 |
 | Home | Contact Us | FAQ | Search & Site Map | Link to Us |
 | Sign In | Join | Other 45 Sites in Network |
Java Forum / General / January 2008SWT or Swing?
Hi all, I am about to start writing a UI application and am thinking about writing it in SWT. I have experience with Swing but never used SWT. I thought it might be good to get to know another technology. I am writing the application for a windows machine so portability is not a primary concern of mine. I have heard SWT is a bit windows focussed which for me is not an issue. With the improvements with Swing under Java 5 and 6 which technology does the group think I should use? Is speed much better in SWT when compared to Swing? Is SWT a steep learning curve? I belive they both paint on the event thread so bad programming in either would hit UI performance. TIA Richard > I am about to start writing a UI application and am thinking about > writing it in SWT. I have experience with Swing but never used SWT. I > thought it might be good to get to know another technology. Always good to know all the alternatives. > I am > writing the application for a windows machine so portability is not a > primary concern of mine. I have heard SWT is a bit windows focussed > which for me is not an issue. SWT uses native GUI components, but they are available for some other platforms than Windows. > With the improvements with Swing under Java 5 and 6 which technology > does the group think I should use? Is speed much better in SWT when > compared to Swing? As a product strategy, then I do not see much point in SWT. SUN works hard to improve Swing and my impression is that they keep improving it. > Is SWT a steep learning curve? Never tried it personally. But on my TODO list. > I belive they both > paint on the event thread so bad programming in either would hit UI > performance. That is generic for many GUI frameworks. It is the exact same issue in .NET win forms. Arne > richnjo...@gmail.com wrote: > > I am about to start writing a UI application and am thinking about [quoted text clipped - 32 lines] > > Arne Does this involve with any software? On Dec 29, 8:44 am, alexanderg...@gmail.com wrote: > > richnjo...@gmail.com wrote: > > > I am about to start writing a UI application and am thinking about [quoted text clipped - 36 lines] > > - Show quoted text - When writing an application therfore has to write with any software because an application example UI will install that program to help you torwards that domain name > When writing an application therfore has to write with any software > because an application example UI will install that program to help > you torwards that domain name What are you talking about?  SignatureLew > When writing an application therfore has to write with any software > because an application example UI will install that program to help > you torwards that domain name Ah, so you're the guy who writes those manuals. >> When writing an application therfore has to write with any software >> because an application example UI will install that program to help >> you torwards that domain name > > Ah, so you're the guy who writes those manuals. ROFL SignaturePosted via a free Usenet account from http://www.teranews.com richnjo...@gmail.com wrote: >>> Is SWT a steep learning curve? Arne Vajhøj wrote: >> Never tried it personally. But on my TODO list. >>> I belive they both >>> paint on the event thread so bad programming in either would hit UI >>> performance. Arne Vajhøj wrote: >> That is generic for many GUI frameworks. It is the exact same issue in >> .NET win forms. > Does this involve with any software? I'm sorry but your question is not clear. By "this" do you mean that you have to be careful about putting GUI actions on the GUI thread? What sort of "involvement" are you talking about? All software that interacts with a GUI has to be careful that GUI events only occur on the GUI thread. SignatureLew >With the improvements with Swing under Java 5 and 6 which technology >does the group think I should use? Is speed much better in SWT when >compared to Swing? Is SWT a steep learning curve? I belive they both >paint on the event thread so bad programming in either would hit UI >performance. In terms of complexity it goes AWT, SWT, Swing. SWT has better performance. However, if you want help, Swing in where everyone is. SignatureRoedy Green Canadian Mind Products The Java Glossary http://mindprod.com On 29 Dec, 21:50, Roedy Green <see_webs...@mindprod.com.invalid> wrote: > On Sat, 29 Dec 2007 07:18:38 -0800 (PST), richnjo...@gmail.com wrote, > quoted or indirectly quoted someone who said : [quoted text clipped - 11 lines] > Roedy Green Canadian Mind Products > The Java Glossaryhttp://mindprod.com Thanks for the replies. I know that SWT is available on other OS than Windows although I seem to remember James Gosling saying it was very windows centric. That was one of the criticisms he mentions if I remember correctly. Roedy: Does SWT have better performance that Swing after Java5. I know Sun spent a lot of time trying to improve its performance? I am not too bothered about individual tests for this but just a general feel. TIA >Roedy: Does SWT have better performance that Swing after Java5. I know >Sun spent a lot of time trying to improve its performance? I am not >too bothered about individual tests for this but just a general feel. IIRC the main SWT app in the universe is Eclipse. Nobody does a big app like that both ways for testing. Perhaps you will have to create some representative app to see if it worth the hassle. SWT came out before swing. So most of the impetus came from people eager for fine control when Swing was not available. It has withered somewhat since now Swing has even more elaborate functionality. Further video hardware is breathtaking compared with what we had only a few years ago. Once is it fast enough, further speed increases don't matter. About the only people who have an insatiable lust for video power are the gamers who want ever-higher resolution 3D renderings that move faster than the eye of my age can follow. Any Java App I have written, when run on my machine is fast enough that it feels effectively psychologically instantaneous. The delays are nearly all about the web. SignatureRoedy Green Canadian Mind Products The Java Glossary http://mindprod.com .... >Further video hardware is breathtaking compared with what we had only >a few years ago. Once is it fast enough, further speed increases >don't matter. About the only people who have an insatiable lust for >video power are the gamers who want ever-higher resolution 3D >renderings that move faster than the eye of my age can follow. Surely, at that stage the developers are overriding paint(Component) and the entire matter of widget toolkit (AWT/SWT/Swing) has become irrelevant, no? Either that or the app. leverages the power of JOGL, or Java3D, or (..insert other non Sun rendering APIs here) - which ends up in the 'same difference' for all 3 widget toolkits. SignatureAndrew Thompson http://www.physci.org/ > ... > [quoted text clipped - 14 lines] > > Message posted viahttp://www.javakb.com Roedy/Andrew Thanks for the replies. I use Eclipse and it does seem very responsive. It is interested that you said SWT was before Swing. I didnt realise that. It is also a very good point about graphics card/ processor speed - it will only get faster. I suppose both can be slow with bad programming but I think I will opt for Swing at the moment. Thanks! ... >...I suppose both can be slow >with bad programming .. <mock serious>Grasshopper, you have reached enlightenment.</mock serious> >..but I think I will opt for Swing at the moment. > >Thanks! You are welcome to any minor part I might have contributed. :-) SignatureAndrew Thompson http://www.physci.org/ richnjones@gmail.com wrote: >> ...I suppose both can be slow >> with bad programming .. > <mock serious>Grasshopper, you have reached > enlightenment.</mock serious> That's the key, really. Much of slowness in Java GUIs historically has been due to misunderstanding of threading issues, a matter common to all GUI libraries. More generally, GUIs suffer more from bad design and layout than they do from slow platforms. Many gurus have taught that algorithm affects speed more than anything else. It's all about the big O - n-squared is bad, n*log(n) better, log(n) great, and we adore the Holy Grail of O(1). Getting work off the EDT is the big secret to Swing, it seems. Getting graphics onto the EDT is the other one. SignatureLew > SWT came out before swing. So most of the impetus came from people > eager for fine control when Swing was not available. It has withered > somewhat since now Swing has even more elaborate functionality. I'm pretty sure that Swing came first (around 1998). The Eclipse site says that IBM considered using Swing for Eclipse and then decided to build SWT instead. The main reason not to use SWT is that it is neither part of the standard library nor pure Java, so you need to bundle it with your application (with different native libs for each platform that you deploy on). This may be worth the hassle if SWT gives your application something significant that Swing doesn't, but usually Swing is perfectly adequate. Dan. SignatureDaniel Dyer http://www.uncommons.org >> SWT came out before swing. So most of the impetus came from people >> eager for fine control when Swing was not available. It has withered >> somewhat since now Swing has even more elaborate functionality. > I'm pretty sure that Swing came first (around 1998). The Eclipse site > says that IBM considered using Swing for Eclipse and then decided to > build SWT instead. I believe it was in between. Swing was on the way, maybe even available in prototype, but some feared that it would be a complete failure. And, at the time, Swing wasn't part of standard Java, either. SignatureJohn W. Kennedy "But now is a new thing which is very old-- that the rich make themselves richer and not poorer, which is the true Gospel, for the poor's sake." -- Charles Williams. "Judgement at Chelmsford" > SWT came out before swing. Swing was released with Java 1.2 in 1998. SWT was made open source by IBM in 2001. It must have existed earlier, but I think that was when it became available for people to use. Arne On Dec 30, 5:59 am, richnjo...@gmail.com wrote: > On 29 Dec, 21:50, Roedy Green <see_webs...@mindprod.com.invalid> > wrote: [quoted text clipped - 19 lines] > windows centric. That was one of the criticisms he mentions if I > remember correctly. SWT is Windows-centric iff (sic) you are running it on a Windows platform. It is extremely Linux centric when you run on Linux, and it couldn't be any more Solaris-centric when you run it in Solaris. If you happen to be developing under the Mac, ... Oh, well, I ran out of OSs. You get the point. -Ramon > On Dec 30, 5:59 am, richnjo...@gmail.com wrote: > [quoted text clipped - 32 lines] > > -Ramon Well thanks for the replies. I am still unsure. I was swaying towards Swing but learning a new technology is tempting. The Eclipse RCP look like it might be worth a look... > Well thanks for the replies. I am still unsure. I was swaying towards > Swing but learning a new technology is tempting. The Eclipse RCP look > like it might be worth a look... NetBeans has a rich-client platform also. <http://www.netbeans.org/> SignatureLew richnjones@gmail.com schrieb: > Hi all, > [quoted text clipped - 13 lines] > TIA > Richard SWT alone has not a chance to stand any comparison with SWING .. so don't try to compare it to SWING this way.. If you want to learn a new Technology that might give you some help in developing an UI application try learning the eclipse RCP. It will give you a lot of help with common isuues.. branding, native starter, update functionality... --- One benefit over SWING imho with doing things in the wrong thread is that SWT will immediately throw an exception if you access ui elements from the wrong thread. On Dec 29, 11:18 am, richnjo...@gmail.com wrote: > Hi all, > [quoted text clipped - 13 lines] > TIA > Richard The widget that best displays and illustrates the difference between Swing and SWT is the File Chooser. Take a look at the "Open file..." menu item in Eclipse. It looks like the real thing, because it is the real thing. Unlike the Swing File Chooser you will be able to view the files by Thumbnails, Tiles, Icons, List and Details. http://www.eclipse.org/swt/ I have always used Swing (due to the Sun's gravitational influence and inertia) but I look forward switching to SWT in my next projects. Many users are surely turned off and consider Java an "inferior product" (as defined by economists) when they experience the crappy Swing emulation of a real native widget. -Ramon > Many users are surely turned off and consider Java an "inferior > product" (as defined by economists) when they experience the crappy > Swing emulation of a real native widget. This is not what makes Swing applications crappy, IMHO. It's when a Swing programmer does not properly work with the event dispatching thread, running long-running tasks on it. This causes the GUI to feel slow and unresponsive. I started learning SWT a while back - still have a book on it - and I just didn't find the same level of flexibility that Swing offers. > Many users are surely turned off and consider Java an "inferior > product" (as defined by economists) The term "inferior good" in economics means that demand decreases when income increases - it does not say anything about quality. The same term outside of economics usually mean poor quality. Arne > > Many users are surely turned off and consider Java an "inferior > > product" (as defined by economists) [quoted text clipped - 5 lines] > > Arne You are right in that the term is only marginally related to low quality, but to the *perception* of low quality. Some textbook examples are the Pentium that couldn't divide and the IBM Peanut with its atrocious keyboard (or the Ford Edsel). IBM was involved in both cases. Even after IBM was pressured by the press and customers into giving a free replacement keyboard, the first IBM home offering was dead, killed by its own progenitor. During the Pentium ordeal IBM had already morphed into the great (or at least wiser) company that it is today and offered to replace all the "broken" CPUs at its expense, which left Intel no option but oblige. The Swing designers and apologists are quick to blame its faults on "bad programmers" but they never say that the Swing successes are attributable to "good programmers" but to "good design on our part" instead. -Ramon > The Swing designers and apologists are quick to blame its faults on > "bad programmers" but they never say that the Swing successes are > attributable to "good programmers" but to "good design on our part" > instead. Citations? SignatureLew > > The Swing designers and apologists are quick to blame its faults on > > "bad programmers" but they never say that the Swing successes are [quoted text clipped - 5 lines] > -- > Lew Just go a couple of postings back in this thread, Lew: [Joe Attardi:] "This is not what makes Swing applications crappy, IMHO. It's when a Swing programmer does not properly work with the event dispatching thread, running long-running tasks on it. This causes the GUI to feel slow and unresponsive." I hope you are not going to ask me for citations about Java success being the inescapable result of "good design". Those citations are left as an exercise. :-) -Ramon ps: You may replace "Java" for "Swing" above. >>> The Swing designers and apologists are quick to blame its faults on >>> "bad programmers" but they never say that the Swing successes are [quoted text clipped - 20 lines] > > ps: You may replace "Java" for "Swing" above. You cite someone pointing out that crappy programming makes Swing programs bad, sure. But you haven't shown that Joe Attardi is either a Swing designer or a Swing apologist. (Let alone substituting "Java" for "Swing" above.) You failed to show that his comment isn't justified. You failed to show that he was "quick to blame its faults", both by not showing that it was Swing's faults that he was addressing, and that he was being "quick" instead of deliberating long and carefully. You also failed to show that this is a general pattern, instead citing only one person who is neither designer, nor apologist, nor incorrect in his assessment, nor quick to reach his conclusions. Your claim was general: that the designers (plural) and apologists (plural) do this behavior. Where is your evidence? Oh, you don't have any. Yes, I am going to ask you for citations about Java success being the inescapable result of "good design". You made an outrageous claim, and you leave it up to *us* "as an exercise"? Come on! You need to back up your own assertions with your own evidence. I assert that you won't find it. Now put up or shut up. SignatureLew > >>> The Swing designers and apologists are quick to blame its faults on > >>> "bad programmers" but they never say that the Swing successes are [quoted text clipped - 41 lines] > You need to back up your own assertions with your own evidence. I assert that > you won't find it. > Now put up or shut up. I'll take the second option. -RFH > the IBM Peanut with > its atrocious keyboard ...which happens to be the direct technological ancestor of every keyboard made today... SignatureJohn W. Kennedy "Only an idiot fights a war on two fronts. Only the heir to the throne of the kingdom of idiots would fight a war on twelve fronts" -- J. Michael Straczynski. "Babylon 5", "Ceremonies of Light and Dark" > > the IBM Peanut with > > its atrocious keyboard > > ...which happens to be the direct technological ancestor of every > keyboard made today... John: I am sure we are talking about two different artifacts. Circa 1981 IBM had a keyboard research facility in Zürich. At the time, the only devices which had keyboards were something really ancient. I bet several readers have heard of them, they were known as "typewriters", more specifically the "IBM Selectra". I hope I don't have to tell the story of the original IBM PC (Model 5150), it was based on the Intel 8088 chip and it came with a great, metallic, heavy keyboard. That was an *office* computer. The masses clamored for a home computer and IBM feared (they had no idea how accurate their worst nightmares were) that a decent home computer with a good KB would eat into their profits. Hence the IBM PC jr aka "peanut" was born. It had a plastic- but-decent, Zürich designed, *optional* keyboard, and one standard that was aptly named "the Cliclet keyboard" by those nasty magazine writers (I was proud to have the PC World first issue, it was stapled, several months later it was the thickest magazine ever put together by an US press). Allow me to place the keyboard in question in the ancestry, evolutionary terms that you introduced... Let's just say that the Chiclet KB was a wrong turn in the genealogical KB tree, and as expected, didn't leave any descendants. http://en.wikipedia.org/wiki/Chiclet_keyboard http://en.wikipedia.org/wiki/IBM_Personal_Computer -Ramon > I am sure we are talking about two different artifacts. Circa 1981 IBM > had a keyboard research facility in Zürich. At the time, the only > devices which had keyboards were something really ancient. I bet > several readers have heard of them, they were known as "typewriters", Sorry, but that's incorrect. I remember using a "glass teletype" terminal in 1972 on an ICL 1902S: it was a direct replacement for a KSR-33 teletype and was smaller, faster and silent. By the mid 70s all mainframes and minicomputers used terminals with electronic keyboards and cursor addressing on the display. Independent sources of generic terminals were well established by 1980. Manufacturers included Televideo, Wyse and DEC. Remember the VT-100? These were in use on a lot of microcomputers such as SWTPc, NorthStar and Imsai by 1978. The Apple II and Commodore PET had electronic keyboards built in. The only thing that IBM introduced was a keyboard design that reported key up/down and the "scan code" (i.e. key co-ordinate) to the computer for translation into ASCII. This translation and shift key handling stole CPU cycles but did allow the keyboard to generate EBCDIC as well as ASCII and/or handle non-English by simply swapping the key tops. All the earlier keyboards did the ASCII conversion and shift key handling etc internally in hardware before outputting ASCII codes. The cheaper ones output characters in parallel on a ribbon cable: more expensive detached keyboards used a UART to output a serial character complete with parity and stop bits. > I hope I don't have to tell the > story of the original IBM PC (Model 5150), it was based on the Intel > 8088 chip and it came with a great, metallic, heavy keyboard. That was > an *office* computer. I think they were more worried that it would displace their current small business computers. Using an 8088 with a low clock speed stopped the PC from being obviously faster and cheaper than their other small computers. By the time the PC-AT appeared (8086 at 8 MHz) the Baby/36 program (a System/36 software emulation) could run S/36 software faster than any model in that range. I knew people who used Baby/36 as their program development system because it was cheaper, faster and more convenient than using a real S/36. In the UK you almost never saw an IBM PC used at home: people were buying PC clones instead due to the price difference. IBM kit started appearing at home only when used PCs came on the market as companies upgraded. Signaturemartin@ | Martin Gregorie gregorie. | Essex, UK org | > I am sure we are talking about two different artifacts. Circa 1981 IBM > had a keyboard research facility in Zürich. At the time, the only > devices which had keyboards were something really ancient. I bet > several readers have heard of them, they were known as "typewriters", > more specifically the "IBM Selectra". ???? Ever heard about VT52 and VT100 ? They had a keyboard. And I would not call them typewriters. Arne >> I am sure we are talking about two different artifacts. Circa 1981 IBM >> had a keyboard research facility in Zürich. At the time, the only [quoted text clipped - 7 lines] > > They had a keyboard. And I would not call them typewriters. Not to mention, from IBM alone, the entire 3270 series, the entire 5150 series, the 3767 and 3770 series, the 129 keypunch, the 50 tape inscriber, and other devices. SignatureJohn W. Kennedy "...when you're trying to build a house of cards, the last thing you should do is blow hard and wave your hands like a madman." -- Rupert Goodwins >> They had a keyboard. And I would not call them typewriters. > > Not to mention, from IBM alone, the entire 3270 series ... which, IMHO, had the best typing feel ever. And they could double as a seriously weighty club when one is having a BOFH moment. :-) sherm-- SignatureWV News, Blogging, and Discussion: http://wv-www.com Cocoa programming in Perl: http://camelbones.sourceforge.net >>> They had a keyboard. And I would not call them typewriters. >> Not to mention, from IBM alone, the entire 3270 series > > .... which, IMHO, had the best typing feel ever. Not the early 3277 and 3275 units with keyboards subcontracted by Microswitch. They were bleedin' awful -- dead touch, and an unreliable electromechanical clicker device. But IBM soon replaced 'em. I much preferred the prior generation -- the 2260 -- with a keyboard that had the splendid feature that, if the keyboard was software-locked, the cursor was in a non-input area, or, for whatever reason, the keyboard was inhibited, the keys literally could not be pressed, except for what you needed to get out of the jam. It was also literally impossible to press two keys at once (except, obviously, shift keys), but they rolled from one key to another sweetly while typing at speed. But those keyboards cost a /lot/ of money, and that was in 1960s dollars, too. In those days, though, almost all IBM equipment was leased anyway. SignatureJohn W. Kennedy "Sweet, was Christ crucified to create this chat?" -- Charles Williams. "Judgement at Chelmsford" Ramon F Herrera wrote: >> Many users are surely turned off and consider Java an "inferior >> product" (as defined by economists) > The term "inferior good" in economics means that demand > decreases when income increases - it does not say anything > about quality. The same term outside of economics usually > mean poor quality. >> when they experience the crappy >> Swing emulation of a real native widget. Which leaves the questions of how many users actually are turned off by Swing widgets, and whether that's actually Swing's fault. Evidence? SignatureLew > Ramon F Herrera wrote: > >> Many users are surely turned off and consider Java an "inferior [quoted text clipped - 13 lines] > -- > Lew Lew: Let's make some narrowing assumptions, as I have nothing to say about threads and techie things of that nature. My issue is the user experience, more specifically (and unfortunately because it is a fact of life in more than 90% of Java desktop usage) on Windows. Many of you out there jump to the defense of Java (or parts of it, such as Swing) from the vantage point of never having the misfortune of having to use Windows. I wish I could have that degree of freedom. The bone I have to pick is Swing's emulated nature vs. SWT's use of "The Real Thing (*)" as illustrated here: http://www.eclipse.org/swt Further narrowing: I have nothing to say about most of the widgets which look and feel the same under Swing or under SWT. The best possible discriminating widget is the File Chooser. In SWT we have always been able to get an FC dialog and create new folders, list the directory's contents from many views, etc. SWT designers left that functionality where it belongs, in Redmond, WA, I am resigned to accept. The "Open File..." in Eclipse has always behaved as it should and it is as fast and responsive as it can possibly be. Compare with NetBeans's "Open File...". In NB 5.5 you could create a folder AS LONG as its name was "New Folder". The NB JFileChooser is custom-made, and I have never seen it anywhere else. This is a cause of *confusion* leading to the *perception* of Java being on the fringes, an "inferior product" like the IBM peanut or the Pentium that couldn't divide. I have a folder with several hundred MP3 songs, and it takes the Windows *native* shell a long while to list (and re-list by header clicking) all the song titles, durations, etc. I would never be so cruel as to assign the job of listing those songs to Swing. Eclipse and SWT perform such job to my complete satisfaction. -Ramon (*) (TM) Coca-Cola, with apologies. >> Ramon F Herrera wrote: >>>> Many users are surely turned off and consider Java an "inferior [quoted text clipped - 23 lines] > such as Swing) from the vantage point of never having the misfortune > of having to use Windows. I wish I could have that degree of freedom. I use Java in Windows all the time. I'm sure "many of us out here" also do - it is the single largest-selling OS, hm? My experience isn't yours. But that's beside the point. You made a claim about "many users" and backed it up with your own personal experience and judgments. Where's your evidence about "many" users? SignatureLew > The bone I have to pick is Swing's emulated nature vs. SWT's use of > "The Real Thing (*)" as illustrated here: Does this mean that SWT only works in a Windows environment? Signaturemartin@ | Martin Gregorie gregorie. | Essex, UK org | Martin Gregorie schrieb: >> The bone I have to pick is Swing's emulated nature vs. SWT's use of >> "The Real Thing (*)" as illustrated here: >> > Does this mean that SWT only works in a Windows environment? no SWT is just limited to a few diffrent OS/UI-managers.. > Martin Gregorie schrieb: >>> The bone I have to pick is Swing's emulated nature vs. SWT's use of [quoted text clipped - 3 lines] > > no SWT is just limited to a few diffrent OS/UI-managers.. Any idea which ones? IOW, is it really worth using? Signaturemartin@ | Martin Gregorie gregorie. | Essex, UK org | >> Martin Gregorie schrieb: >>>> The bone I have to pick is Swing's emulated nature vs. SWT's use of [quoted text clipped - 5 lines] > > Any idea which ones? Win/x86 Linux/x86 Linux/PPC Solaris/SPARC HP-UX/IA-64 AIX/PPC MacOS X/PPC & x86 Arne Arne Vajhøj schrieb: >>> Martin Gregorie schrieb: >>>>> The bone I have to pick is Swing's emulated nature vs. SWT's use of [quoted text clipped - 15 lines] > > Arne on unix GTK and Motif are supported... though one should be aware that the combination of Solaris and Motif are no longer supported by the eclipse RCP .. so if you want to use it with this you will be more limited. Free MagazinesGet these publications absolutely FREE for up to 12 months. There are no hidden fees and no obligation. Simply choose a title, complete the application form and submit it. Read more ...
|
Reply to this Message
|
 Sign In
 Join
 My Latest Posts My Monitored Threads My Blog My Photo Gallery My Profile My Homepage Start New Thread Enable EMail Alerts Rate this Thread
|
©2008 Advenet LLC Privacy Policy - Terms of Use
This website includes both content owned or controlled by Advenet as well as content owned or controlled by third parties.