Want to read Slashdot from your mobile device? Point it at m.slashdot.org and keep reading!

 


Forgot your password?
Close
typodupeerror
×
Java Programming

IBM releases VisualAge for Java for Linux 3.0 77

Posted by Hemos from the coming-to-linux-first dept.
Quite a number of you have written in with the word that IBM has released Visual Age for Linux v3.0. As well, a number of you noted that v3.0 came out for Linux before certain other OSes *grin*. As for obtaining it, you can't get it for free but you can come pretty darn close - get on VADD for $50 to get VAJ for Linux. Happy Coding.
This discussion has been archived. No new comments can be posted.

IBM releases VisualAge for Java for Linux 3.0 More

IBM releases VisualAge for Java for Linux 3.0

Comments Filter:
  • This is truly excellent. Not only does Linux now have an excellent "RAD" tool, but it also has an excellent Java tool. Think we'll be seeing a lot more Java tools geared to the Linux-world? Can Freshmeat handle the stress? Has the whole world gone mad?? Tune in next week...

    -----------

    "You can't shake the Devil's hand and say you're only kidding."

  • I'm looking at starting to play about with Java and being used to Visual Studio on Win32 (spit) I was wondering if this sort of thing would suit my needs. If anyone has experience with such environments I'd be interested in their comments.
  • > I've used this product under Windows NT and it sucked,

    Name one product which doesn't suck under Windows NT. Around here we've used Visual Cafe, and it sucked much worse than the Visual Age that we use now

    > I don't see any reason why version 3.0 should be any better under Linux.

    I do see a reason: Linux. At least, under Linux, VA won't have to rebuild its workspace whenever we switch from summer time to winter time and vice versa.

  • I don't think VAJ on Windows is any more bloated than JB, VC or Net Beans. It was, but they've caught up now ;-)

    It is the best of the Java IDEs IMHO.

    Tom

  • I've used Visual Age for C++ under NT (and VA for Smalltalk!) and I would rate it OK. As someone mentioned earlier, it's a little bloated and slow at times. For Java, I prefer Visual Cafe by Symantec...Now if only THEY would port to Linux.

    IDE's are fine, I'm just wondering when we will get a decent JDK port for Linux..No use having a good IDE if your JDK is too slow, or hogs too much memory or doesn't run from a common place on the file system

  • I've used VisualAge on NT too, and I didn't like it... too much of a feature bloat, with a clumsy interface. I suppose hardcore people like me are more comfortable with the command-line JDK compiler...

    On the other hand, it's good that these "user-friendlier" apps are being developed for Linux. At least Linux is slowly making inroads to being a desktop environment for non-techs. This is good, and should be encouraged, because Linux is about choice, and there should be a wide range of different choices that suit different people, eg., command-line interface for techies like me, and "dumber" GUI apps for people who don't want to bother with learning arcane commands.

    Adding more choice is always a good thing, because there is no such thing as a one-size-fits-all when it comes to systems. (It's more of a one-size-fits-few.)

  • I'm not too interested in these news at all. In ways, I think that these proprietary vendors harm the future development of our beloved system.
    So why is this? Primarily because I see that the same problems that exists on other systems are now coming to ours. Soon enough, what we have won't be THAT much different. Perhaps the most harmful trend is when people make copies of these proprietary programs to share with their friends. Suddenly the industry will turn to them and call them pirates. In our community, we try to teach people that sharing software with your friends is the right thing to do. Now, we'll have hunters doing their best to enforce upon us the idea that this is as bad as murdering and robbing people on high seas.
  • I've always liked VisualAge better than VS. The amount of coding I've done for Windows is very limited, but VisualAge (for C++, haven't really used VAJ yet) was a lot more intuitive. At least for me. At any rate, I'm happy there's a decent JDE for Linux now.

    numb

    PS> Please feel free to correct me if there are and decent java development environments for Linux that I'm not aware of.
  • The page linked to for purchasing had this tidbit:
    • "Operating System: Windows 95 or NT 4.0 (service pack 3); OS/2 Warp Version 4; Server Version 4 or Advanced Server Version 4"

    I'm not seeing linux there ;)

  • I downloaded the BETAs and it was quite hard to get used to. Big ugly buttons, and I didn't like the way each method was edited in it's own window (didn't MS give that up after QB :P)...i didn't bother to check to see if it could change (you can actually do thatin VB).

    I much prefer J++ as an IDE, even if it's far far behind cause it has a very nice editor.
    I've managed to get Swing and RMI working with J++ so it keeps me content :) Very fast compiler too.

    I think Microsoft might lose part (well more than is already lot) of me to Linux if they started porting Visual Studio & IE5 to linux :).
  • I used one of the beta's of VA and it was ok, but I just can't get over not being able to see all my code at once.

    I've been using CodeWarror java for Solaris at work, and it works fine (I still use emacs for editing though... :) It lets me look at all the source code I want, is extremeley easy to dump a ton of java code into and just start using it, and the debugger actually works. (I'm not sure how well the debugger works in VA, as never got to do any serious testing with it)
  • I'm happy to see that IBM decided to take the step and release VAJ 3.0 for Linux. However, the only benefit in this whole scenario is the recognition of Linux. I've worked with VAJ 2.0 NT for several months and this whole IDE concept is flawed. IBM has too many proprietary classes for it to be flexible. The IDE is bloated as hell, its slow and basically annoying.

  • Re:Is this package any good? (Score:3)

    by hey! ( 33014 ) on Friday November 12, 1999 @03:06AM (#1539745) Homepage Journal
    Pretty darn good, if you ask me.

    I've never tried Symantec's offerings, but VisualAge for Java is head and shoulders better than Sybase's PowerJ.

    By rights this ought to be a VB and PowerBuilder killer. Both VB an PB have very cluttered IDE's and languages, whereas the Visual Age IDE is smaller and more logical. Java is also truly object oriented as opposed to VB, and it takes care of garbage collection which has to be done manually by PB.

    The only reason I'm not more excited by this is that I'm getting really sick of relying on tools vendors for support and upgrades. If there's a bug that affects you, you end up twisting slowly, slowly in the wind while some PHB decides on new "features" to put into the product. If IBM open sourced this, I'd be in heaven.



  • I've used VisualAge under Win95 & NT, and it was a little sluggish (mind you, it was on a pentium 200, so the machine itself may be to blame). Does anyone know if the linux port is any zippier?

    Kawa is a much better IDE for small projects, VisualAge is kind of intended for a team building a large project. I heard a rumour about linux Kawa port a few months ago, but nothing seems to have surfaced.

    Dana
  • My company uses VisualAge almost exclusively for Java development. I happen to like it a lot. Here are some pros and cons:
    Pros:
    1) Object Browser code viewer encourages good object-oriented development.
    2) Powerful interactive compiler, which means new code you write is parsed directly into the running application vs haveing to edit-recompile-restart an application to see changes you have made in code.
    3) Excellent runtime inspector and debugger - far more advanced than what I have seen from the competition.
    4) Scrapbook to execute arbitrary snippets of code easily.

    Cons:
    1) Uses an old (1.1) JDK (unless this changed in the 3.0 version). Not a big issue for me because I want my Java code to be availible on many different platforms, and "Java 2" apps are restricted to Windows and Solaris (yes, I know there is a beta 1.2 for Linux but we can't require our customers to use beta software, and anyway we need to support Macintosh and several other UNIXes).
    2) Object Browser interface will be unfarmiliar to someone who is used to other tools e.g. Visual Studio that present a file based vs object based view of the code. Everyone I know has complained about this at first, and later decided that it is really better this way (but maybe that is because only people who are inclined to liking it continue to use the tool).
    3) Very big footprint; requires a fairly powerful computer to run it (I would say *at least* 64 megs of RAM, 96 for a comfortable experience).
    4) GUI builder has some bugs in it.

    hope this helps.

  • disagreement is good (Score:3)

    by The G ( 7787 ) on Friday November 12, 1999 @03:13AM (#1539749)
    One poster likes VisualAge, another thinks it sucks.

    One poster likes Visual Cafe, another thinks it sucks.

    Some folks think emacs is a great IDE, some think these folks are on GnuCrack.

    The point is, the more IDEs we have for Linux, the more of these people can work on Linux, and the better off Linux is.

    Diversity of uses and users is Linux's strengh; diversity of apps is what we need to get even stronger. Even if you think that newest FooBar Visual Widget Toolkit +- sucks, it's making Linux that much more appealing to people who don't think that it sucks, and we all get a bit stronger.

    Freedom of choice. It's that good thing, remember?
    --G
  • Sorry, the IDE won't make your code better-all it can do is make you more productive.
    Seriously, this is a very difficult problem in principal for Java, because no two VMs actually exectute the same Java code the exact same way.
    So, if I'm developing in VisualAge for Java (or JDK for that matter), and the code that I develop is being executed Netscape's (or Internet Explorer's, or Java Plugins)'vm, there could very well be a difference that could cause a problem, or reveal a bug in your code that was not encountered in the development process.
    No substitute for QA, I'm afraid.
    No big application is portable in the absolute sense of the word, rather it is a matter of degrees.
  • Maybe you can suggest a good book to read about VAJ then. I just know at the place I worked at we had no support, no books and they just said "here is IBM's main help number for this, call them because we can't afford books." Between trying to get the Servlet builder to work properly ( as opposed to just writing out the 200 lines of code for the silly thing in the first place ) and inconsitancies between the IDE and what was actually being output I was near pulling my hair out.

    As with most things though, it was probably the bad experience and that I'm used to having the control of coding everything by hand. It is more than likely I just didn't know how to use the tool, so if it is so good where should I look for information? I really would be willing to try it again, I just don't see the point in breaking my back to learn it.

    Cheers...
  • To be honest, I'm kind of confused as to how a proprietary piece of software will damage Linux.

    You might argue that the danger is that product 'XYZ' will start to be used by a large percentage of the Linux population, and become something of a standard in itself. However, even then, there's nothing that's really forcing you to use product 'XYZ', except when trading information, data or files with an 'XYZ' user. And even then, the very nature of the Linux community and code is such that open-source software will be able to compete and perhaps outdo that product.

    A good example would be Photoshop -vs- The Gimp. Imagine if today, Adobe shocked us all and announced that they had a Linux port for a mere $100. Would you see all development on the Gimp cease? Hell no, if anything the competition and chance to show up the big boys on the same platform would drive our Gimp developers to even better heights.

    Finally, as to software piracy, I'd say that Linux isn't about trying to "teach people that sharing software with your friends is the right thing to do" but more towards providing a damn fine, customizable platform in which free GPL software and non-free proprietory software can both exist, allowing the users to decide which to use...

  • The link provided points to VA 2.0 for 50 bucks...
    Proper link: VADD - How to buy [ibm.com] it seems to be $149 minimum.

  • Great news, but... (Score:3)

    by snoopdave ( 21953 ) on Friday November 12, 1999 @03:40AM (#1539756) Homepage

    I am an IDE type of guy. I love class browsers, integrated debuggers and GUI designers. Unfortunately, I have not yet found a Java IDE that works well enough to replace my favorite editor (CodeWright or vi) and command line tools (make, javac and jdb) for daily use.

    Why is this? I have tried most of the major Java IDEs out there including JBuilder, Visual Age, Visual Cafe, Netbeans and even Visual J++. Most of the IDEs have some nice features but all of them are bloated memory hogs. Maybe faster CPUs and cheaper memory will make them usable someday.

    That said, I think it is great that IBM is supporting Java on Linux. To support the Visual Age and Websphere products, IBM is providing their fast and stable Java VM free-of-charge under Linux. Can't complain about that!

  • Perhaps the most harmful trend is when people make copies of these proprietary programs to share with their friends. Suddenly the industry will turn to them and call them pirates.In our community, we try to teach people that sharing software with your friends is the right thing to do.

    Sure, but I hope we also acknowledge that if we're not allowed (by licence agreements we've agreed to) to share things, then we shouldn't. IBM have every right not to make their software free etc, and have every right to call us pirates if we break licences we agree to.

    I hope the open source community's attitude is, "Look how great it is to let people share!" rather than "We're going to steal your software if you don't agree with our principles."

    If you disagree with the licence that software comes with, don't use it - it's (usually) as simple as that, as far as I'm concerned. Things like DVD encryption cracking may be exceptions - I'm afraid I don't know enough about that to comment.

    Jon

  • Books on VAJ, Servlet Builder, etc, (Score:1)

    by Anonymous Coward
    I think that VAJ is the best Java IDE, and I've tried the others. As for books on VAJ, there are many and you can get them for free in PDF format at http://www.redbooks.ibm.com/rbol.html, on the Visual Age Developer's Domain site, and various other IBM websites. I have MANY binders at work filled with these manuals and most are well written and tell you what you need to know.

    As for Servlet Builder, I don't recommend using it and IBM doesn't strongly recommend it either. It's one of those ideas that sounded good in theory but didn't work out in practice. VAJ is still a terrific tool for writing servlets by hand and has by far the best test environment for servlets going.
  • I think the $50 is for the competitive upgrade price, the $149 is the full retail price.

  • > Soon enough, what we have won't be THAT much different.

    Then it's up to you to make it different. That's the beauty of Open Source Software. You have the liberty of starting a project that will offer something different. JUST TAKE THAT OPPORTUNITY!

  • Then why isn't Linux listed in the system requirements section?
    Besides this, I don't have VA 2.0 for Microsoft 'platforms' (...) so, as probably most of you, I'm more interested in the retail price.
  • A Kawa port would be great, but I don't think Tek-Tools are very interested in doing it. I have written to them and asked for a port, but they haven't replied. There was also a thread on their Kawa support news group where several people expressed their interest in such a port, but the Tek-Tools people didn't feel like answering that either.

    Anyway, Emacs with color coding of the Java code is all you really need, isn't it?

  • Emacs (Score:3)

    by DanaL ( 66515 ) on Friday November 12, 1999 @04:28AM (#1539764)
    I kind of like VisualAge, though. Features like compilation-as-you-go, the respository and auto-versioning are kind of nice. Maintaining different versions has saved my life on projects in the past.

    I imagine you could probably makes Emacs do that, but having to learn elisp to add functionality to my java development environment just seems wrong somehow :)

    Dana
  • "An Entry Edition of Version 3.0 for Linux will be available at no charge from VADD late in November."

    This continues IBM's tradition of offering entry-level versions of VisualAge (for Java and Smalltalk, anyway) at no charge (free beer).

    Fascinating strategy. The traditional route is to sell your tools to schools, and get kids hooked there. IBM may be doing that, but they're also offering it to us who are picking up these technologies on our own.
  • unless the Linux version is of lower quality. That said, the more commercial tools that are released for Linux, the easier it is to sell to corporate suits as a viable alternative. The key here is assimilation through patience. Very few corporations will be willing to buy into the entire free software concept up front. But many would consider installing Linux if their favorite commercial tools were available for it. Once a free operating system is installed, that opens the door for free alternatives to commercial products. All you have to do is have a quality free alternative ready to offer when a commercial product has a show-stopper bug or an unacceptable timeline for a needed upgrade. One would hope, over time, an obvious pattern would emerge ;).

  • Is it just me, or is anyone else annoyed by the way that when a commercial product is released with "Linux support" the implicit assumption is "x86 Linux only"? Since I only run non-x86 Linux (primarily LinuxPPC [linuxppc.com], although also AlphaLinux [alphalinux.org]), I'm more than a little irked at that bias. Given the fact that Linux is a multi-architecture operating system, I'm tempted to say it amounts to false advertising.

    Will software vendors ever wise up and say "x86 Linux" when that's all they mean? Or are we doomed to be stuck with the assumtion the "Linux" means "x86 only" unless otherwise specified as long as x86 is the dominant CPU architecture?

  • I evaluated VAJ when the first beta came out, and it was already darn good. OK, it takes about 15 secs to start up, but Visual Studio is no speed demon in that department either. Once it's up it's solid as a rock - it always has been, right from the first beta.

    VAJ can do something that's worth the price of admission alone: you can recompile parts of a program while it's running and the changes take effect immediately. Cutting incremental development time down to practically nothing. This feature is very easy to get addicted to, believe me.

    VAJ uses native UI components, so the UI's run fast. Well, there are a whole lot of things I like about it, too many to mention. There was a very noticable speed up between version 1 and 2. I haven't tried version 3 yet, but I'm looking forward to it. If I had an enterprise Java project to do, I'd feel comfortable using this tool, no doubt about it.

    As far as using it to get your feet wet, it's not such a bad idea. The ramp-up time is pretty good - I got my first sample app running in about 10 mintutes, following the wizard-kind-of-thingy that comes up when you first start it. The nice thing about the initial tutorial is - it holds your hand, but not so much that you don't have to think. Very fast ramp-up.

    Oh - another thing that is really cool - where's the compile option? There is none. To compile a program, or part of one, you just edit and save the file. Saving the file compiles it. Wow.

  • I also have a lot more trust in command-line tools than GUIs. There are few automated tools that don't make me shiver when I click on a button. But recently, I had to integrate an existing database schema with Container Managed EJBeans in WebSphere, and my opinion of VisualAge for Java changed totally. Without VAJava, the mapping would have been nearly impossible or would have taken hours for every single bean. Their persistency tool for EJBs, at least, does a wonderful job and I would not consider deploying complex Entity beans without VisualAge.

    I however hate the way they split your object view into various method windows. In summary, I only use VAJava for deployment, not development.

    Kk.

  • Speaking of the JDK, according to this (from the faq) link:

    http://www7.software.ibm.com/vad.nsf/Data/Docume nt0553?OpenDocument&p=1&BCT=1&Footer=1

    they 'support' jdk 1.2. Does this mean they've actually provided a full java 2 platform for linux?

    If so that would be great, if not I'm not going to bother with it. One of the previous version's biggest annoyances was that you were stuck with jdk 1.1. Which is why it quickly found it's way to the garbage can (sorry, no /dev/null on winnt). I was impressed with it's way of connecting beans though.
  • Perhaps the development wouldn't stop on the Gimp if adobe released photoshop but it would cause most of its more serious users to abandon it. 100$ is nothing for a tool you use to make a living on.

    The main motivation for creating the Gimp was because there wasn't a good pixel editing program on Linux. If that reason is taken away, what exactly is the motivation for creating it? Providing a free alternative?

    Propietary software won't damage linux but will establish it as a mainstream OS. I don't see that as a wrong thing.

    In fact the coming of/need for propietary software signals an important thing: the partial failure of OSS. Apparently OSS can't serve the need of normal users in providing programs to run on the OS (wordprocessors, programming environments, spreadsheets, browsers). Though there are OSS variants of all the things I mentioned they all have propietary counterparts that are better.
    If you want to create documents you can use TeX, if you want a handy tool to use to create your document you'll need to obtain a propietary one.

    I realize I'm pissing of a lot of people here by claiming that OSS partially failed. But I fail to see its success on the desktop (I'm not talking about server applications). Sure there are a lot of projects (the gimp, K office, mozilla) but they are all unproven (unlike linux itself and the lower level stuff that runs on it).

  • Re:Is VisualAge that great?? - duh, yes. (Score:1)

    by Anonymous Coward
    It's fine and dandy to use a command-line interface when you're building an application with a couple of hundred lines of code. Now move up to the big leagues where multiple programmers are working in tandem on a project with thousands of lines of code and the need for repository management- VAJava becomes a necessity. I hope you aren't insinuating that VAJava is a "dumb" GUI app. I've seen far dumber GUI IDE's. Also, being hardcore doesn't translate into being anti-GUI. GUI apps or OSs at their best mean better productivity for the user. Being "command-line" for the sake of feeling like a true techie is strange. Use the best tool for the job, GUI or command-line based. For large, enterprise Java applications, VAJava is that tool.
  • IBM VisualAge Java 3.0 is awesome. I have used 3 Java IDEs - VAJ, JBuiler, and MS Visual J++. VJ++ simply stinks, so I will not waste time on it. JBuilder is nice, but doesn't do anything special. IBM VAJ gives you a VERY nice GUI builder which requires NO programming to create a fully functional program, is by far the most OO environment, has an awesome debugger, and is available on more platforms than any other IDE (Win32, Linux, OS/2, OS/390, AIX, etc.). In the Win32 version they are supposed to have JDK 1.2 support, but I don't think that happened in the Linux version.
  • 1.0 Prerequisites

    This edition of VisualAge for Java, has the following hardware and software prerequisites:

    * Linux - RedHat 6.0, Caldera 2.2 or SuSE 6.1
    * TCP/IP installed and configured
    * Pentium (R) processor or higher recommended
    * SVGA (800x600) display or higher
    * 120 MB disk space (minimum)
    * Frames-capable browser, such as Netscape Navigator Version 4.04 or higher. Recommend browser is Netscape Navigator 4.6
    * Java Development Kit (JDK) (TM) 1.1.7 for deploying all applications

    #include "disclaim.h"
    "All the best people in life seem to like LINUX." - Steve Wozniak
  • I realize I'm pissing of a lot of people here by claiming that OSS partially failed. But I fail to see its success on the desktop (I'm not talking about server applications). Sure there are a lot of projects (the gimp, K office, mozilla) but they are all unproven (unlike linux itself and the lower level stuff that runs on it).

    You're contradicting yourself here. By saying they are "unproven", you're admitting they have yet to fail. Your point, at best, can be summarized as "OSS has yet to succeed." You would need to go much further to establish that it has failed or will fail.

    For that matter, it makes no sense whatsoever to describe OSS has being a success or a failure without defining "successful at what" or "failed to do what". I find it funny to hear people talk about what Linux has to do to become successful when it is already wildly successful (a far greater success than I ever imagined it to be when I first started using it).

    If by failed you mean "OSS failed to replace all proprietary software on the majority of computers by Nov 1, 1999" you are indeed correct, and your statement is undeniably true. But it's not a very significant truth.

    If you mean to suggest "OSS will fail to replace all propriety software on the majority of ordinary users' desktop computers, ever" that's debateable, but I believe you're correct. However, I've never thought the point was to replace 100% of all proprietary software on all people's computer. Rather, it's to give people the option to run that way if they so choose. I don't want to be forced to buy a $100 imaging program because I have no alternatives. I do want to be able to buy a $100 imaging program is I decide I want to! If your vision of OSS is to drive such proprietary alternatives out of existence, I must say I oppose your vision of OSS. I still want to have that choice, I just want make sure it is a choice and not a necessity.

    --

  • Simplicity [datarepresentations.com] by Data Representations is a Java development environment written in Java. Not only that, but it includes its own JVM, also written in Java. What does this mean? It means that it can run on any Java platform (like OS/2, Mac, and non-x86 Linux).

    It also has one very cool feature: it can actually execute your program while your editing it - the running program gets updated in real time as you type it in! It's freakin' amazing! If your app displays a dialog box with some text in, and you edit the corresponding line of code, the dialog box changes while your editing that code, character by character!

    Not only that, but the company really likes alternate computing platforms. They had a booth at Warpstock [warpstock.org] a few weeks ago, and they had a PC running OS/2 and an iMac, both running Simplicity.

    I've been programming for over 20 years, and very little surprises me these days. But when I saw the Simplicity demo last year, my jaw dropped. It is that cool!.

  • I hate MS as much as anyone, but I've got to admit that J++ is the best Java IDE that I've used so far. From what I know about VAJ, it keeps your code in its own repository instead of just keeping your source in regular files, and that make me...uneasy. I prefer to keep my project in plain text files so that I can edit them with anything. (btw, if I'm totally wrong about VAJ, please let me know...) J++ has a nice tree structure for files and objects, and the auto-complete of methods means I don't have to reach for the manual or load up javadoc if I forget the api. The auto indent and the different color for javadoc comments is also a nice touch.

    Then again, the only reason why I can still use it is because we're doing 1.1 development. As we all know, the chances of J++ coming up with a 1.2 version is quite slim. I can also get Swing to work with it, but it does not work the same as Sun's or IBM's.

    I know that other IDE's exist that have some of these features, but I have yet to see one that does it as well. JBulder comes in a close second, but it still feels a little more bloated than J++.

  • I'd have to tend to agree with you on this. I'm not familiar with this latest VAJ, because of all the bad things that were in previous versions i stopped using it.
    There have been some people making comments in threads on this that VAJ is the way to go for large projects, i'd have to argue against this. VAJ generates code that is damn near unreadable by anything other than VAJ, so if you really are making a LARGE piece of software, something that is going to be around for years, you better make sure you keep an old copy of VAJ around if you ever need to edit anything. This generated code also makes a real hassle if you plan to try and make the software open source--the only way to read your open source software is with a propietary IDE.
  • Oh - another thing that is really cool - where's the compile option? There is none. To compile a program, or part of one, you just edit and save the file. Saving the file compiles it. Wow.

    Yikes! You can turn that off, right? I habitually save every time I pause to think for a moment, which is frequently right in the middle of writing something such that if a compile were attempted at that point it would produce a large number of errors. Not that I would mind that as much as it wasting time running the compiler and chewing up processor cycles and doing whatever else it does (popping up and output/error list, for example). This "feature" alone would stop me from using this product...

    --

  • Some folks think emacs is a great IDE, some think these folks are on GnuCrack.

    These are, of course, not incompatible statements. Just because we're on GnuCrack doesn't mean emacs isn't a great IDE.

    Besides, that's unfair. You make it sound like emacs is an addiction. It is not.

    It's more of an obsession... :-)

    --

  • VisualAge seems to be The One for the beginner programmers. It seems to help you to organize all your stuff, and do half of the work for you, and keep it neat and clean all the time.

    But, as you start having a clue, or even worse, see it first time when you do have a clue, unpleasant surprises await you down the road.

    Disclaimer: all the facts below are based on my experience with not the latest version of VA, your mileage may vary.

    1. VA doesn't let you control your source code. It enforces its own understanding of how it should look like, specifically:

    - Own indentation conventions. It does let you specify the tab size, but all effect of that option is how the code *looks* within VA. If you export your code, it reformats it according to its own preferences, and the version I used didn't care about Javasoft coding standard and just blindly set the indentation to one tab (8 characters), even if you specified the tab size to be 4.

    - Method ordering. Sorted alphabetically, which is incredibly stupid. My own preference is to group the methods by relevance. Worse than that, when I imported my code into VA, it rearranged everything.

    - Autogenerated javadoc comments for the method signatures. Like, @param p java.lang.String, as if I'm blind and don't see it myself. If I want comments, I put them there, if I don't, I might rely on the fact that javadoc makes a reference to the parent method.

    Consequence: the code you produce is not standard, which makes cooperative development difficult. Someone else's code you modify gets screwed up, see below why it's bad.

    2. VA provides a poor cooperation with the version control systems. It may be possible, but within a time I could afford to dedicate to that task, it wasn't possible to make it work with CVS, which is a de-facto standard for UNIX. Plus, the abovementioned source code management completely screws up the diffs. And God forbid if your team has some people working with VA, and some other who don't - VA forces everyone to behave as it wants. Classical "Vendor Lock-Up" antipattern.

    3. VA provides poor runtime support. The version I used didn't even produce the line numbers in the exception trace. The reason is understandable - it doesn't have the line number concept as such due to its way of maintaining the code, but how do I care? I want to know where it happened. And don't tell me I can use the debugger, because there's no intelligent way to use it in the servlet development in particular and any heavy multithreaded system development in general.

    I could ramble on and on, but these facts alone made me deeply despise the VA and I recommended my company to ban it completely.

    Of cource, IDE is great for the completely clueless because it allows to replace the "educated decision" pattern to be replaced by "educated guessing", but there's a danger of locking up the mindset to the extent that whomever affected will not be able to even *see* the limitations, save to overcome them.
  • "By saying they are "unproven", you're admitting they have yet to fail. "

    You got me here :) What I was trying to say is that so far OSS seems to be succesfull in other areas than the desktop. I have a strong feeling that the commercial software software industry wants to move on to making higher level software. By that I mean they are starting to realize that the market for lower level software like operating systems or simple mail servers is satisfied. Microsoft, apple or sun is not going to win any customers by rewriting the kernels of their operating system (though arguably that would be a good thing in some cases) because most users simply don't care about what powers their wordprocessors and webbrowsers. What is a profitable area is the stuff that runs ontop of an operating system though even there you see that some applications are so common that they no longer offer the fat profit margins they used to have.
    Software is subject to inflation, it gets less profitable over time. Nobody is willing to spend 100s of dollars on an OS that just runs your system, it needs to have something extra to be worth that amount of money. The same applies to a lot of other programs. Take development tools for instance. Nobody is paying money for just a compiler, in many cases you get the compiler for free. What brings in the money for development tool producing companies is the stuff that runs on top of the compiler (i.e. the IDE, debugger and prefab components).

    The trend I'm seeing is that the stuff that has dropped below a certain profit level is open sourced. In a way OSS has increased the inflation speed by 'attacking' products that are still profitable.

    This view I presented here sort of contradicts the accepted view on this site that OSS software is going to dominate everything. From where I'm standing, OSS gets to implement and reinvent the stuff that is no longer profitable enough. In many cases the reinvented version is better than the original so I see it as a good thing.
  • To compile a program, or part of one, you just edit and save the file. Saving the file compiles it. Wow.
    Yikes! You can turn that off, right? I habitually save every time I pause to think for a moment, which is frequently right in the middle of writing something such that if a compile were attempted at that point it would produce a large number of errors.

    I don't know if you can turn it off - you probably can, but I never felt the need to. For one thing, VAJ remembers your edits whether you save them or not - that how it should be, right? If you do save something with errors, it's no big deal, the program just stops until you fix them. It's really nice, trust me. You can work in the traditional way if you want, but this new way of working really beats heck out of the old way - it's fast.

    On the other hand, you'd want to exercise some care if you were actually working on a running system that was doing something important, but bear in mind how hard it would be to even contemplate doing this without something sophisticated like VAJ. I don't think most folks appreciate just what a high-tech feature this is.
  • Not really. I thought it was fast enough on an average PC.
  • Does the professional version have database and RMI support now, or do you still have to pay out the butt for that?

    Last time I tried VAJ, it was version 1.1 on OS/2. It was pretty cool, but had some annoyances. Like the fact that it just happily pasted error messages right into your code! Of all crappy ideas..... hopefully that's been changed since then.
  • Hi.

    Just to clarify, this is a fantastic IDE that supports Java2. Now, if they'd just release a fast jdk2, we'd be all set!

    Seriously though, VAjava makes programming in Java seem truly Object Oriented. It reminds me of Smalltalk Express, my first IDE, it had incremental compile (as does VAJ) and stored an "Image" leaving messy file management and bizarre symantics (FILE.java has to be FILE class) out of the eyes of the developer.

    Very cool stuff when you're just learning. I don't know why VAJ isn't used in every university, as it is free to them.

    Ben
  • I downloaded and installed the previous beta version for both SuSE and Win95. The two products were identical in look, feel and operation, except that the Linux version seemed a tad faster.
  • If you set a breakpoint in your debugger, you can edit code while debugging in VisualAge. Quite cool.
  • While most of your points are good ones, here's my response to some of them:

    - Enforced indentation & generated JavaDoc. I think this is somewhat of a nitpick - VA gives you lots of options for the different indentation conventions in Java. If you want to use a non-conventional indentation, you're being stubborn.

    - Method ordering: I can see the benefit of this, but what's important to YOU may not be important to someone else. A better tradeoff is just what VA chose: alphabetical ordering, since you can select which method you want to look at.

    - Poor cooperation wtih Version control: Sorry, I don't agree. The pro version integrates directly with PVCS & ClearCase, and I'm using it with CVS very easily. (though it's not integrated in this case)

    - Poor runtime support: Line numbers are irrelevant in VA. And yes, it actually is fairly useable with multithreaded systems. The enterprise version hooks in with servlets very easily (we're doing this now on one of my projects).


    Most experienced (10+ years) OO designers that I know prefer VisualAge for Java/C++/SmallTalk because it's the most expressive and flexible environment to program in... yes, it does tend to encourage ALL developers on a team to use VisualAge as well since it's not as file-based of an approach, but who the hell wants to use files anyway? Your job is to program, not to fudge with your source files.

    Furthermore, the packages/classes/methods metaphor is a proven Smalltalk ui concept that is the most efficient way for learning a system through its source code.

    In fact, most beginners do not use VisualAge because they find it too hard to use!

    anyway, this is my opinion.
  • You make it sound like emacs is an addiction. It is not.

    It's more of an obsession... :-)

    I'd classify emacs as more of a religion than anything else. I mean, 20+ meg compressed package for a text editor?


    ---

Slashdot Top Deals

How many QA engineers does it take to screw in a lightbulb? 3: 1 to screw it in and 2 to say "I told you so" when it doesn't work.

Close