse since day one) and native dialogues (I have a lot of software installed that connects to the context-menu, e.g., in "File Open" dialogue). From what I hear, font anti-aliasing will be in JDK 1.6 - but not the native dialogues. Steve, is that correct? If so, that still puts Swing at a disadvantage in my book.
  32. Re: No need to be afraid...[ Go to top ]

    There are two things missing in Swing in JDK 1.5 that are very important to me: font anti-aliasing (which I've had in Eclipse since day one) and native dialogues (I have a lot of software installed that connects to the context-menu, e.g., in "File Open" dialogue). From what I hear, font anti-aliasing will be in JDK 1.6 - but not the native dialogues. Steve, is that correct? If so, that still puts Swing at a disadvantage in my book.
    Anti-aliasing is there. I am not sure what you mean regarding the dialogues - not something I have worked with. The functionality you require may well be there in the Java Desktop Integration Components.
  33. Re: No need to be afraid...[ Go to top ]

    Anti-aliasing is there. I am not sure what you mean regarding the dialogues - not something I have worked with. The functionality you require may well be there in the Java Desktop Integration Components.
    Since everybody has LCDs these days, he's talking about subpixel rendering. Pre-Java 6 fonts look horrible on an LCD.
  34. Re: No need to be afraid...[ Go to top ]

    That's right. I use three LCDs (notebook + 2 external LCDs) and have Windows anti-aliasing. One thing that I immediately noticed when I downloaded Netbeans for a trial is that the fonts where jaggy and looked different from any native app on my machine.
  35. Re: No need to be afraid...[ Go to top ]

    What I mean with "native dialogues" is that the "File open" dialogue of Eclipse is the real Windows "File Open" dialogue, not the emulated-but-not-really-right looking Swing dialogue. That is why when I right-click on these files, I get my Windows context menu with the regular Windows context menu plus all the software that installs to this context menu.
  36. Re: No need to be afraid...[ Go to top ]

    .....
    We've learned our lessons and worked hard to add features our community was asking for (yes better editor is the last large missing piece, check out latest daily builds, better editor is on the way)... now it's time for Eclipse to get some important work done.
    Roman, As a long time NetBeans user, I have to ask you how you define "large" here: If you consider "refactor" as a large piece, then Editor is definitely NOT the "last" one missing. We need NetBeans refactor to catch up with Eclipse/IntelliJ. And we need it NOW! I understand it takes time to play catch-up game, but how long has it been since the Refactor project started with NB, two years? And just as it caught up about half of the features that are already there with Eclipse/IntelliJ, it stalled at 5.0 - no new refactor features added in 5.5 at all! Ok, 5.5's focus is Java EE, but what about 6.0? No new feature, either! Why, "becasue we are switching the underlying refactor engine to Jackpot". I don't care what potentials a new technology (Jackpot or anything else) brings, if it doesn't bring values to the end users fast and solid, it does NOT matter!!! So please, catch up on Refactor already!
  37. Re: No need to be afraid...[ Go to top ]

    .....
    We've learned our lessons and worked hard to add features our community was asking for (yes better editor is the last large missing piece, check out latest daily builds, better editor is on the way)... now it's time for Eclipse to get some important work done.
    Roman, As a long time NetBeans user, I have to ask you how you define "large" here: If you consider "refactor" as a large piece, then Editor is definitely NOT the "last" one missing. We need NetBeans refactor to catch up with Eclipse/IntelliJ. And we need it NOW! I understand it takes time to play catch-up game, but how long has it been since the Refactor project started with NB, two years? And just as it caught up about half of the features that are already there with Eclipse/IntelliJ, it stalled at 5.0 - no new refactor features added in 5.5 at all! Ok, 5.5's focus is Java EE, but what about 6.0? No new feature, either! Why, "becasue we are switching the underlying refactor engine to Jackpot". I don't care what potentials a new technology (Jackpot or anything else) brings, if it doesn't bring values to the end users fast and solid, it does NOT matter!!! So please, catch up on Refactor already!
  38. Dont think so[ Go to top ]

    Current trends does not support the claim Eclipse will "become a household name by year 10". I have used eclipse for three years and found good for standard java development. But when it comes to new technologies JSF, EJB3, Web Service NetBeans is way better. Today Netbeans provides integrated IDE for JSF, AJAX, EJB, WebServices, BPEL. What else I could ask for for free!! Eclipse is too much of a pain downloading all plug-ins and then troubleshooting when they dont work as expected. Netbeans also has improved a lot in recent past and has integrated view for Runtime environment! I can understand why Java developers are moving towards Netbeans.
  39. Que sera sera...[ Go to top ]

    Honestly, I don't know. The one thing that I take for granted is people do not behave rational :-). I would have thought that Eclipse would totally destroy the IDE market, but since Netbeans 5 I am not so sure anymore. And there are brilliant tools out there that just lack marketing, like Oracles JDeveloper....we'll see.,
  40. Wishful thinking[ Go to top ]

    These days I moved back to NetBeans for Swing, Web and Java EE5 development and am much happier. Eclipse does not have as much competitive advantage over NB as it had, although it still gets a lot more press and industry support. But, for me, I like the integrated "everything works out of the box" Netbeans experience much more. Eclipse could learn from it, Callisto didn't nearly achieve that.
  41. would Net Beans be where it's at today[ Go to top ]

    Here is a dumb question. If eclipse didn't appear, would Net Beans have gotten to where it is today? I think the competition has helped accelerate and improve free IDE's.
  42. Here is a dumb question. If eclipse didn't appear, would Net Beans have gotten to where it is today?

    I think the competition has helped accelerate and improve free IDE's.
    Peter; I think this definitely works both ways. Just like history has shown, nothing advances technology quite like a "war" economy. The pushing back and forth has made both camps work harder, and I credit both with innovation as a result.
  43. MyEclipse[ Go to top ]

    Hi Jens, Off topic question but you didn't answer my question on the MyEclipse thread so asking again :-) When is MyEclipse going to support portlet development and deployment to portal servers?
  44. Re: MyEclipse[ Go to top ]

    When is MyEclipse going to support portlet development and deployment to portal servers?
    Satadru; We are going to be exploring this issue in early '07. But, as with every feature, we want to be sure to include the most effective solutions, which requires a bit of critical analysis. We should have a more concrete position in the coming months. Thanks! -Jens
  45. Eclipse Ubiquious?[ Go to top ]

    I doubt it. I think it is a great platform and I'm very excited by the wide range of uses that exists for eclipse. I also make sure to keep up to date on it. Personally I use Netbeans for my Java development, other then my corporate job which requires JDeveloper (BTW I like JDeveloper and have no beef using it). At this point I'm only using Eclipse for ROR development using the RDT and WTP toolkits.
  46. In five years from now, we might have 1,2 or 3 new players that are not even started up yet. All the commercial IDE might have gone open-sources or the other way around. And genetic science might have come up with a chicken with shark theeth. Hammers are very simple tools that have a little more than 5 years of existence if my informations are accurate. If we apply the same simplistic logic of the author, there should only be one universal hammer available today. Everybody would happily have flatted thumbs and index fingers using that same hammer on 5 inches and half inch framing nails (ouch!!!), sheet metal shaping, etc, and, of course, nobody would ask themselves if there could be a better hammer for the job. The reality is that IDE's are nothing but sophisticated hammers and each of them have their strenghs and weaknesses. Eclipse was for a time the only serious open-source player. With NetBeans 5.0+, this situation is turning around very fast. I even came accross a Java book where all examples and howtoos were written for Netbeans (at aPress I think). This would have been unthinkable one year ago. I think Eclipse has some great features but I simply don't buy the plug-ins approach. I largely prefer the Netbeans approach where the main functionnalities are core features yet, specialties can be implemented as modules. And this approach seems to gain momentum every day. My whish is that all players in the IDE field provide interoperability so the choice of the IDE would be a personnal preference within a multi-developpers project. It is somewhat possible today but improvements are still needed. I hope that Eclipse people react intelligently to this shift and spend their time in improving their product to face this new challenge. That way, in five years from now, there will still be more than one hammer available... Jacques Ledoux
  47. Your post and others seem to make the assumption that we are simply comparing IDEs. That isn't really what the opinion of the article is about at all. Rather, it has more to do with vendor adoption of Eclipse as a platform / framework. The fact is that the plug-in approach will come to dominate as the right approach for building highly modular software. Eclipse's plug-in approach has transformed into an OSGi services approach, and before you knock these as plug-ins, go do a bit of research on how OSGi handles classloading and deployment. NetBeans may be nice to use from the perspective of a business application developer building Java apps. Eclipse, as a platform for vendor integration, however, is superior. Eclipse, as a community for vendor cooperation, however, beats NetBeans. The ubiquity Eclipse will enjoy will be ubiquity among tool vendors as a marketplace to play in. I'm sure you'll even be able to use NetBeans to code up that OSGi service for your next-gen Spring+Equinox server stack . But then, this is just opinion, also. Regards, Michael Murphree
  48. ...is the huge number of plugins that aren't Java related. Eclipse has pretty much become the standard alternative to Emacs for creating development environments. The other thing, to a lesser extent, is Eclipse as a RCP. Netbeans is starting to focus on this, but is playing catchup, and has in the past been hindered by Swing.
  49. Looks like NetBeans is doing well[ Go to top ]

    Hi, Before a year or two, if the same topic was posted, i doubt anyone would have mentioned about NetBeans in an Eclipse thread. Yes, Eclipse was the IDE of choice at that point of time. But things are changing very fast. Almost everyone who replied in this thread mentioned about NetBeans. That itself is a victory for NetBeans. We were very reluctant initially to switch to NB. But dude, we just couldn't resist the offerings like Matisse, Java EE 5 support, Profiler etc. Looks like NB has learnt the lessons. Looks like NB is doing well indeed. May be in few years from now, i foresee NetBeans replacing Eclipse as the major Java IDE.
  50. Eclipse is only an IDE, not a new techology, If five years ago jbuilder was free and open souce,I'm sure that jbuilder would be more popular than eclipse. "Netbeans is more than an IDE" Agreed. IMO,eclipse is only a framework, in order to develop a web application, we have to buy plugin(e.g. myeclipse),whereas, we can develop web application with NB easily. developer tutorials
  51. ...is the huge number of plugins that aren't Java related. Eclipse has pretty much become the standard alternative to Emacs for creating development environments.

    The other thing, to a lesser extent, is Eclipse as a RCP. Netbeans is starting to focus on this, but is playing catchup, and has in the past been hindered by Swing.
    Frank; You've hit on a big issue. Eclipse is successful not only because of its core Java components, but also because it truly is a platform for all sorts of applications. IDE flamewars are all fine and good, but when you envision the success of the Eclipse platform, not of Eclipse exclusively as a Java IDE (though this is how many folks see Eclipse), the picture gets bigger. Tools-only vendors will find difficult to compete with Eclipse in the many infiltrated verticals as a result of RCP and non-Java areas as you mentioned.
  52. Tools-only vendors will find difficult to compete with Eclipse in the many infiltrated verticals as a result of RCP and non-Java areas as you mentioned.
    This is why JSR198, or some similar mechanism, could be important. Why shouldn't tools be portable between IDEs? Hopefully use of NetBeans (and other IDEs) will continue to grow their share of the market, then at a certain point tool and plug-in vendors will put the pressure IDE developers for this kind of portability. I suspect I am being far too optimistic.
  53. Just check out where netbeans was 2 years ago and where it is now. And also check out netbean's maven2 support through Mevenide. The only true advantage of eclipse is the "compile on save" and error reporting. I just dont see eclipse evolve as much as netbeans and i believe that the main factor for this is eclipse's arrogance. Anyway a developer just chooses the IDE that fits his personal development way. And if you combine the features of the current players in the market you have the ultimate IDE: Eclipse does some things , Netbeans some others, IDEA some others. Competition is essential to further improvement of any IDE and this is the best for the developers( just to have choices). if eclipse continue to hide behind its finger i think that it will just lose the game. and just because i dont want to start a flame war i use both netbeans and eclipse and yes eclipse rcp is better than netbeans but until when?
  54. Well, I expected that such a debate can't but turn into a flame, nevertheless some more objective points could be tried - please? :-)
    Pre-Java 6 fonts look horrible on an LCD.
    For instance, I don't understand the recurring point about anti-aliasing. Maybe I live on Mars? I'm running a MacBook Pro, a MacMini with Ubuntu and an old desktop with Windows. I've run NetBeans 5.5 (but it was the same with 5.0). All the three with Java 5, not Java 6. I'm seeing antialiased fonts everywhere. What am I doing wrong?
    That doesn't change the fact that pre-Java 6 Swing looks like crap.
    Maybe I'm still on another planet, but I'm looking at the Swing Sightings page and I'm seeing a lot of Swing applications. Yes, there's that blind-made Maple. But there are a lot of others. Some are so-so, others appear very good to me (maybe it's also a matter of programmers' skills?). Can you elaborate on the "crap" concept? Maybe it's too subjective :-)
    Your statement that open source developers aren't willing to touch Swing could not be more clearly shown to be wrong.
    Well, this is probably one of the most interesting points. My opinion is that if I need some special UI component, I just search java.net or other places and I find it. For instance, this summer I needed a component to embed Google Maps in my application. I found JXMapViewer from SwingLabs (that delivers a lot of other advanced components, such as date pickers and so on). Integrating it in my application was a breeze. Perhaps it's just because I'm definitely on the Swing side, and I don't know about SWT: does exist such a component for SWT? I could perfectly be wrong (please contradict me), but I don't see the same amount of extra UI components in the SWT world. On java.net I see daily a lot of works about new cool widgets (besides SwingLabs guys it's well worth while mentioning Kirill Grouchnikov's work), and they are all Swing-based. And now to the Eclipse vs NetBeans. Talking about the IDE. I've been using for years, at least since 2003 but probably even earlier, and I'm really grateful to the Eclipse people for having delivered such a beautiful IDE for all this time - in the free world there were really no good alternatives. I had evaluated NetBeans in 2002, and soon get rid of it. When 5.0 came out, I was pretty skeptical about it. I had to change my mind, even though there were still many points behind Eclipse. With NetBeans 5.5. most of them are on par. If the trend goes on, I suspect that with 6.0 the gap could be completely filled. In the meantime, I think that NetBeans 5.5 is much superior for the J2EE stuff. Also some customers of mine for which I did some comparative evaluation think the same. Is it NB still buggy? Probably there's still some annoying bug, but they're catching up, and Eclipse isn't immune too. For instance, some weeks ago I was driven mad by trying to set up the JBoss Plug-In - there were some blocking bugs that I've learnt to patch manually from JBoss maling lists. Is this an Eclipse blame? Probably not, but the out-of-the-box experience for J2EE in the end is absolutely superior for NetBeans. NetBeans vs Eclipse RCP? Until this year I wasn't really much involved in RCP stuff. I had evaluated Eclipse in mid 2005, but I really didn't like at all to learn a new API since I already knew Swing. As I didn't have any professional need, I just gave up. This year my needs changed and I started using NetBeans RCP. I like it very much, I appreciate the point that I'm just using enhanced Swing and that I find a lot of extra components that can be easily integrated (see above). I'm aware that Eclipse is superior for some aspects (e.g. OSGi) but they are not relevant for me today, and in perspective I think NetBeans will be able to catch up. PS I should also mention that when I switched from my G4 Mac to the Intel Mac, NetBeans and all my GUI stuff just worked. For Eclipse I had to wait until 3.2. Ok, a microprocessor switch isn't something that happens every year, nevertheless...
  55. Hi
    Maybe I'm still on another planet, but I'm looking at the Swing Sightings page and I'm seeing a lot of Swing applications. Yes, there's that blind-made Maple. But there are a lot of others. Some are so-so, others appear very good to me (maybe it's also a matter of programmers' skills?). Can you elaborate on the "crap" concept? Maybe it's too subjective :-)
    All opinions here are subjective. I took a look at the Swing Sightings, and I've got to agree with Frank - they look crap to me too :^). Superior features could induce me to use a swing app though, despite the looks. For example IntelliJ IDEA is Swing, looks crap, but is still a very good app. BTW as for SWT, well I tried to use it once, and found it limiting. It looks great on Eclipse, but not that good for your own bespoke stuff I find. There you go a couple more opinions :^). Paul.
  56. Hi


    Maybe I'm still on another planet, but I'm looking at the Swing Sightings page and I'm seeing a lot of Swing applications. Yes, there's that blind-made Maple. But there are a lot of others. Some are so-so, others appear very good to me (maybe it's also a matter of programmers' skills?). Can you elaborate on the "crap" concept? Maybe it's too subjective :-)


    All opinions here are subjective. I took a look at the Swing Sightings, and I've got to agree with Frank - they look crap to me too :^).
    I really can't believe that anyone can think that of all those applications. Are you seriously saying that you think that the truly stunning Podereso Manager looks "crap"? Or you think that Oculus - producing major business visualisation tools has an interface that looks "crap"? Or that the International Children's Digital Library written by the University of Maryland has a "crap" interface? I like debate, and I have more respect for those who are prepared to admit that these things are subjective, than those who make broad objective generalisations like "Swing is crap on all platforms", but I have to say that I find such sweeping generalisations as you have made hard to take seriously, as, I would imagine, would the developers of those interfaces I have mentioned.
  57. Hi


    Maybe I'm still on another planet, but I'm looking at the Swing Sightings page and I'm seeing a lot of Swing applications. Yes, there's that blind-made Maple. But there are a lot of others. Some are so-so, others appear very good to me (maybe it's also a matter of programmers' skills?). Can you elaborate on the "crap" concept? Maybe it's too subjective :-)


    All opinions here are subjective. I took a look at the Swing Sightings, and I've got to agree with Frank - they look crap to me too :^).


    I really can't believe that anyone can think that of all those applications. Are you seriously saying that you think that the truly stunning Podereso Manager looks "crap"? Or you think that Oculus - producing major business visualisation tools has an interface that looks "crap"? Or that the International Children's Digital Library written by the University of Maryland has a "crap" interface?

    I like debate, and I have more respect for those who are prepared to admit that these things are subjective, than those who make broad objective generalisations like "Swing is crap on all platforms", but I have to say that I find such sweeping generalisations as you have made hard to take seriously, as, I would imagine, would the developers of those interfaces I have mentioned.
    Hi Steve, I'm not trying to be controversal, but all Swing apps to me me look poor. They just have this dull finish to them. Not bright and dazzling like native apps. I've noticed that on the Mac Swing apps look better then they do on Windows, but you can still tell the difference straight away. I was following the debate as an impartial observer (like I've said I've got my issues with SWT too, unless it's improved significantly recently). And from what has been said I was curious to see if Swing had somehow got better. From the screen shots IMO no. So I don't know if Franks preferred adjective of crap is the best word to use, but I know what he means. Paul.


  58. Hi Steve,

    I'm not trying to be controversal, but all Swing apps to me me look poor. They just have this dull finish to them. Not bright and dazzling like native apps.
    Perhaps you are right. In spite of all the hundreds of different looks and feels produced for Swing, some hacked by amateurs, some by skilled professionals, and in spite of the best efforts of GUI designers from some of the most successful software houses, all Swing apps have a 'dull finish'. All of them. Perhaps a bug report should be submitted for Java 7 (probably too late for Java 6)... next time you work on Swing ... don't forget to bring the shiny.
  59. I'm not trying to be controversal, but all Swing apps to me me look poor.
    Hmm... usually I'm really skeptical about sentences with all or never. :-)
    They just have this dull finish to them. Not bright and dazzling like native apps.
    Does Aerith look dull to you? Have you ever tried MoneyDance, or MagicDraw?
    I've noticed that on the Mac Swing apps look better then they do on Windows, but you can still tell the difference straight away.
    Really? On the Mac, there are just few components that surely can defined crap, for instance the JFileChooser (but they can be replaced with the Quaqua package). AFAIK NeoOffice has a GUI made with Swing and the purpose of NeoOffice is to replace the GUI of OpenOffice (which surely looks crap on Mac OS X). Still convinced of that all? Do you think that NeoOffice is dull?