BEA to release Apache Beehive: Workshop open sourced

Discussions

News: BEA to release Apache Beehive: Workshop open sourced

  1. BEA is expected to announce that they were open sourcing major parts of their BEA WebLogic Workshop tool-set. The open source product is code-named "Beehive" and will be part of Apache. Rumor is that the parts being open sourced are the Controls, and Page Flows, and Metadata for Web Services. XMLBeans is already open of course.
    Alfred Chaung's view on this was expressed in a recent interview as follows: "I think the opportunity and also the challenge at hand is how do we continue to proliferate Workshop in an exponential way."
    Update: View the information from BEA itself at: dev2dev Beehive Project

    Read more in Will BEA's "Beehive" Save Alfred's Stock Price?

    In related news: BEA aims to turn it around

    Threaded Messages (41)

  2. Few people use this MS like tool.
    I tried it a while ago to see how it works, but I got into these behind the scene generators. This is like learning a new language - bea style.
    Hey, this is just my opinion - it reminds me about those MS tools that take over your mouse :)
  3. Actually, I encourage you to look at the BEA site. Those articles were released before our press release, and the JDJ article got a few things wrong.

    http://dev2dev.bea.com/technologies/beehive/index.jsp

    The main thing is we are not open sourcing the IDE or the designer. We are opensourcing the Meta-data (using JSR 175) driven programming model on the server side. The three main technologies that are in here are :
        Controls
        Page Flows
        Metadata for Web Services

        Will Pugh
        Principal Technologist
        BEA
  4. Are you releasing Weblogic app server and portal also as open source?
    Otherwise what is the big deal in open sourcing an application concept migrated to JAVA world through BEA from Microsoft. Workshop in JAVA world seems like the snake-head fish in USA!!. Hey still it is a development but not good for JAVA or Open source community!!. Do you have something else dude?
  5. Be grateful[ Go to top ]

    BEA does not "owe" the FOSS give-me-a-free-riders anything.

    The fact that they are opening up some of their stuff should be seen as exactly what it is: a gift from BEA to improve their customer's experience and service.

    If this gift adds value to the FOSS community, be thankful.
  6. Be grateful[ Go to top ]

    BEA does not "owe" the FOSS give-me-a-free-riders anything.The fact that they are opening up some of their stuff should be seen as exactly what it is: a gift from BEA to improve their customer's experience and service. If this gift adds value to the FOSS community, be thankful.
    Plus they figure this way they'll get some free ideas and coding time from developers out there.
  7. Be grateful[ Go to top ]

    Plus they figure this way they'll get some free ideas and coding time from developers out there.
    We are absolutely hoping to get new developers and new ideas brought into Beehive!!!

    Since, we are open sourcing this under the Apache License 2.0, these ideas go to the entire community (not just BEA). Free to anyone that wants to use, look at, or modify the code.

    This goal is to make J2EE easier. Not just BEA easier.
  8. Shame...[ Go to top ]

    People should be grateful for free stuff! Not everything is free!!

    If you can't afford something, then don't sh!t on it. Use the free stuff then! Everyone needs to make a living! You go BEA!!

    BEA has good Stuff!! If you like IBM, Oracle or whatever, then just use that!! Whatever makes your job easier!!
  9. Are you releasing Weblogic app server and portal also as open source?
    No. We are not open sourcing our J2EE server or our Portal product.
    Otherwise what is the big deal in open sourcing an application concept migrated to JAVA world through BEA from Microsoft. Workshop in JAVA world seems like the snake-head fish in USA!!. Hey still it is a development but not good for JAVA or Open source community!!. Do you have something else dude?
    I truly don't understand your point here. There is a lot of momentum in the Java community to simplify programming on J2EE. There also seems to be a lot of momentum to use Annotations and Meta-Data to do so. We did significant work in this area in Workshop 8.1 with controls, page flows and JWS. We are now open sourcing these innovations to help make the Java community stronger.

    If you think simple, elegant programming models are a Microsoft "application concept", I think you are mistaken.
  10. Check out xdoclet.sourceforge.net
    This is what you have missed.
  11. BEA is not open sourcing the application server or portal server. This was covered in the announcement for the open source project named beehive.

    - Beehive is about providing an industrywide programming framework for enterprise applications written on top of J2EE and Web containers.

    - Beehive is an open-source software project designed to deliver a cross-container, ease-of-use programming model and application framework for J2EE- and SOA-based applications. Beehive includes support for JSR 175 metadata annotations, the Java controls framework for creating and consuming J2EE components, a simplified Web services programming framework, and the Struts-based Java Page Flow technology for creating Web-based user interfaces and applications."
  12. Please learn to write properly[ Go to top ]

    If you can't write the English language properly, please don't waste our time by posting here. I expect people who post here to be able to write at a professional level (or at least at a level that indicates they have passed the 6th grade).
  13. The main thing is we are not open sourcing the IDE or the designer. We are opensourcing the Meta-data (using JSR 175) driven programming model on the server side. The three main technologies that are in here are :    Controls    Page Flows    Metadata for Web Services    Will Pugh    Principal Technologist    BEA
    Thanks BEA!
  14. I have used controls and Pageflows in Workshop for a while now. I felt hesitant using these proprietary extensions and didn't feel any love in mentally investing in something I couldn't use outside of Workshop. I also felt BEA was moving away from their standards-based philosophy. Even if the proprietary tools helped develop faster, there was that design smell of using something unmaintainable. Not anymore.

    For BEA, this is a useful selling point to write J2EE apps faster without incurring any proprietary overhead. They have also given back to the community.

    This effectively gets them out of the IDE wars for core J2EE development. They still retain their Portal and Integraton tools.

    (Please verify through TSS's new IP-tracking feature that I don't work for BEA. A feature inspired by the Partriot Act no doubt.)

    andrew
  15. I'd like to see a study comparing the productivity of a developer using Ant, XDoclet, Struts, Hibernate and Tomcat compared to a developer using Weblogic and Workshop.
  16. I'd like to see a study comparing the productivity of a developer using Ant, XDoclet, Struts, Hibernate and Tomcat compared to a developer using Weblogic and Workshop.
    Well I don't have any hard stats for you but as someone who doesn't work for BEA but does know and use and contribute to all the o.s. projects you mentioned (well I haven't contributed to hibernate so far) plus over a dozen other ones you don't mention, I can say - provided you have enough memory in your machine :-) - that Workshop is at least as productive as any other environment I have used. Of course we still make heavy use of a lot of open source including ant, parts of struts (non-netui, sorry BEA), checkstyle, pmd, jrefactory, canoo webtest, junit, junitee, sqlunit, comons-logging, http client and many others. These are all integrated with workshop as plug-ins or through ant. UML is the only thing which isn't there in our environment. We do that separately when we need to.

    I really like WSAD 5.1 too - it has a bunch of stuff built-in which you have to add yourself with Workshop - but it also has a significant cost.

    Anyway, back the the central thread, I think this move by BEA is great. There are many things done right in BEA's controls which we can now learn from and build upon. Plus the things we don't like we now can fix! As a community, J2EE needs to make progress in ease of use and BEA's approach is certainly one of the best attempts so far at achieving that.

    Paul.
  17. BEA should drop workshop .[ Go to top ]

    I love Weblogic 8.1 server but I simply hate weblogic workshop I havent met anyone who even use it . I think BEA is wasteing a lot of resources on product that no one uses. I would like BEA to provide plugins for Eclipse project . I dont think anyone will buy weblogic server if they try out workshop.
  18. BEA should drop workshop .[ Go to top ]

    Roger, as my earlier post said. The main benefit to Workshop is the tight integration with controls, page flows, processes, and a visual web service designer. However, the core IDE and Java editing doesn't compare to Eclipse. So, it would be nice to make the additional features plugins to Eclipse. I doubt BEA would support IBM though when they're at war in the marketplace.

    As an experiment, open up a relatively large file in Workshop, then try to do code completion on a method in it. It will be ridiculously slow. I reported this to BEA whose suggestion was to turn off code completion (uggh). It's a very inefficient IDE, the extra features which have previously been WL-specific (and may continue to be if Beehive isn't adopted by other servers exactly) are the only value it brings.
  19. Workshop is very important[ Go to top ]

    We're standardizing on BEA WL Platform 8.1 at my company and Workshop was a big part of that decision. It is a very good application-developer IDE. Of course, Eclipse is better as a full-featured IDE, and we use that too.

    Rolling the Workshop features into an Eclipse plug-in would be one of the best decisions BEA could make, IMHO -- but I'm still happy with the results we get out of Workshop, particularly for weblogic integration support, but also the EJB and Portal support. It eliminates most of the "stupid" complexity of EJB and JSP development.

    They're competing quite well with Visual Studio .NET in this way - vs. BizTalk or ASP.NET WebForms. Eclipse is more of a system coder's tool and WSAD is very pricey to get you up to the features Workshop has.

    Perhaps open sourcing the underlying Workshop components , as they're doing here, will enable smooth Eclipse support.
  20. I'm interested to see what is actually released open source. The core IDE of Workshop does not compare to Eclipse or other Java IDEs. Some of their extra features are nice (PageFlow, Process Definers), but those are useful only if you're using Weblogic Server with Weblogic Integration. So, I don't see why they would release those features open source, unless they want to make Workshop free for Weblogic Server customers, which would be surprising.

    They should really just ditch Workshop for Eclipse as a superior core IDE and make the additional features as plugins... fat chance.
  21. I believe this was mentioned briefly at TSSS as well. My main reaction to this is: Who's going to work on it? Open Source developers tend to work on products they use. The type of developer that works on Beehive is the same kind that uses Beehive.

    The best thing will be the integration with other app servers. Not needing BEA to run apps built in Beehive will probably be the biggest win.

    Cheers,
    Jordan
  22. I believe Chip Tyler and Arun Patel have already asked for commit rights on the project.
  23. I think you have it right. Straight from the FAQ:

    "
    Q: Will Beehive run on other application servers?
    A: Project Beehive will run on Apache Tomcat, the reference implementation for Java Servlet engines. Beehive can help multitudes of Tomcat customers scale their applications by easily connecting to industry-leading infrastructures such as BEA WebLogic Platform. BEA WebLogic Server is the only J2EE-compliant server on which Beehive runs. However, we anticipate additional ports of Beehive will be created through the open-source development process.
    "

    Of the pieces opened up, I don't believe NetUI is of real use (we don't need *another* very-similar-to-struts web framework). Java Controls are so-so, but they only get really useful in the VB-style programming like Workshop, so I'd be surprised if they are useful outside of Workshop. Web Services annotation I can't say much about either way, except that I thought J2EE 1.4 already makes pretty good headway in standardizing how app servers implement web services.

    Unfortunately, I don't really see this taking off as it doesn't seem very useful. But still, glad to see BEA open source some proprietary software, it doesn't hurt.
  24. Who's going to work on this?[ Go to top ]

    BEA was trying to bring a vendor lock on JAVA and J2EE by introducing a VB model, so that 1000s of dummy managers can try to feel the hot and warm of JAVA programming concept by drag and drop approach of workshop based development process. However BEA completely failed in the concept of buying corporate software purchase deals by teaching managers about how to develop J2EE by drag and drop approach. Now they are dumping it into open source community. I am not sure what is their intention.
  25. Who's going to work on this?[ Go to top ]

    Why do I feel the same?!?
  26. Who's going to work on this?[ Go to top ]

    Of the pieces opened up, I don't believe NetUI is of real use (we don't need *another* very-similar-to-struts web framework).
    NetUI is built on top of struts. The idea is that it uses meta-data to make programming on top of struts more simple and more productive. We are looking to enhance struts not compete with it.
    Java Controls are so-so, but they only get really useful in the VB-style programming like Workshop, so I'd be surprised if they are useful outside of Workshop.
    Java Controls are a peice that we've had a lot of interest in. The main use to experienced delveopers that it allows them to write much less repetitive wire-up code, and allows them to package up functionality to other users more easily.

    Controls are also useful to ISVs who are looking to be able to better package solutions aimed at all kinds of developers.

    Controls are also useful to IDEs other than just Workshop. Controls make tooling easier. I would hope that this is something that could translate into plugins built for both Eclipse and Intellij.

    To get a better feel for these, I would suggest looking at the documentation on the Database control and the Web Services control that we shipped in Workshop 8.1. Both of these will be open sourced in Beehive, and customers have found both to be very useful.

    http://e-docs.bea.com/workshop/docs81/doc/en/workshop/guide/controls/service/navServiceControl.html

    http://e-docs.bea.com/workshop/docs81/doc/en/workshop/guide/controls/database/navDatabaseControl.html
    Web Services annotation I can't say much about either way, except that I thought J2EE 1.4 already makes pretty good headway in standardizing how app servers implement web services.
    The Web Services project will implement JSR 181 (which is making good headway).
  27. The type of developer that works on Beehive is the same kind that uses Beehive.
    I think it's a great opportunity for Java developers in the community to work directly with BEA developers, just like BEA developers work on existing open-source projects (Groovy, StAX, XMLBeans, etc.) Also, IMHO controls have proven they are a viable and established component framework, and I've seen many frameworks come and go...some never making it out of spec-land (CORBA Components). With an open-source project for these, community developers will have the opportunity to 'scratch the itch' and make their mark on the framework.
    The best thing will be the integration with other app servers. Not needing BEA to run apps built in Beehive will probably be the biggest win.Cheers,Jordan
    I'd say a big win is for the control provider, knowing that they can provide controls and services on a platform that can benefit all Java developers.

    Garrett Conaty
    BEA Systems, Inc.
  28. I'd say a big win is for the control provider, knowing that they can provide controls and
    >services on a platform that can benefit all Java developers.

    And what was the reason do not provide them (controls) before that? Why do you think the open code will help? There is no obvious correlation I think. Actually I am asking as a control provider :-)

    Dmitry Namiot
    Coldtags suite
  29. Who's going to work on this?[ Go to top ]

    And what was the reason do not provide them (controls) before that?
    The Beehive effort has taken a while in the making, although this is getting announced right now, the energy and work behind making this happen has taken much longer. It is worth pointing out, that we are not simply open sourcing the code we shipped in 8.1. We are open sourcing the code we will ship in our next release.
    Why do you think the open code will help? There is no obvious correlation I think. Actually I am asking as a control provider
    We think open sourcing this code will help for two reasons:
      1) Part of this announcement is that Beehive will run under Tomcat (So it already runs on one more server :)
      2) Since it is open source, we think developers will port Beehive to run under the platforms they need.
  30. Who's going to work on this?[ Go to top ]

    I can see potential benefits for opening the code to beehive, I've looked at NetUI before and it would be good to take a look under the covers of the tool. But why are you using Apache to host the project? Why would anyone want to take a product and drop it into the worst developer site on the internet. Why don't you do your project justice and host it at Codehaus for example.

    I don't have any affiliation with CH, but at least I know that when I download a library from their site it will be well documented and (usually) of production quality. Compared to the dire state of a large amount of Jakarta projects, CH tends to be useable and relevant.
  31. Interesting. We chose to approach Apache first because we've have a very positive experience working with them on XMLBeans, and we believe that they have the largest community of developers. We like the CH guys too, (see http://stax.codehaus.org/), but our perception was that Apache has broader reach among developers. Would be interested in TSS's clans' take on this.

    George Snelling, BEA
  32. Tomcat[ Go to top ]

    Tomcat is not a j2ee server.
  33. 2 cents:
    Assuming someone is gathering requirements, No matter if it is the apaches or BEA product management:
    In order to provide a minimal level productivity, Worhsop must be able to work smoothely with less than 1GB of RAM.

    Yuval.
  34. While I am not opposed to BEA open sourcing part of their prodict, I genuinely wonder if there is a larger desire for these metadata facilities without the proper GUI tools available.

    After all, there main appeal is that I draw the webflows instead of editing XML files, that I can use editors to configure the metadata facilities for webservices. The whole idea of controls is to snap them into some VBish IDE to increase productivity. Just the metadata facility takes you about as far as xdoclet. Even if editors spring up for say eclipse or idea, then they will only provide part of the programming environment that I currently have using WebLogic Workshop.

    Also, the metadata facility does not create running web services does it? There is some infrastructure required (JAXP implementation for one thing). Will the generator create an implementation agnostic webservice or will I just be able to create client controls? Will the netuix controls be Open Sourced? In my opinion they are the actual added value of the WebLogic environment, and with JSF still going nowhere, I would like to see a "industrial strength" web gui component framework in the open source.
  35. Workshop is free[ Go to top ]

    I would assume that you would be able to use the Workshop IDE (which is free) to develop BEEHIVE applications -even if you endup running them on TomCat.

    Perhaps someone from BEA product management can confirm that this is technically Feasible.

    Also, Workshop Professional Edition is about $1000 and you can deploy limited-scale Server, WLI and Portal Applications into Production with that.

    Matt
  36. IDEs and MetaData[ Go to top ]

    I would assume that you would be able to use the Workshop IDE (which is free) to develop BEEHIVE applications -even if you endup running them on TomCat.
    Exactly. You can download Workshop free from BEA and use it to develop Beehive applications against WebLogic Server or Tomcat (or potentially others -- there is an extensible API). Moreover, with the introduction of standard (JSR 175) meta data, we expect a number of tools to be able to easily support Beehive application development.
    Perhaps someone from BEA product management can confirm that this is technically Feasible.Also, Workshop Professional Edition is about $1000 and you can deploy limited-scale Server, WLI and Portal Applications into Production with that.Matt
    Right again.

    Carl Sjogreen
    WLW Product Manager
  37. There's no doubt that metadata-driven applications require IDE support. I think that metadata-driven applications will become an integral part of J2EE development in the future and that open-sourcing the Workshop framework will accelerate the process of developing standards for metadata-based components as well as provide incentives for independent developers and companies to create plug-ins for the different IDE's to generate the required metadata graphically.

    However, whether you think that this is a good or bad move by BEA probably depends most on whether or not you think that metadata-driven applications are a good thing.
  38. Why not more of a good thing???[ Go to top ]

    First of all, I think BEA has made a very sound and very visionary decision in open sourcing the WLW Framework.

    There is a *LOT* of really inventive thinking behind WLW IDE and WLW Runtime - and best of all - it's a complete product (no offence - but XDoclet has been lying still for a long time, and XDoclet still use the un-intuitive template-language and still lacks a lot of support for servers like WL and others - not to mention crappy documentation). XMLBeans, Java Page Flows, Java Web Services and Java Controls really rocks in my opinion. Did you know that XMLBeans has build-in support for Schema, XPaht 2.0 and XQuery 1.0?? JDOM, DOM4J, SAX, DOM, whatever - go home!!! XMLBeans makes working with a XML doc as easy as working with plain Java-models. And controls makes a lot of the really boring plumbing code go away - like setting up everything needed to send a JMS message! Java Page Flows are really intuitive as well - think of it as a very easy to use version of Struts - especially using sub page flows is really nice. Finally Java Web Services are very easy to create with WLW Runtime - haven't used JAX-RPC, so I can't make a comparison here.

    I think me and by fellow developers can benefit a great deal from the WLW framework, as it makes development a lot quicker. And with the release of Beehive you don't even have to work for a huge company (with lots of $$$) nor accept vendor lock-in to use the framework. In my opinion a good mix of a VB-like-ID with the power of true J2EE spinning underneath is just great! What I really like is, that you are not constrained to use the VB-like features only - no I'd use WLW for the 80% standard stuff in an app. that WLW handles really well, and go down an abstraction level to the good ol' J2EE-components for the last 20% where I need more flexibility. (And with EJB and JMS Controls traditional J2EE code integrates seamlesly with the WLW Runtime coding model).

    I guess BEA will benefit a lot from OS'ing Beehive: More users will learn the WLW Runtime (ah.. sorry - Beehive) - which is good for BEA, as WL Portal, WL Integration and WL Liquid Data is all build on top of the WLW Runtime/Beehive. You could say, that BEA is betting it's business on the WLW Runtime. Thus a really bad scenario for BEA would be, if the J2EE community would rally around something like the Spring Framework, which would leave WLW Runtime as something you'd only use when creating WL Portals / Business Processes (like it's already happening with Hibernate compared to EJB). Furthermore the feeling I get from a lot of developers is, that they don't want to use WLW because of the vendor lockin - which will change with the release of Beehive. (And I personally really look forward to studying the runtime that makes my WLW-apps fly).

    As I see it, Beehive will be a win-win for BEA and the Java/J2EE community. Not sure that MS is going to like it though.

    The only thing I don't understand is, why BEA doesn't open source the Workshop IDE as well. It's the only viable alternative to Eclipse, and would have the comparative advantage of using the "official" Java-GUI kit Swing. As I see it, BEA would be able to win on an open source version of the WLW IDE - and BEA would have nothing to loose (I guess?). My only theory explaining why BEA doesn't OS the WL IDE is, that BEA actually want some open source freaks to make Eclipse support Beehive (and thus bring the Eclipse community to Beehive), which might be more likely to happen with a closed source WLW IDE???

    Hey all you BEA-folks reading this discussion - could you enlighten us on the reason why you don't see open sourcing the WLW IDE as a good idea??? Please :-)

    And would you (BEA) plan to support a project targeting Beehive-support for Eclipse??? (Which is still way better than WLW IDE when it comes to writing Java code).
  39. We are using Weblogic Workshop / Integration in our current project, and I very, very like the ideas which you've introduced there. But the Workshop IDE itself is buggy and consumes too much resources. I can't comfortably work with it on my P4-2.4/1G RAM workstation, because Weblogic Server with Integration + Workshop IDE consumes 1.1G of memory just to launch. And even when the server is stopped, Workshop is very unresponsive, unlike other IDEs on market.

    I like your stuff, but there's still a job to do on it.

    But if you decide to open source Workshop IDE, I think it can be boost to J2EE development. I don't think it will make a negative impact on your business: your commercial customers still need your technical support and infrastructure services. And all of the rest can already try entire Weblogic Platform for free and do everything they wont as long as they don't use it in commercial projects.

    So, what do you say?
  40. Portlet Toolkit[ Go to top ]

    I see promoting the Beehive toolkit and web application infrastructure as antiproductive to the process of arriving at a common, easy to use, web application framework. From what I can make of the page flow technology, it is not an extension to Struts but is trying to consume it. I do not want to see a Controls vs. JSF war instigated either. There are Java and web services standards developing roots that will provide infrastructure for more advanced web applications, JSR 168 and WSRP. BEA is obviously worried about being comoditized, but they should provide leadership that is not disruptive. My suggestion for BEA would be for them to support these portlet standards in their portal server and give us open source tools to create standardized portlets.
  41. I see promoting the Beehive toolkit and web application infrastructure as antiproductive to the process of arriving at a common, easy to use, web application framework. From what I can make of the page flow technology, it is not an extension to Struts but is trying to consume it.
    Please elaborate. What do you mean by "trying to consume it"? I find Page Flow has taken the best from Struts, and then added/changed some things to make Struts a lot easier to use. If that is what you mean by "consuming", I think BEA should take you comment as a compliment. The only thing that has kept me from using JPF is the vendor lockin that came with it - a concern that is gone when BEA releases Beehive. I urge you to take another look at JPF.
    I do not want to see a Controls vs. JSF war instigated either.
    Do you know what controls are?? Controls and JSF will SUPPLEMENT each other nicely - but they are two VERY different things!! (Actually BEA says, that they will include JSF in Beehive)... Take a close LOOK at controls, and THEN tell us what you think.
    There are Java and web services standards developing roots that will provide infrastructure for more advanced web applications, JSR 168 and WSRP. BEA is obviously worried about being comoditized, but they should provide leadership that is not disruptive. My suggestion for BEA would be for them to support these portlet standards in their portal server and give us open source tools to create standardized portlets.
    BEA Portal already support JSR 168...

    Do you actually know anything about the BEA Platform, or are you just making some wild shots?!?!
  42. I used to think the problem with Weblogic Workshop was it's proprietary nature, and so making it open source should help.

    But now I think the problem that seems to prevent tools like this (Weblogic Workshop, Oracle JDeveloper, the SOFIA framework) from making a bigger impact then they do is that Java developers don't really want GUI toolkits.

    It's strange, in other environments tools like this do very well (ASP dot net, Cold Fusion, Visual Basic, Delphi, Powerbuilder, etc...), even though they are 100% proprietary, but Java developers seem to want to hand code everything and write as much low level plumbing code as possible, even though it makes application development take twice as long.

    For example, the SOFIA framework (http://www.salmonllc.com/sofia) gives programmers a DreamWeaver plug-in to visually edit JSPs, but most Java guys still hack through and write the JSPs with a text editor, even though it takes much longer to produce the same result.