Article:  |
A Look at the Eclipse IDE | |
| Subject: | For the next version: port it to Swing | |
| Date: | 2003-11-15 16:02:27 | |
| From: | anonymous2 | |
| Looking at the quality and speed of Swing in 1.4.2 days and also looking at the speed of Swing applications like JBuilder or IDEA the Eclipse people and especially IBM should end their small mindedness. People often criticize Swing as being too complicate when compared to Cocoa or WindowsForms. Well they should take a look at SWT. This thing is already outdated. IBM please remove it. You strengthen only .NET with this fragmentation and you gain nothing. And start to think about a Swing implementation using a Java2D implemented on top of OpenGL (hardware accelerated). Isn’t it planned for 1.5 ?? | ||
Showing messages 1 through 2 of 2.
-
For the next version: port it to Swing
2003-11-17 05:09:46 neuro1 [Reply | View]
-
For the next version: port it to Swing
2003-11-17 14:05:37 anonymous2 [Reply | View]
Did you really try JBuilder (free personal edition available) or jEdit, which maybe a better choice if you come from XEmacs. :-) I do not want to argue which IDE has a better concept. (as I think jEdit, JSwat and ant are the very best environment ) but just want to stress that you cannot criticize neither speed nor memory consumption of Swing in JDK 1.4.2 as demonstrated by these two professionally done apps (just to give an example).
Swing has a good concept as it can be implemented on a high level (to shield you from the bugs of Win32) by drawing its own widgets. And if you do it clever, like Sun did in the latest version, you can make a perfect fake of the native widgets. They are pixels on the screen after all. Just name me a relevant difference of the Windows look and feel on XP to the native widgets!
Sorry, but I do not buy your statement, that a single customer jumps off because of Swing. Maybe they do because of bad GUI programming. I want to stress again, that the concept of Swing allows that Java uses OpenGL for all its drawing. By that way you get screen drawing with hardware acceleration. (M$ will “invent” this in .NET when Longhorn arrives in 2006). There is no reasonable way to think SWT can improve anything. And I bet M$ will find many opportunities to thwart it, if would get relevance, as it is so deeply enmeshed with native code. By its low level concept (hello again event loop, haven't seen you since the days I did Mac programming many years ago) you get in touch with unnecessary complexity and platform specific bugs (as I understood the reason for stopping AWT development). Swing only needs a window from the OS it runs on. This concept is hard to beat.
If you do a fork of a technology you should have very good reason. And there is no real case against Swing, apart from long standing convictions. They really need a review, as things improved meanwhile.
By the way, if IBM really want to improve things, it should grab the futuristic ideas from Cocoa not the API concepts of the past of GUI programming.
About SWT, we've been writing a couple of applications based on SWT and yes we did the absolutely right choice. Our competition is using .Net/C# solutions, most of our clients are still under Windows (for now:) and without SWT we could not stand a chance. Our applications are fast (I mean damn fast) and not very pretentious on resources (heck it even works reasonably on an Pentium 166 with 32 RAM) and absolutely look great, no wander out clients love them all with the full power of Java. Without SWT we would be out of business by now.
True do it is still work in progress so beware of bugs.