[Jon Nermut]

It looks as though IBM are planning on dumping a whole heap of WSAD web tool functionality into open source via the eclipse web tools project.
See:
http://www.cs.huji.ac.il/~dar/wdt/ibm.html

I’m wondering if MyEclipse is planning on integrating with this stuff or competing with it? The IBM stuff looks really good, but looks like there is limited support for application servers initially (only Tomcat and websphere). It will probably be several months away at minimum as well. But it definitely looks as though you will have some free competition…

[Ivar Vasara]

@jnermut wrote:

It looks as though IBM are planning on dumping a whole heap of WSAD web tool functionality into open source via the eclipse web tools project.

Genuitec (the company behind myeclipse) is a member of the Eclipse organization, and is very aware of the IBM donation. I’d already pointed out the unofficial announcement to them in these forums ( see https://www.genuitec.com/forums/topic/eclipse-news/ )

@jnermut wrote:

I’m wondering if MyEclipse is planning on integrating with this stuff or competing with it? The IBM stuff looks really good, but looks like there is limited support for application servers initially (only Tomcat and websphere). It will probably be several months away at minimum as well. But it definitely looks as though you will have some free competition…

I can’t see anyone not leveraging the IBM donations.. anything that is officially part of the eclipse project would be the standard upon which others can build (the beauty of eclipse’s plugin architecture..) and by providing a competing implementation they’d be not only be competing with the official standard, they’d also be competing with any third party development efforts. ( bad idea )
Myeclipseide has suffered a major blow in that much of their development effort is about to become redundant.. but.. they’re obviously an agile company and given their track record and experience, I expect them to continue providing high quality , low cost value add.
I just wish they’d come up with an official statement. 🙄

[Todd Williams]

Since I’ve answered similar questions a few times, I’ve decided it’s now a FAQ (at least for me) so I’m posting a general response here so that it will be part of the public record and easily available to anyone that is interested in the topic.

To set the context of my reply, my name is Todd Williams and I’m the VP of Technology at Genuitec, LLC and our representative on the Eclipse Board of Stewards. Therefore, the comments here can be considered to be the official position on the relationship between MyEclipse, Genuitec, and the Web Tools Platform Project (WTP).

First, Genuitec will be participating in the WTP, primarily in the areas of the J2EE project model, Universal Server Tooling (UST), and server deployment. UST is actually a capability we’ve been lobbying for inclusion into Eclipse for about a year now. We are excited that we can now help address these areas through the WTP. I’ve already had several in-depth discussions with the Eclipse Platform PMC lead, John Wiegand, to discuss the project’s needs in both these areas. We’re expecting a synergistic relationship between the WTP, Genuitec, and MyEclipse. Genuitec will contribute to the WTP in an effort to ensure that it becomes a high quality source of flexible and extensible web tooling infrastructure.

This thread references the proposed contributions by IBM, so I’d like to add a few comments on that. One of the big problems that I have with the proposed IBM contribution as it is, is that it is all based on a very rigid project model, which will be completely incompatible with many existing project structures. In fact, when we designed the MyEclipse project model, we looked at WSAD’s and discarded it for precisely this reason. Since then, although our project model is already more flexible than the one embodied in the proposed IBM contribution, we’ve been working hard to make our model even more flexible and extensible so that it will be usable by the largest audience possible. For example, you’ll see full support for links in 2.7 GA and an even more flexible structure in 2.8.

It’s this type of project model flexibility and extensibility that must be part of whatever model is adopted into Eclipse. Remember, Eclipse’s main mission is to provide an interoperable infrastructure that all tools can leverage. The WSAD project model can’t be accepted “as is” since it does not fulfill this key design objective and would be a step backward for both MyEclipse and others. That said, I should also point out that the MyEclipse project model isn’t good enough as it is either. Collectively, we can do better than any of us can do individually, but we need to begin a design process with that in mind. That’s why we’re all at the Eclipse Foundation to begin with.

Given that the underlying project model must be redesigned, any tools built on top of it will need significant adjustment also. So, while IBM’s proposed contribution is “a solution”, if you’re building an end product, it is not “the solution” if you’re building infrastructure. However, please don’t take that comment to imply that the IBM contribution is not valuable, because it certainly is. We just need to design the correct solution for Eclipse first, and then look for contributed code that helps us to realize it faster. Starting with code and working backward will only ensure that Eclipse’s wonderful frameworks devolve into a hodge-podge patchwork quilt of whatever contributions could be expediently thrown together and called “good enough”. This can’t be allowed to happen or Eclipse will soon cease to be a compelling platform for innovation. Instead, Eclipse must continue to be the best example of a carefully woven tapestry of interoperable frameworks possible. We can only do this by designing first, and that’s one reason the new Eclipse Foundation contains a Requirements Council and Architecture Board to facilitate this process.

I’ll now change topics to how the WTP will impact MyEclipse. The MyEclipse team’s mission is to integrate, extend, enhance, and productize useful J2EE capabilities as soon as they become available, in any applicable open source projects. Open source is one of the raw materials we distill in order to manufacture our products. At some point in the future, the WTP will be one more high-quality, open source resource for the MyEclipse team to assimilate and extend. As a result of our ongoing involvement within the Eclipse Foundation, the WTP will benefit from Genuitec’s experience in the J2EE tooling space while MyEclipse’s customers will continue to receive the very latest in reliable, supported, and inexpensive J2EE development capabilities built on Eclipse technology.

If you’d like to follow my involvement with the Eclipse Foundation more closely, you can find additional information in my blog at http://www.jroller.com/page/todd_e_williams/Weblog.

Regards,
Todd

[Steve Bromley]

I agree with Todd on the downsides using the existing stuff from IBM.

I work for a company that is using WSAD. We have found that it is virtually impossible to get an existing project to work using the J2EE functionality provided in WSAD. We have been able to get MyEclipse to work in our existing project structure without much difficulity. If a rigid, predefined project structure is required to get the plugins to work, then for the most part, MyEclipse will become unusuable for anything but new projects.

I think that IBM should have followed the architecture used by MyEclipse. In my opinion, it would have been a lot more usuable. MyEclipse has a much better idea. 😀

[Scott Anderson]

Steve,

I personally think it would really help if more J2EE developers from large shops start posting their views about how important a flexible project structure is on the web tools newsgroup. Tell them why WSAD didn’t work. We need to let everyone know that the WSAD model is a non-starter and that as a community we expect something much better from Eclipse.

Todd’s going to fight the good fight and he’s a pretty persuasive guy, but the more voices that harp on this topic, the better. Besides, I think he’s going to be laying lower than normal since he’s running for one of the two Eclipse Foundation board positions that are now open for add-in providers. 🙂

[Jon Nermut]

Thanks Todd and Scott for the replies. I certainly agree with you that a flexible project structure is very important for J2EE development. The nature of J2EE is that there is no “right” way to architect your project, but a million different conflicting ways.
The J2EE project that I am leading at the moment was originally created in JDeveloper. We quickly realized the limitations of this rather lame IDE and wanted to switch to Eclipse. But the swtich only happened because I was able to get Eclipse + MyEclipse to build and deploy our project without any major refactoring of the project structure or build scripts. Once I had it working convincing management to change over was relatively easy – based on $. The switch most likely wouldn’t have happened if we had to start from scratch with a new project structure – it would have been just to much pain and down time.
Where I am interested in the IBM contributions are the little added extras that they have thrown in above and the basic html/jsp editors and server connectors, i.e. the XSD editor and Web Services support. Hopefully these plugins will not have too many dependencies and can be used by themselves – in which case we will take Genuitec’s approach and grab the best useful plugins from where ever we can.
As we are an OC4J shop we will definitely be sticking with MyEclipse for the foreseeable future, as no doubt it will be the best part of a year before anything comes out of the Web Tools Project that supports OC4J as well as MyEclipse.

[Author]

Posts