BEA and Codehaus have partnered to create Controlhaus. What is hosted there?

BEA WebLogic Workshop Control Pack, a set of freely available packaged and open-source controls designed to help enable developers to easily tap into the Web services capabilities of the Amazon, eBay, Federal Express, Google, PayPal and UPS platforms.

The BEA WebLogic Workshop controls are immediately and freely available for download at http://dev2dev.bea.com/technologies/controlpack/index.jsp. BEA is sponsoring an open source community to foster collaboration and further improvement on service controls. The community will be hosted at controlhaus, a sister site to popular open source community codehaus, at http://www.codehaus.org.

Read the press release: BEA Offers Free Packaged and Open-Source Components to Help Build Applications for Service Oriented Architectures

Visit Controlhaus

I'm confused now, why are BEA using Apache for the Beehive project and also setting up with Codehaus? I know the two primary developers on Aspectwerkz, a Codehaud project, work for BEA on JRocket, and I'm glad to see them do more work with Codehaus (who do produce some very useful products), so where does Apache come into all of this? Why are they donating any code to Jakarta?

Perhaps I'm missing an obvious, political point, but I'd rather see more people join in with Codehaus and Open Symphony who produce fewer, higher quality projects than Jakarta's raft of deprecated/discontinued projects.

Back on-topic, it is great to see the bigger vendors realise the benefits of opening more of their code. As a developer I find it encouraging to see the bigger companies raising the bar on the quality of OSS projects, especially in relation to the documentation and packaging, elements that can often determine whether a project is useable or not.

Jakarta is part of Apache, a US non-profit organisation, not controlled or dominated by any individual or company, with over 100 committers (http://jakarta.apache.org/site/whoweare.html). Some committers are from Sun, IBM and BEA but they do not dominate. New projects must demonstrate that they are not geographically or company dominated.

Open Symphony is a much smaller grouping, led by two individuals (http://www.opensymphony.com/team.jsp) with a small team of people.

Codehaus is also a smaller grouping, led by one indiviual (http://www.codehaus.org/Manifesto).

Controlhaus is an unknown - although linked to Codehaus, it describes itself as a sister-site. Since all the code originates from BEA, there may be questions of company dominance.


Each person must decide for themselves as to what they feel is the best way to run a project, or an OSS group. Personally, I'm not that fussed by the group, and more interested by the quality of the individual project - so I'll use Ant, Tomcat, Commons, Xerces from Apache, Jaxen from Codehaus and OSCache from OpenSymphony.

Stephen, committer Jakarta Commons

Presumably this has something to do with the fact that, according to Cameron Purdy, BEA hired Bob McWhirter to help them out with their Open Source efforts.

I could see them choosing to go an other-than-Apache route for smaller, lower profile, projects simply to avoid the hoops that they have to jump through for the ASF otherwise. Granted, this is only my opinion and I could be completely off base as to their reasoning. I'm not saying that the Apache way is bad at all, in fact it works quite well to make sure that the legal issues are well understood and the legalities have been satisfied, but I can see it being too much of a hassle for some projects to consider.

Personally, I'm not the least bit troubled by the group that it was released through. Companies giving software to the community is a good thing no matter what the release vector happens to be, as long as licensing terms are reasonable.

Beehive has a specific purpose and a limited audience, it is not applicable to all projects. Jakarta projects are often aimed at lower levels, ie APIs (commons) and frameworks (struts), as Beehive sits above this layer wouldn't it have been a better candidate for Codehaus? In fact wouldn't the webservice API's/components have been a better candidate for hosting at Jakarta?

The benefit of the smaller organisations is that they are more focussed on smaller goals. Jakarta seems to have a finger in almost every Java-flavoured pie, I doubt there are many Java developers who haven't used Jakarta code. I also doubt that many of these developers have been happy with the stablility of the Jakarta APIs, I know I have had to refactor many projects to get around the deprecation that they are so fond of.

.Controlhaus is an unknown - although linked to Codehaus, it describes itself as a sister-site. Since all the code originates from BEA, there may be questions of company dominance.

BEA sponsored the creation of the controls as well as a community site. We felt it important to partner with an independent entity to encourage ISVs and developers to have their own projects and specifically their own control. This is something encouraged by the codehaus model, where project leaders are autonomous (essentially a web of trust) for the projects which they lead and commit code.

Likewise, in controlhaus we want to encourage both corporations (eBay/Apache/Amazon/UPS/FedEx) as well as independent developers to manage their own section of controlhaus (e.g. http://ebay.controlhaus.org

Garrett Conaty
BEA Systems, Inc.

One of the important reasons we wanted to go to Apache with Beehive is the synergy of communities that are there.

Beehive builds on top of Struts, makes extensive use of XMLBeans, uses Velocity for code generation, and the Web Services will run on top of Axis. We also expect there will be synergies with projects such as Geronimo in the future.

With all the Apache technology we use, we felt it made sense to let Beehive sit aside it's ancestors. Also, we were hoping that people in some of these communities would be interested in helping :)

I'm confused now, why are BEA using Apache for the Beehive project and also setting up with Codehaus?

My bet is that they want to have FSF or FSF-like licenses for some of their projects for the same exact reasons that JBoss.org prefers FSF or FSF-like licenses. Althought the BSD/ASL zealots like to think so, BSD/ASL is not a one-size-fits-all.

I don't believe that this decision (the split between Apache and Codehaus) was licensing motivated.

The decision to put the base Beehive framework into Apache was largely based upon the belief that there was some good synergy with other projects that live there (Pageflows are built on top of Struts, Controls use Velocity for code generation, and the Web Services implementation will leverage Axis).

There may be some Controls put into the Beehive repository in the Incubator, but generally only as examples. It (Apache) will be more the place for the framework to live than a general respository for contributed control types built using the framework. I think the feeling was Controlhaus might serve better for this latter function.

Of course, we are just trying to get things started in the right direction. If we are wrong, then I'm sure it can be made right.

Cheers!

-- Kyle Marvin

Graham,

Both Apache and Codehaus represent the Java community well, and there's lots of cross-pollination between them. I think that audiences are attracted to each for different reasons (e.g. they have a different 'feel'), but in the end multiple communities are good for Java developers.

One of the key factors in our decision to establish controlhaus with codehaus is the leadership and vision of 'Bob the Despot'. Controlhaus is the first site that provides cohesive project functionality (version control/issue tracking/forums/wiki) with a good interface.

Garrett Conaty
BEA Systems, Inc.

open-source controls

I couldn't tell right off the bat - Are these Web controlss(if so can they easily be made into portlets)? Desktop Components? API's that you create your own UI?

http://www.dev2dev.com/controlpack.com

BTW, Link doesn't work.

BTW, Link doesn't work.

Quick search on BEA's site turned up the following URL:
http://dev2dev.bea.com/technologies/controlpack/index.jsp

BTW, Link doesn't work.

Quick search on BEA's site turned up the following URL:http://dev2dev.bea.com/technologies/controlpack/index.jsp

Yeah, saw that. Forgot to post. Thanks. My mind was more on how do I use this stuff.

Thanks, I updated the URL in the main message.

I couldn't tell right off the bat - Are these Web controlss(if so can they easily be made into portlets)? Desktop Components? API's that you create your own UI?

Most of the controls in the service control pack, and on controlhaus are web service controls. These controls provide a simple POJO interface to the vendors' web service APIs, and putting into metadata often used parameters (e.g. authentication tokens).

You can absolutely turn these into portlets...for a great example, check out the salesforce portal demo at http://sforce.controlhaus.org, which shows how easy it is to create a full featured portal using these controls.

The great thing about them is that you can compose business processes, page flows, and even new controls, using these freely available, open-source controls.

Garrett Conaty
BEA Systems, Inc.

Thanks!

Its great to see a vendor finally take the route that MS$ is taking with Visual Studio. MS$ has a great opensource community contributing newer and newer controls everyday for ASP.Net. They are leaving the Java camp in the dust with reference to webapp development. This is primarily because the Java camp keeps reinventing the wheel with all the different Webapp frameworks that are not very compatible with each other. Choice is good, but fragmentation is not.
Instead of furthering one or two mature & capable frameworks, every camp has their own lowest common denominator set of tags and controls.
Sometimes, a benevolent dictator role played by a vendor to foster healthy advancement in one area of technology is a good thing.

Lets applaud BEA's move for doing this.

Don't get me wrong, the Java community is always leading the way in innovation around libraries, techniques, patterns and almost everything else. But when it comes to rallying around a set of developer tools that make tough things easy for newbie developer types, we still have a long way to go.

What's next? A community fostering a similar set of controls for JSF and Sun's Java Studio Rave. That would be great too.

What's next? A community fostering a similar set of controls for JSF and Sun's Java Studio Rave. That would be great too.

In the Beehive Controls architecture when you author a Control you are simply authoring a specialized type of JavaBean that brings along some special powers, like the ability to set properties via JSR-175 annotations, as well as a resource management and extensibility model that makes them particularly useful for building a client abstraction to external resources (whether web services, EJBs, databases, ...).

There is no reason Controls can't be useful anwhere JavaBeans are useful. For example, accessing a Control from a JSP page is as easy as using the <jsp:useBean> tag.

We are hoping that Project Beehive enables many more interesting use cases for Controls. A key design goal is to make them as useful as possible in a wide variety of environments. Hopefully what is happening in Apache and Controlhaus can get the ball rolling, but we'll need help!

-- Kyle Marvin
   Developer, Beehive Controls Project

There is no reason Controls can't be useful anwhere JavaBeans are useful.

Porlets?

The shipping Weblogic Portal exposes many personalization services via Controls... so I think it is safe to say that portlet use cases should be entirely possible.

Cheers!

-- Kyle

The shipping Weblogic Portal exposes many personalization services via Controls... so I think it is safe to say that portlet use cases should be entirely possible.Cheers!-- Kyle

Found a link explaining it - http://dev2dev.bea.com/technologies/beehive/overview.jsp

Found a link explaining it - http://dev2dev.bea.com/technologies/beehive/overview.jsp

You can find out even more at The Beehive Project page, including a Controls overview and programming model links in the Controls Sub-Project section.

Its great to see a vendor finally take the route that MS$ is taking with Visual Studio. MS$ has a great opensource community contributing newer and newer controls everyday for ASP.Net. They are leaving the Java camp in the dust with reference to webapp development. This is primarily because the Java camp keeps reinventing the wheel with all the different Webapp frameworks that are not very compatible with each other.

Yes, but WinForms is still a page paradigm. And that is what the plethora of Java web frameworks are attempting to solve.

So, like where are all these OpensSouce ASP.Net controls? I can only find $$$ ones. Pretty much everything MS related is. Not that that is bad.

WinForm and WebForm are 2 different things.

There are ton of free asp.net control on the internet. Not
only control, but GUI control where you can drag in and play.

GCL

WinForm and WebForm are 2 different things.

Oops. Mentally fat fingered it. Meant WebForms. I hope I am know the diff. I've done WebForms and am about to do WinForms.

There are ton of free asp.net control on the internet. Notonly control, but GUI control where you can drag in and play.GCL

OpenSource and Free aren't necessarily the same.

I can find a few, but none very useful. Check out Echo and Echopoint for Java OSS controls. So again - where?

Yes, but WinForms is still a page paradigm. And that is what the plethora of Java web frameworks are attempting to solve.So, like where are all these OpensSouce ASP.Net controls? I can only find $$$ ones. Pretty much everything MS related is. Not that that is bad.

Have you seen Infragistics? Their controls are available with source and their grid controls have NO match in the Java world. Microsoft has tons of experience in building a component oriented framework. ActiveX and COM controls on the windows platform are hugely popular. They have a commercial market around it and that is key.

Java frameworks may be trying to "solve" the page paradigm, but I ask, what's lacking? With the control architecture, .NET presents a pretty neat paradigm. All java frameworks do is invent new ways of writing config driven applications. Its a nightmare to develop and manage.

Shameless plug of course as I work there, but we have a lot of components too:
http://www.servletsuite.com/jsp.htm

and it is Java and it is not a framework ;-)

>.NET presents a pretty neat paradigm
agreed about ASP.NET

Dmitry Namiot
Coldbeans

Have you seen Infragistics? Their controls are available with source and their grid controls have NO match in the Java world.

Yes, there are plenty of components out there for ASP.Net and some with their source available for sale. But you said opensource. Infragistics isn't.
 

 Microsoft has tons of experience in building a component oriented framework. ActiveX and COM controls on the windows platform are hugely popular. They have a commercial market around it and that is key.

Is it key? Why is MS starting to OpenSource somethings? Because it is only part of the key.

I've done (and do) development in VB, ASP, ASP.Net (C#/VB.Net), WebForms[starting] and Java (Swing/Web). In the Java world, I have never lacked for controls. I'm sure though someone can find an example of a missing control. And really, how many Charting controls do you need? I only used but a few VB 3rd party controls. And I have great story about 3rd party ActiveX controls - DLL hell - nuff said? :)

Java frameworks may be trying to "solve" the page paradigm, but I ask, what's lacking? With the control architecture, .NET presents a pretty neat paradigm. All java frameworks do is invent new ways of writing config driven applications. Its a nightmare to develop and manage.

What's lacking in ASP.Net? An MVC and OO paradigm. Granted many Java frameworks are a nightmare to develop and manage. But some are not. Have you tried Echo, MillStone, WebOnSwing, ... ? ASP.Net has its pains too. Try pulling it all into SharePoint. Then compare it to something like Liferay. Have you tried figuring out IIS6 security?

I'm not going to say that ASP.Net is not cool. MS did a good job (at the very least they did a good job of making people think they did a good job). But it is not perfect and there is more than meets the eye - "All that glitters is not gold".

The orignal poster of this part of the thread ("Emulating Microsoft here") said "MS$ has a great opensource community". I have yet to see it. So "Show me the no Money!". :)