BEA to introduce "Cajun" in 2002


News: BEA to introduce "Cajun" in 2002

  1. BEA to introduce "Cajun" in 2002 (23 messages)

    BEA has announced plans to introduce a new product in 2002 called Cajun, a new software tool that will enable less sophisticated developers, perhaps with 20 or 40 hours of training, to build relatively inexpensive, non-mission critical programs that can automate work specific to a department or individual.

    Sounds like a GUI website builder. Could this product compete with BEA's own partners such as Altoweb?

    Press Release
    BEA Systems introduces product for new market to analysts
    NEW YORK, Nov 15 (Reuters) - Shares of business software maker, BEA Systems Inc.(NasdaqNM:BEAS - news), on Thursday fell about 4 percent, despite a meeting that one analyst called ''long-term positive'' and demonstrated the company's ability to develop new products to maintain its market leadership position.

    Shares of BEA closed down 59 cents or about $15.95, after a 41 percent run-up on its stock price in the nearly two weeks preceding Tuesday's release of its earnings and its annual meeting with analysts.

    ''My sense is that it's trading rhythms,'' Legg Mason analyst Paul Krieg said, adding that some investors were taking profits from the increase.

    BEA is the leader in the market for application servers, software that forms a basis on which developers build their applications. Its flagship application server, WebLogic, competes with WebSphere, the close No. 2 application server made by International Business Machines Corp. (NYSE:IBM - news)

    Both WebLogic and WebSphere are based on J2EE, the architecture that uses Java as its code. Java was developed by Sun Microsystems Inc.(NasdaqNM:SUNW - news) and has been a favorite among developers and is an alternative to Microsoft's products.

    Until now WebLogic has been the domain of highly skilled high paid developers who build sophisticated programs that run businesses.

    However, during the meeting with analysts, Alfred Chuang, a BEA co-founder named chief executive last month, introduced a new product called ''Cajun,'' expected to be on the market in the first half of 2002.

    The new software tools will enable less sophisticated developers, perhaps with 20 or 40 hours of training, to build relatively inexpensive programs that are not critical to run a bossiness but ones that can automate work specific to a department or individual.

    Cajun can allow BEA to go into a large company and offer them products for a range of their needs.

    ''I think this is a very cohesive strategy at going after the framework level of what companies need,'' Krieg said. ''I think that BEA did a very good job outlining why they are in the leadership position in the market today, and outlining their strategy ...raising the competitive pressure on everyone else. It's a strategy that I think that makes a lot of sense and what will be most critical in the next six months, nine months and maintains the pressure on the competition.''

    Threaded Messages (23)

  2. BEA to introduce "Cajun" in 2002[ Go to top ]

    Interesting - although this offering doesn't look like it would go as deep as AltoWeb. Does anyone know of any other competitors to AltoWeb?
  3. BEA to introduce "Cajun" in 2002[ Go to top ]

    Fair warning on this comment: I work for Oracle.

    For another Altoweb-like solution, you might be interested in taking a look at Oracle's J2EE framework, Business Components for Java, which going down the feature list of Altoweb pretty much has a similar breadth of functionality.

    This includes modeling support, EJB 1.1 support, out-of-the-box tag libraries and JSP beans, performance monitoring, declarative business rules, XML datasets, support for any J2EE certified application server and in the imminent release candidate of Oracle9i JDeveloper, the ability to publish the EJB's produced by the framework as Web Services. While it works quite nicely as a stand-alone class library, it also has a gui in Oracle9i JDeveloper (and earlier releases).

    Something I believe Altoweb is trying to do as well is map their framework to J2EE Design Patterns as one of their product demos shows how the Sun Petstore could be built using their solution.

    If this interests you, check out how Oracle's J2EE framework maps to standard J2EE Design patterns in this paper released last week:

    Interestingly, the writer, Steve Muench, maps the framework capabilities directly to coded samples in the Sun Petstore tutorial to show the value of the framework.

    With Oracle's framework, sometime soon Cajun and Altoweb, there is clearly a plethora of frameworks emerging. I am sure others will chime in with others.

  4. Competitors to AltoWeb?[ Go to top ]

    It's hard to say what's competitive. AltoWeb has a lot in it, so it depends on what your requirements are - what benefits you are looking for.
    1) What your service consists of - how many integration points between external packages?
    2) Message format/Interop: How much emphasis do you give to Web services as the message protocol - as opposed to CORBA or MQ? How many services are external to your organization?
    3) How important is graphical service/application assembly,
    4) How important is performance?

    AltoWeb is a Business Process Manager (BPM) that runs on top of a J2EE server and also provides Web service management. At design time, a BPM enables you to describe business processing - some require no programming. At run time, BPMs are like sophisticated message routers, moving messages through a series of services. Logic servers are somewhat similar to BPMs.

    Here's a short list. Most are headed toward Web services (everyone is headed toward Web services though people have a number of different pictures of what Web services actually are).

    - Bowstreet
    - Versata BPM, Versata Logic Server
    - DCH Bluebridge Logic Server+BPM
    - Level 8 Geneva App Builder

    But also consider:
    - NewVis
    - Epicentric
    - Component X
    - WakeSoft + a UML tool
    - OptimalJ
    - SonicXQ
    - Fiorano
    - Integration offerings from IONA, Sun (used to be called Forte Fusion, now called iPlanet Integrator)
    There are others. Start here:

    Web service enabling tools:
       Read the whole page.
       It gets better as you get closer to the bottom.

    BPMs, Logic servers, Enterprise Tools:

    About BPMs: Business Process Management Institute:

    Article on Instant J2EE (right side):


    Rich Katz
  5. BEA to introduce "Cajun" in 2002[ Go to top ]

    I work for Versata (

    Versata provide a very sophisticated product set that allows the building of end-to-end J2EE applications, or just thre rapid creation of the server-side components and integration with best of breed content vendors. This can be achieved very rapidly due to the innovative design-time business rules that are used to enable the creation of server-side J2EE components. At deployment Versata also deploys framework services that encompass best-of-breed J2EE patterns such as Session-entity facade, value object pattrern, lightwieght classes etc.

    Many toolkits are available that extend the Versata proposition such as the XML Integration toolkit which allows multiple transaction to be passed and processes via XML over either HTTP (AKA Web Services / SOAP) or JMS.

    Although I am probably seen as biased as I work for the company I think it is a powerful proposition as it compliments existing Java skill-sets and allows companies to get to market quicker and maintain applications more rapidly (change /add a rule rather than changed / amend the code).
  6. Versata - avaoid at all costs[ Go to top ]

    Been there done that.

    I spent the beter part of two years fighting wiht versata. What a pain in the ***.

    Major issues was lack of documentation, lack of support, their tech support couldn't answer anything deeper than how do you change the color of a field.

    BTW we build an app the beat the pants of what anyone else has built with versata, and last I heard they were still running it in house.

    Bench marked some EJB code to replace versata under JBOSS and it beat Versata solution by about 8 times, and was stable. (JBoss 2.2.X vs BLS 4.0) Upgrading from 4.0 to 5.0
    caused us serious pain and was still in progress when I did the bench mark.

    Versata took about 2 years to ship a Java 2 solution, and even then it forced you to use 1.2, 1.3 had been out for some time.

  7. Versata - avaoid at all costs[ Go to top ]


    I've got say I'm surprised by your comments and would be interested to know what period they relate to and what version / product set.

    You refer to JBoss as if it would be a direct comparison to Versata which is a little confusing. Versata is not an Application Server. It works on best-of-breed Application Servers - IBM Webpshere 3.5.4, and BEA WebLogic 5.1. Currently beta versions are being put thorugh their paces that support Websphere 4.0 and WebLogic 6.1. These are expected to be released around Q4 2001. We also have a CORBA verscion of the product which supports Borland's Visibroker - I think this is what you are referring to in your comments.

    In terms of performance, we recently rebuilt the IBM Trade2 reference application in Versata and not only did it take 2 to 3 weeks to complete rather than six months, but the performance achieved was comparable to the initial reference implementation.

    This thread is probably not the right place to discuss the merits of Versata, and my initial entry was to show that there are already sophisticated tools that allow companies to leverage their investment in WebLogic.

    My last point would be that Versata has been about since 1996, so we were there at the beginning and a lot of what we have learnt on our journey through to today has gone into the product, so if you get a chance have another look at it, you might be surprised.
  8. Versata - avaoid at all costs[ Go to top ]

    Hi James,

    well as I said I was compairing the BLS 4.0 to jboss. The BLS was what the Versata Server was called in version 4. This was before Versata tried to run on top of other app servers. And yes it used visbroker, but as corba does not have a component model it was more versata than visi broker. And at the time, versata sold it as a appserver. Of course given how well it worked I can see why you would back away.

    Yes versata was there from the start (1996) but they never seemed to learn. My favorite example of versata code quality was the VSResultSet. Now a VSResult set was an interface and the api had methods that would accept a VSResultSet. When you have a method that takes an interface as an argument you should be able to pass in any implementation. Any Java 101 course will tell you that.

    But the versata code did a cast to a VSResultSetInternal, their impl of VSResultSet, so you can't create a custom impl of VSResultset. Even better VSResultSetInternal is a final class, so you can't extend it! Yes!

    So versata's stellar coding style had an method that took an interface where there could only be one implementation.

    That's just one horor story, I have many others! Versata, avoid at all costs!

    PS I'm a big fan of the for Versata internal use only approach to java docs. :)

  9. Versata - avaoid at all costs[ Go to top ]


    I don't want to get in a 'thread war' with you where I respond to your criticisms of Versata and you come back with something else. You obviously have some very fixed opiniions of the product from the point in time that you used it. Like any product it has moved on. I remember the first version of Jbuilder that I used that was awful, but the current release is probably one of the best Java IDE's on the market.

    If you want to converse direct I'd ve quite happy to. My email is James_Liddle at Versata dot com.
  10. Versata - avaoid at all costs[ Go to top ]


    "I don't want to get in a 'thread war' with you where I respond to your criticisms of Versata and you come back with something else"

    That's good cause the list is long :) I'm not sure what sort of response you could have for code like that, other than to hang your head in shame. We're not talking polish here, were talking to the core bad code. Of course fixing it would be good, but in two years I saw more bad versta code than good, so that's a tall order.


  11. Versata - avaoid at all costs[ Go to top ]

    Versata is all about *not* coding (or at least as little as possible). Using this (IMHO) minor Java point as your main criticism of the product would suggest that perhaps you did not use it properly (i.e. as a Java IDE instead of a declarative environment). There are thousands of developers around the world who are quite happy using Versata, please don't knock down a perfectly good product.
  12. Versata - avaoid at all costs[ Go to top ]

    I think this answer is a litte trite - anyone who has used Versata know that at some point you need to get involved in some Java coding.

    I think the real issues are:

    1. Adherance to Java coding conventions
    2. Adherance to J2EE - basically there are still a lot of CORBA naming etc in Versata. Also it seems to me that the logic is in Entity Beans, not in Session Beans where it should be. At least this was the case in the version 5 of the product for Websphere
    3. Connectors to other things other than Databases
    4. Get rid of that awful HTML presentation technology - become a framework

    Also the price is wayt excessive on a per CPU basis - it's not worth 5 times the App Server.

    However when all is said and done it is a very good product - not one I'd recommend to expereienced developers but certianly for companies who need to delvier enterprise Java prodjects quickly
  13. Versata - avaoid at all costs[ Go to top ]


    just a quick response to some of your points here:

    1. We are just at the point of releasing version 5.5 of the product which is a substantial release of the product which I am sure addresses any concerns you have as to Java conventions and J2EE compliance. Also we are working with IBM to allow Versata VLS to be 'plugged into the IBM Workbench Toolset.

    2. Actually deployed objects are normally 'lighweight classes' (as per new EJB 2.0 spec') with a Session Bean as the transactional controller aka the Session-Entity facade pattern. If the developer wishes to deploy Versata objects as EJB's then Versta Data objects become Entity Beans and Versata Query Objects become Session Beans.

    3. There are a lot of connectors that allow access to other systems. We have worked with MQ, Tuxedo and various Back Office Applications on projects where needed. It is just that the default connectors that are installed with Versata are for all the major RDB vendors.

    4. The default Versata Archetype technology has been extended with Presentation Design Extions (PDX) which are much more feature rich, and there is also a toolkit that allows clients to work with JSP if that is what they require. To this end we have worked on projects where Versata is implemented with a Struts framework etc.

    We are not 5 times an App Server price, and as always our pricing model is flexible depending on the project and requirement - ISV requirements etc.

    I also believe it is a an excellent product as do many companies who have implemented Versata within their strategic platform. It is much more than a server-side framework because you also get the innovative design time features that allow rapid build, deployment and change of Versata Projects due to the business rules techology ('What' not 'How').

    I fell new products, such as Cajun, and framework vendors, such as Wakesoft, show that Developers and companies have real requirements that abstracts the complexity of application design and delivery.
  14. I work for Instantis - I've been looking at the discussions surrounding Cajun, Versata, web services, the Petstore app etc. and I'm surprised that folks haven't looked at our platform called SiteWand. While Cajun is touted as the "non-mission critical app" server, SiteWand has all the same advantages even for mission critical apps!

    The successor to the Petstore app in JDK is Duke's bank. See if this is not a compelling comparison: original Duke's bank code was 5000 lines and you be the judge of how long it would take you to develop from scratch. We have recreated the same app using browser-configured "engines" on top of a SiteWand platform in a few days (2 days to be precise). Check out our applications gallery and go to our resources section if you would like to see our whitepapers. Key features of the platform:

     * Build and run scalable, robust Web applications
     * Fast, component based creation - 10x faster than custom coding
     * Run-Time automatically provides instant deployment,
       transactions, sessions, concurrency, etc.
     * Easy Web services creation
     * Requirements: J2EE environment, JDBC to relational database

    Instantis SiteWand was developed by key architects from IBM's Transarc, Websphere and HP e-speak.
  15. J2EE Application Frameworks[ Go to top ]

    This discussion is very interesting. I have just begun research into application frameworks.

    One of the more interesting frameworks I have run across is called realMethods. It was built from the ground up on top of the J2EE design patterns and includes framework components for each tier of your application. It also integrates with web-tier frameworks like Struts. I have never heard of some of the frameworks mentioned here so it has been very helpful in learning more about frameworks.

    Some other frameworks we are researching include Expresso, WebAccel, and Kona.

    <strong>Chuck Kiefriter</strong>
    Nielsen Media Research
  16. J2EE Application Frameworks[ Go to top ]

    Being an actual Cajun, I'm not sure whether I should be insulted by this new product, since it is a software tool for "less sophisticated developers".
  17. J2EE Application Frameworks[ Go to top ]


    Another intersting framework is from a group called Evergreen. The framework is more oriented towards commerce, but still seems pretty strong.

    You should also check out something from a company called Wakesoft. There framework is called an architecture server and it seems to be a pretty thorough framework. I attended a webinar and downloaded the product and am currently evaluating it. I also talked to them at the Java Developer's Conference and they were really nice and helpful people.

    Togethersoft also does some very rudinmentary framework like things, but you have to use there product to get advantages out of it.

    I'm glad Expresso is on your list, it's a cool product.


  18. J2EE Application Frameworks[ Go to top ]

    I`m currently using struts framework for my project, however, as you know, struts focuses on the controller and view. As for model, I have to handle on my own. I tried expresso framework, however it is so complicated and lacking documentation(In order to get more support licence fee? Reasonable enough..However I`m just a college student).
    I`d like to ask about your opinion in adopting frameworks, and if possible, those which can work with struts because I`ve worked with struts for four month and like it;-)
    By the way, has anyone used Castor JDO implementation?
  19. J2EE Application Frameworks[ Go to top ]

    Hi, Charles.-

    I see that you were looking into some frameworks and I also agree with you that realMethods looks very interesting. I was wondering if you guys decided in a framework and which one and maybe you can tell why if you did?

    TDC Kabel TV
  20. BEA to introduce "Cajun" in 2002[ Go to top ]

    Also, wonder if it's going to be something like cold fusion?
  21. BEA to introduce "Cajun" in 2002[ Go to top ]

    My idea is that this is something build on top of their existing WLI and WLP products. They have the workflow engine, they have the integration stuff, they have the webflow engine, etc. If they add something like a Lotus Notes document database or something and make a nice integrated GUI they could pretty easily leverage existing technologies into "Cajun".

    I think what BEA needs to do to get out of the infrastructure/appserver/middleware-market and start making some bigger bucks is to get more into the CRM/ERP/etc-market like Oracle did with Oracle Applications. But that's just my 5c.
  22. BEA to introduce "Cajun" in 2002[ Go to top ]

    I think the industry and analysts agree that something is needed on top of application servers in order to help people build applications. We have been looking at some of the alternatives for some time, and AltoWeb seems to have the best approach by far.

    It's narrow minded to think just of web applications, or just of web services. There are many other genres of applications that need to be built (messaging based, wireless, etc) that many of the proposed solutions (including Cajun) do not address.

    Apologies to James, but Versata just does not cut it for anything other than data entry applications. The lack of business logic is a real negative point. The Oracle solution also seemed to be too Oracle centric.

    I also think that Richard is wrong. AltoWeb seems to be more than just a BPM application. They support the entire application space, from data integration to presentation and allow presentation in the form of Web Services or XML based services, as well as the usual JSP (to HTML / WML). They also support much of the application lifecyle, such as testing and deployment, that other vendors ignore.

    I would strongly recommend downloading the 'complete environment' version of their product. It comes complete with a JBoss app server (one to watch) and hypersonic database and allows you to get up and running and building apps very quickly. We downloaded, installed and started working on their platform in just 7 minutes. Not sure anyone could beat that.

    As for BEA moving into applications, sorry Jon, but I don't think it would work. History has shown that as soon as you start to spread your wings, you defocus and devalue your core offering. It could be argued that Oracle has devalued their database by moving into apps, and they still have only a tiny market share (of CRM for example) compared with the big players. The same may also be true for BEA and Cajun - only time will tell.


  23. BEA to introduce "Cajun" in 2002[ Go to top ]

    Being an actual Cajun, I'm not sure whether I should be insulted by this new product, since it is a software tool for "less sophisticated developers".
  24. BEA to introduce "Cajun" in 2002[ Go to top ]

    Being an actual Cajun, I'm not sure whether I should be insulted by this new product, since it is a software tool for "less sophisticated developers".

    Think about how people living in Cairo, Whistler or San Francisco must feel...