Jbuilder vs Eclipse

[Bally Banwait]

Hi all,

I know the question of which IDE is asked each week,
but i'm in a little bit of a pickle.
I used JBuilder to create a program for my dissertation purely on the basis of that i knew how to use it. But i have been asked to describe the benefits of Jbuilder over eclipse and justify the use of Jbuilder. However my reason of "i knew how to use it" is not acceptable. I have no experience of Eclipse, and at this time do not have the time to learn it.
I was wandering if there are any Jbuilder users which could tell me some of the benefits that it has over Eclipse.

Thanks

[Ilja Preuss]

Originally posted by Bally Banwait:
But i have been asked to describe the benefits of Jbuilder over eclipse and justify the use of Jbuilder. However my reason of "i knew how to use it" is not acceptable.

Why not?

What if you found out that using Eclipse would have been better, if you have had known it, too?

[Bally Banwait]

Most research that i have found so far, all supoorts that Eclipse is better than Jbuilder. Unfortuantely are uni, are a bit mean so they want everything to be justified and i didn't quite think about that when i started programming.

Any ideas.

[Bharat Ruparel]

I find it hard to believe; you have been asked to justify using JBuilder instead of Eclipse in your "doctorial dissertation"? Don't they have anything better to do?
Bharat

[Ilja Preuss]

Originally posted by Bally Banwait:
Unfortuantely are uni, are a bit mean so they want everything to be justified

And why wouldn't "I was familiar with it, knew that it would suffice and wanted to concentrate on the actual subject of the dissertation instead of learning to use a new IDE" suffice?

What operating system did you use? Do you need to justify that, too? :roll:
[ April 11, 2005: Message edited by: Ilja Preuss ]

[Ernest Friedman-Hill]

Originally posted by Bharat Ruparel:
your "doctorial dissertation"?

He didn't say "doctoral dissertation", and I truly hope that it isn't; perhaps this is an undergraduate project or thesis.

But in any case, the only way you're going to be able to defend the position that JBuilder is better than Eclipse is to learn at least a little about Eclipse. Why not download it, try it out, and see what you think?

I suppose you could trot out the argument that if you pay for JBuilder, you get commercial support, although that's a weak argument in this case.

My condolences, in any case.

[Bally Banwait]

Yeah it is a undergraduate dissertation. To be honest, this was partly my fault at the start for not testing the two applications as i did mention earlier. Although this isn't a huge part of the project and certaintly not the main topic. I am just adding the finishing touches to the project and just needed a quick overview of why i used it. The justification i think that i want is that what benefits JBuilder would have to offer to the project. In question to the operating system. Since i am using Java it is therefore platform independent, but i suppose if you were doing something related to that, then you would require justification.

[John Boufford]

I think that your reason of � I knew how to use it� is relevant. In fact it was a key factor in choosing Eclipse as the IDE for a GM project I worked on in October of 2004.

I am a contractor and I have had to use whatever IDE the customer supported. IBM�s Visual Age, IBM�s WSAD, Oracle JDeveloper, Borland�s JBuilder, Compuware�s Optimal J, and Eclipse. Each IDE offers pro�s and con�s. For instance Optimal J was the best at Model Driven Architecture. However it only worked well with new development.

I would answer the question in terms of the Business; your instructors may be trying to make the same decision or checking on your decision making criteria. So here are some thought starters on supporting JBuilder.

Training, having your developer proficient with the IDE is important.
Company stability, focus on Borland long term experience with compilers such as Turbo C and Turbo Pascal. Include the fact that they will probably be around 10 years from now when the re-write needs to take place.
Infrastructure. Most likely you had a copy of the JBuilder installed.
Cost, Eclipse's j2ee add on requires a yearly licensing fee. I�m sure you could go into detail on the pricing model for both compilers.
Finally discuss the state of the compilers when the decision was made. JBuilder has been a recognized leader in compilers for several years. Eclipse has just reached a competitive state.

Here are some thought starters on supporting Eclipse.

Training, Most new developers are using Eclipse.
Stability, IBM and Borland both supporting and WSAD is build on the eclipse engine.
Ect.

If you choose to cover some similarities, Cover things like they both support UML and JUnit testing. Skip the obvious stuff like Applying color to syntax and stepping though code, or auto complete. It�s like saying a word process has spell check.

Well these are my feelings. I can create code with any IDE, even vi if needed. But somewhere in the organization someone is make these decisions... Most likely they are based on the cost to roll out to 100+ people and the cost to support and get support over the long term. How well it integrates with existing software. Finally what is the learning curve, and how likely will the tool be able to solve future problems.

My IDE of preference is WSAD.

[Jeanne Boyarsky]

Originally posted by Ilja Preuss:


And why wouldn't "I was familiar with it, knew that it would suffice and wanted to concentrate on the actual subject of the dissertation instead of learning to use a new IDE" suffice?

What operating system did you use? Do you need to justify that, too? :roll:

[ April 11, 2005: Message edited by: Ilja Preuss ]

I agree with Ilja. "I knew how to use it" is a perfectly valid reason. Unless your project involves writing a plugin, you could have done the project in vi or notepad (both perfectly valid IDEs.

On a related note, for my Master's degree thesis, my advisor questioned me on whether an interface is a class. After I (and Ilja and Ernest) convinced him that it was, everything was fine. I think they are more interested in you having a thought process and being able to justify it than the actual IDE of choice.

[Jeroen Wenting]

Personally, having used both extensively in a professional environment for years (though JBuilder for a longer period I admit) I must say I far prefer JBuilder to Eclipse.

It's in large part due to the more intuitive user interface (which is personal preference in part but not wholy so) and the better syntax highlighting.
As I rarely use wizards and stuff I won't comment on those, but insofar as I do use them I find the same things there.

Sadly it looks like Borland is going to abandon JBuilder soon. Their decision to join the Eclipse project as a core developer doesn't look good for the future of JBuilder, why keep developing a commercial IDE when you're spending money on developing its main competitor?

[Ko Ko Naing]

Originally posted by Jeroen Wenting:
Sadly it looks like Borland is going to abandon JBuilder soon. Their decision to join the Eclipse project as a core developer doesn't look good for the future of JBuilder, why keep developing a commercial IDE when you're spending money on developing its main competitor?

That's really terrifying news for me. I've developing things with JBuilder for more than four years. I hope that Borland will still be taking care of its JBuilder and continue it.

[Marcos Maia]

Hi,

I�ve used both IDEs on diferent projects. If I could choose myself(and not my employee) I would certainly go fo JBuilder Enterprise, the key points are:

All in one bundle, I can develop from a simple swing app, passing to J2EE an J2ME without instaling any pluging. It�s straighforward.
I has a powerful JUnit wizard integrated, and it helps a lot.
Great EJB designer, I don�t know any tool that has such a good environment for EJB development.(They keep talking about ejb complexity interfaces, dds, and all, but with JBuilder you�re abstracted from this using ejb2.0 from a long time ago).
Great integration with tool for code profiling/tunning(Optimezeit).
Great integration with all ALM - Application Lifecicle Management from borland suit, after all coding is only one part of a success project(requirements may be the most crucial one in fact as doesn�t help to build right the wrong thing - I�ve read this some place I can�t remember now and completely agree).
Very easy to integrate libraries like hibernate and get F1 touch help from them.
Very well integration with Weblogic/Jboss/Websphere app servers.
Great debug environment.
Great integration with together(for me the best UML tool available).
Ant integration with debug ant script capability.

These are some aspects from the top of my mind, notice that the borland tools are not for free ant that can be a great argument against using Jbuilder Enterprise.

About borland supporting Eclipse, in fact as far as I know Borland will develop the new JBuilder version on the top of the core eclipse platform(as IBM already does for it�s websphere studio), this does not mean they will abandon it, in fact I believe it will be very good as we will be capable when using the much powerful than Eclipse, Jbuilder Enterprise, to use ANY plugin done for eclipse already, so we�ll have all JBuilder Enterprise + all Eclipse pluging available to use.

[Paul Sturrock]

...can't...stop...myself...

All in one bundle

I see that as a disadvantage of JBuilder. Why would I want the baggage of supporting a technology in my IDE I'm never going to use?

Great integration with tool for code profiling/tunning(Optimezeit).

Absolutely. A real plus point for JBuilder. Thought I wish you could still get Optimizeit separately from the rest of the stuff JBuilder provides.

Very easy to integrate libraries like hibernate and get F1 touch help from them.

Its interesting you mention Hibernate, since Hibernator (a very useful tool) is not available for JBuilder.

Great integration with together(for me the best UML tool available).

Except its not properly round-trip. Try creating a sequence diagram, reworking the code in the IDE, then see if you can get your sequence diagrams to update themselves. Stand alone Together was a decent product but integrating it into JBuilder has hampered it a bit. I don't know if this is true of everyone's experience, but we have real performance issues moving diagrams. You can cut-and-paste a sequence diagram and wait ten minutes for you CPU to dip below 100% utilization.
My preference is Together stand alone, or Together with its Eclipse plugin.

My biggest issue with JBuilder (I'm referring to JBuilderX here) is performance. I'm bracing myself for a post saying just the opposite from Jeroen, but we experience such poor performance as to render the tool almost unusable. The biggest issue, as I have said, is modelling. But even just as a code editing tool, it just doesn't compare to Eclipse. We have been "in dialogue" with Borland for more than a year now to try to resolve these performance issues, but to no great result. Whatever they suggest, the IDE just is not as quick as Eclipse.

I can't help feeling when you compare a tool which costs $3500 with one which costs nothing, the $3500 tool better be far better than the free one. And JBuilderX just isn't.


(*sigh* And I could have just let that lie...)
[ April 15, 2005: Message edited by: Paul Sturrock ]

[Marcos Maia]

Originally posted by Paul Sturrock:

I see that as a disadvantage of JBuilder. Why would I want the baggage of supporting a technology in my IDE I'm never going to use?

You can disable any features you want(or you don�t need for a moment) in any project you want, in a very easy way in Jbuilder, so I really prefer that as having to configure plugins when needed so I don�t loose my time.

Except its not properly round-trip. Try creating a sequence diagram, reworking the code in the IDE, then see if you can get your sequence diagrams to update themselves. Stand alone Together was a decent product but integrating it into JBuilder has hampered it a bit. I don't know if this is true of everyone's experience, but we have real performance issues moving diagrams. You can cut-and-paste a sequence diagram and wait ten minutes for you CPU to dip below 100% utilization.
My preference is Together stand alone, or Together with its Eclipse plugin.

Looks like you�re not talking about the same together or Jbuilder, I�ve been using them for quite a while now and the code sync is perfect using together Control Center(now together architect 1.0) or Together for JBuilder. And in my machine I have no complaining about performance at all.( I use 1G RAM).
About together plugin for eclipse, as I�ve heard from a Borland friend is the buggiest together in all together family of products.

My biggest issue with JBuilder (I'm referring to JBuilderX here) is performance. I'm bracing myself for a post saying just the opposite from Jeroen, but we experience such poor performance as to render the tool almost unusable. The biggest issue, as I have said, is modelling. But even just as a code editing tool, it just doesn't compare to Eclipse. We have been "in dialogue" with Borland for more than a year now to try to resolve these performance issues, but to no great result. Whatever they suggest, the IDE just is not as quick as Eclipse.

I can't help feeling when you compare a tool which costs $3500 with one which costs nothing, the $3500 tool better be far better than the free one. And JBuilderX just isn't.


(*sigh* And I could have just let that lie...)

[ April 15, 2005: Message edited by: Paul Sturrock ]

Very strange that even borland support couldn�t solve your problems, first time ever, that I�ve heard such a thing.

Despite this all, JBuilder new versions will be build on top of eclipse core platform and as so you�ll be able to use everithing from eclipse plus the Jbuilder superior features.

Also as I said coding is just a small part of developing good software and Jbuilder is focused on integration with tools and features that WILL make easier to build good software. (Of course if you�re a super perfect developer you won�t need anything more then an IDE to build perfect softwares, but this definetly is not the real world for most pple).

You should also consider that in a big software project management feedback is essential and Jbuilder integrated with borland tools will provide this kind of information for a manager(infomation about how many lines of code for each programmer, how many line changes, time to implement a use case, etc).

[ April 15, 2005: Message edited by: Marcos Maia ]

[Steven Bell]

I feel a strange, irresitable, force, sucking me into another fruitless IDE war...

I have nothing against JBuilder, I've used it, I'm sure if I spent alot of time getting used to it I'd like it. What I can't understand is why I'd shell out $3,500 for it. Sorry, that is just a show stopper right there.

[Marcos Maia]

Hi,

this is fun... , on friday just before getting some beers.. hehe..
well I do respect eclipse a lot and as I said I have worked with it in a project and I don�t have nothing against it. But it doesn�t compare with JBuilder. About being expensive, that�s usually not my concern as I�m not the guy who buy the tool.

And also I�m glad borland will build next JBuilder generation on top of eclipse so I will be able to use eclipse plugins while still using JBuilder. Plugins like spindle for tapestry that are NOT available to jbuilder yet.

chers..

[Jeroen Wenting]

My main problems with Eclipse are poor performance and the user interface concepts.
If Borland abandons JBuilder and turns it into a set of Eclipse plugins they'll abandon a lot of what makes JBuilder strong.

If that happens I'll probably stop being a Borland customer and start looking for another IDE(A) when and if JB2005 no longer meets my requirements (which might be several years).

JBuilder might cost more than Eclipse but if the gain in productivity means that cost is recovered relatively quickly it's worth it in a commercial environment.
If I can work say 2% faster/more accurate in JBuilder (and that's a very conservative number, I think the real number is probably closer to 10%) that saves about an hour a week. At $100 an hour (for argument's sake) that I cost the company, that means that the cost of JB2005 Developer (I've no need for the Enterprise version for my work) is recovered in a few weeks time. At an economic life for the product of several years (no need to buy every upgrade at Borland in order to keep qualifying or keep getting support, unlike some other vendors) the gains are many times the purchase price.