Kito D. Mann's Blog

Has Borland seent the light?

I've been a devout fan of Borland's development tools since my pre-teen years (I started programming with Turbo Pascal, downloading code from BBSs). I used Delphi back in the day, and I was a debout JBuilder supporter until about a year or two ago, when I moved to Eclipse.

Even when I was using JBuilder, a friend of mine and I often concluded that Borland should be selling the best suite of Eclipse plugins on the market. While I was glad to see JBuilder 2007 reborn on top of Eclipse, I was intrigued, but not sold. (Granted, I did always love Together, which is integrated, along with OptimizeIt!). Admittedly, part of my resistance was the lack of any additional JSF tooling, which is my primary concern.

With JBuilder 2007, the pricing has come down quite a bit -- the standard version only costs $499, which is squarely aimed at IntelliJ. But Borland has just moved a step closer to the light by introducing JGear, which is basically JBuilder 2007 sold as three separate plugins (plus a server portion for its collaboration tools) that work with existing Eclipse installs. So, if you're using MyEclipse but you want profiling, you can just add JGear Performance. Or, if you're using BEA Workshop, you can add JGear Team client plugin. And the pricing is actually reasonable; the plugins start at $299.

Am I excited? More so than usual. I think Borland really has seen the light. Is it too late? Maybe so. But late is always better than never...

Comments

• I was also very intrigued by this, but when I started looking closely, it seemed that the interesting ones already had open source counterparts. I did not use them myself yet, but the Performance plugin seems very similar to the TPTP project so I am not sure if the pricing is really all that reasonable. Just my two cents.

  eyt*

  Posted by: etheriau on August 17, 2007 at 08:38 PM

• Do they have a plugin to migrate their generated GUI code for older JBuilder versions? See they abandoned every user of their visual designer with the move to an Eclipse based system. At work we literally have hundreds of dialogs and frames created with JBuilder that now have to be maintained by hand. It's possible to hand write good swing code thats easy to maintain, but JB2006 does not create anything like that.

  Isn't CodeGear just an attempt to spin the IDE development part of the company into a separate entity so that they can sell it off? They droped their actual product and refactored into a group of plugins.

  Gee.. thanks.

  I would say the have "seen the light" when they support their customer base instead of making the CodeGear division look like a nice acquisition target.
  

  Posted by: aberrant on August 20, 2007 at 07:24 AM

• I told them squarely they would loose the indy and SME market if they didn't modularise their products and start thinking seriously about their pricing policy almost a decade ago. They didn't listen then, let's see what will happen now.
  I'd not be surprised if "Codegear" gets a bad case of Inprise in the coming years...

  Posted by: jwenting on August 20, 2007 at 11:51 AM

• and oh, I'm not going to pay $500 for the privilige of using Eclipse and a stripped down few plugins when there's IntelliJ for $250.

  Posted by: jwenting on August 20, 2007 at 11:52 AM

• and that's from someone who's been a loyal Borland customer for over a decade...

  Posted by: jwenting on August 20, 2007 at 11:52 AM

• What would really be sweat is if they would offer in particular the Together product as a Netbeans and as an Intellij plugin.

  Posted by: pupmonster on August 20, 2007 at 03:26 PM

• How could borland see the light? I invested a lot of time and effort in delphi. A project they failed to port over to Linux. When they moved over to java things got only worse. Instead of porting their really excellent GUI development tool to java, they cave in and follow the Swing route. Which basically throws away the idea of GUI layout by dragging and dropping unto a canvas, to code everything. Yea, all GUI developers in Suns world code GUI from VT100 terminals connected via dial-up modems.. For years Borland used to be an the fore front of compiler and UI kitts. They just gave all of it up to follow the crowd. So in essence borland has abandon the developer to embrace the corporate & industrial ethos of making profit at all cost. The sad fact is Borland tools where the ONLY tools I ever forked out money because they were of really good quality. Allowing me to quickly develop front ends without loosing focus on my customer's requirements. No we have eclipse, good IDE but a even worse visual control library. The most in eclipse now requires many hours of coding, and reams of code using their RCP. Netbeans on the other hand, we have SWING which also forces the same requirements on the developer. Gone are the days when I can create a Delphi form, with minimal code have a workable application. So thank you Borland.

  Posted by: wildfirexy on August 20, 2007 at 11:52 PM

• The Java GUI paradigms don't translate to the Delphi mindset of dragging and dropping fixed size components on a fixed size form, making fixed size screens for an application that behaves well as long as it is run on a system that uses the same fixed size font as the one on which it was designed.
  In fact that's one of the major drawbacks of Delphi (and indeed all "visual" editors), the other of course being the client/server architecture which promotes strong coupling of data and business logic with user interface code.
  JBuilder did in fact have an option to use a more or less Delphi like D&D approach to Swing development, using a custom layout manager Borland created for the purpose which was superior in many ways to NullLayout in that it did behave somewhat decently under resizing. But that was never embraced by users because it was Borland specific code, not something provided by some other 3rd party (why people would willingly buy a license to use a Layoutmanager but not use one they get free with JBuilder is beyond me).
  Borland died because they forgot to cater to their core users, not because they didn't ship quality products (if only those core users had been able to afford them at "enterprise" prices). Kylix was abandoned because there was no market for it. The Linux crowd didn't want it because it wasn't open source, the corporate users didn't want it because they have no need for Linux GUI applications (especially ones that work only in one or a few window managers, which is endemic of pretty much all good looking native Linux applications).

  Posted by: jwenting on August 21, 2007 at 01:09 AM

• I also like the approach Borland (or CodeGear) is taking. Whether they are polishing their portfolio for an easy take-over or not, will not affect my point of view. They are VERY late to the party, though.

  I've had the opportunity to try the Eclipse-JBuilder and with the JGear announcement, there is even better offerings for different Java development scenarios.

  I can't see where there is "an open source counterpart" to the Together tooling. If this piece is sold for a mere $399, it's a bargain. I tried this myself with their EJB3 modelling (and migration from EJB2!) and found it to be really nice.

  Will CodeGear make it or not? Only time and marketing managers will tell ;-)

  Anyway, Good luck!

  Posted by: gullet on August 21, 2007 at 05:53 AM

• Hi Folks,

  I just saw Kito's post about JGear, which we're pretty excited about. I see a few questions in the feedback and I wanted to chime in and clarify a few things:

  1. One person asked about TPTP vs. JGear Performance. I'd have to recommend you try them out head-to-head on a real performance project. JGear Performance is based on Optimizeit, using the same data collection engine and advanced viewers, updated to use JVMTI. It does things like remote profiling really well, and can even profile an application across VMs.

  2. "Isn't CodeGear just an attempt to spin the IDE development part of the company into a separate entity so that they can sell it off? They droped their actual product and refactored into a group of plugins. "

  Not exactly. For many years those of us who build JBuilder, InterBase, Delphi and other developer-focused products have had to take a back seat to ALM. With the creation of CodeGear we operate as a seperate entity entirely focused on *developer* tools. This is a Good Thing.

  Also nothing has been dropped, JBuilder is still available and that is the best way to get all of the tools together at a better price. The JGear plugins represent the most popular parts of JBuilder and provide a way for folks that use _any_ Eclipse-based IDE to get the same productivity aids.

  These are good, solid tools for professional software developers. There will be free trials available for download when we release the JGear plugins -- give them a workout. If you think they aren't worth it let me know directly and we'll address any concerns with our first update.

  Cheers,
  Joe McGlynn
  JBuilder Team
  

  Posted by: jmcglynn on August 21, 2007 at 01:55 PM

• Joe,

  I'm glad your re-org is working out well for you. I'm sure the plugins are great. I'm frustrated by the lack of support for the old GUI system in the 2007 Eclipse based product. We are stuck with 2006 until we can find some migration path. Eclipse won't do it, NetBeans won't do it, IntelliJ claims to be able to import any GUI but I haven had a chance to try it out. The bottom line is, you've lost a customer, and it had to be a consciouses decision to leave us all behind.

  Posted by: aberrant on August 22, 2007 at 07:44 AM

This work is licensed under a Creative Commons License.