Microsoft Targets J2EE Users with JLCA 2.0

Discussions

News: Microsoft Targets J2EE Users with JLCA 2.0

  1. Microsoft Targets J2EE Users with JLCA 2.0 (76 messages)

    For what its worth, Microsoft just unveiled v2 of their utility (JLCA: Java Language Conversion Assistant) that automatically converts Java code to .Net. This version attempts to migrate JSP code as well, converting elements like tag libraries to Asp.Net web controls.

    Read more about it on Internet News:
    Microsoft Targets J2EE Users with JLCA 2.0

    or go to the Microsoft JLCA web site

    I see some value here for .Net afficionados who want to leverage functionality currently available in the form of open source Java projects by using this conversion wizard. In fact, I think some "ported projects" on Sourceforge do this today using JLCA.

    The article cites CodeNotes.com, which ported their Java codebase to .NET using JLCA. It is worrying to see someone come out and "claim" that "maintaining the system was a nightmare".

    Apparently, support for converting EJB code is expected in the next version.

    Threaded Messages (76)

  2. Do they support Ant script, XDoclet, Struts etc.? Because if they don't, then I cannot really imagine this stuff to make any sense except noise. And BTW who is going to make migration? VB developers who has know clue about the J2EE or experienced J2EE developers who religiously against all MS technologies for any reason or no reason.
  3. I think .NET doe's a great job for us. It is a good place to test experiments,
    before to add new/better/untested features to JAVA. And I think .NET has a lot of good experimets/innovations.
  4. And I think .NET has a lot of good experimets/innovations.


    Such as ? ....

    .Net hardly has anything that really distinguishes it from Java other than a good Development environment. and when J2SE1.5 hits the shelves there will be even less.

    Microsoft hailed Java as dead and as yesterday's technology if that is the case then so is .Net.
  5. .Net hardly has anything that really distinguishes it from Java other than a >good Development environment. and when J2SE1.5 hits the shelves there will be >even less.


    That's because we are including those things into 1.5 :). autoboxing, enumerations, foreach loops as far as language syntax goes. We are adding metadata to annotate classes. We added ADO.NET's ideas with different implementations of RowSet. There are a lot of ides from Microsoft that you can find in Java, and vice versa.
  6. .Net hardly has anything that really distinguishes it from Java other than a >good Development environment. and when J2SE1.5 hits the shelves there will be >even less.

    >
    > That's because we are including those things into 1.5 :). autoboxing, enumerations, foreach loops as far as language syntax goes. We are adding metadata to annotate classes. We added ADO.NET's ideas with different implementations of RowSet. There are a lot of ides from Microsoft that you can find in Java, and vice versa.

    Yes, innovations is a positive side of this framework it is inspired by JAVA, but the most new 1.5 language features are inspired by .NET.
    I do not think this tool was a good idea, it must be a lot of better ways for integration.
  7. And the EJB 3.0 specification seems to be taking ideas from the way .NET distributed objects are implemented (the big thing here being the metadata). I think that is good as it will greately simplify the way these are written. All in all .NET is causing a flury of improvements to Java and J2EE. Hopefully, this is a cyclical thing and developers on both sides will benifit as the environments take good ideas from one another.
  8. E.g. go to Coldtags suite and convert ASP.NET controls into JSP tags :-)
  9. Well, the fact that so much of what's in 1.5 mimics the interesting things that .NET brought to the table should tell you that .NET has something to offer. For example, "foreach" constructs, autoboxing, constants, and (I think) attribute marking.

    The fact of the matter is that Java and .NET both have a lot to offer and both can be more or less appropriate in any given situation. I like the fact that there's competition in this space. For my own personal reasons, I tend to prefer Java solutions and the vast number of useful libraries out there (I feel that the open source community will never be as vibrant for .NET as it is for Java due to the nature of environments where .NET is more appropriate). But I'm perfectly happy working with either and can appreciate the strengths each has. Maybe I'm patting myself on the back too much, but I think that's a pretty healthy attitude to take.
  10. I think this is one of the business strategies. But one thing is sure that J2EE design patterns and concepts are truly matched for OOAD expert professionals. The major question is that if I am an expert on C,C++ and JAVA, then why should I go for Dot Net and also if JAVA language can able to fulfill business requirements. I think this tool will be useful for those people, who do not know how to write programs on C,C++ and JAVA. So if some projects are already developed in JAVA and owner of the project wanted to convert that project into Dot Net and also the professionals, who are going to work on that project do not know how to write program on JAVA, for them this tool will be useful.
  11. Why .NOT?
    http://www.StupidVideos.com/?VideoID=53

    This make SUN CEO look relatively smarter:
    http://www.StupidVideos.com/?VideoID=255

    .V
  12. Nicer link to video[ Go to top ]

    http://216.127.90.17/~admin1/Low_WMP8/Microsoft.wmv
  13. And other video[ Go to top ]

    http://216.127.90.17/~admin1/Low_WMP8/Steve_Balmer.wmv
  14. There seem to be two primary reasons CodeNotes.com made the switch:
    1. Environment Configuration
    2. Licensing Fees

    Environment configuration is a pain, but can be managed, especially when the environments are well understood, as they are with a hosted application like a Web site. Licensing fees shouldn't require switching languages, just vendors.

    I'll bet that their development environment (that is, what it was like to work there as a developer) was the real source of their problems.
  15. Question on article[ Go to top ]

    Quote from article:

    CodeNotes said converting the Struts framework, which relies on EJB, was an issue.

    "Struts nested tags is an area where the JLCA won't convert exactly the way we expect it to," McGovern said. "Fortunately for us, most of the time that we have nested tags related to Struts, we're using an iteration tag combined with some sort of formatting tag. The good news is that these are very easy to replace with ASP.NET controllers. We can rip out almost all of the nested tags and replace them with a simple grid control."

    Even so, McGovern said the important thing when coming up against a trouble area like this is getting the functionality, not the code.

    "It's not as important that our ASP.NET site have exactly the Struts framework and the Struts class libraries and everything that you're used to," he said. "What it needs to do is mimic the same functionality that Struts gave us."



    Question:

    Since when does Struts rely on EJB's? Perhaps the Struts Actions they used called EJB's per business requirements but Struts and Struts tags do not use EJBs. He goes on and says that Struts was difficult to migrate because of EJBs and talks about the work around for tags. Was this just meant to drag EJB's in the mud? EJBs have nothing to do with Struts.
  16. Question on article[ Go to top ]

    Since when does Struts rely on EJB's? Perhaps the Struts Actions they used called >EJB's per business requirements but Struts and Struts tags do not use EJBs. He >goes on and says that Struts was difficult to migrate because of EJBs and talks >about the work around for tags. Was this just meant to drag EJB's in the mud? >EJBs have nothing to do with Struts.


    Yet again we have an example of how Microsoft FUD in action, They had to drag in EJB since it's a typical J2EE buzzword, by pulling EJBs they affectively smear Struts/EJB for potential customers who don't know better...
  17. EJB's in CodeNotes[ Go to top ]

    Quote from article:

    >
    > Question:
    >
    > Since when does Struts rely on EJB's? Perhaps the Struts Actions they used called EJB's per business requirements but Struts and Struts tags do not use EJBs. He goes on and says that Struts was difficult to migrate because of EJBs and talks about the work around for tags. Was this just meant to drag EJB's in the mud? EJBs have nothing to do with Struts.

    Being Rob McGovern, I can say that the article author misquoted me. I never said that Struts relied on EJB's or that we used EJB's in CodeNotes. I said two things:

    1. JLCA v2.0 does not convert EJB's. So if your application uses EJB's, you will have to rewrite these pieces by hand (until possibly JLCA 3.0)
    2. At the time we did the original conversion, the JLCA did not handle nested tags very well. However, that was a bug we submitted and Microsoft has since fixed.

    I'm happy to answer CodeNotes questions. However, I have to say that I'm not thrilled with the article on Internet News. The author took bits and pieces of other material and did jumble together parts that didn't mix.
  18. Reduced costs...[ Go to top ]

    On this page http://www.microsoft.com/resources/casestudies/CaseStudy.asp?CaseStudyID=14258 it is claimed that moving from Struts to .NET has eliminated several 3rd party software licenses. Obviously Struts and JSP/Servlets do not require a license to use (and they are 2nd party I *guess*).

      Can someone shed some light as to what "several 3rd-party software licenses" and "The Java version uses several different third party products, each with its own upgrade and patch schedule" are and how they are eliminated by moving to .NET?
  19. Reduced costs...[ Go to top ]

    From the site developer...

    Moving from Struts didn't eliminate any 3rd party license. Moving CodeNotes from Java to ASP.NET resulted in the elimination of third party licenses for Jive Fourms, JRun, the JChat module, and the Lucerne Search engine and I may be forgetting one or two more.

    I will cheerfully try to help clarify any such technical questions.
  20. Reduced costs...[ Go to top ]

    I take that back, moving from Struts did eliminate the Apache license associated with the Struts codebase. However as this is an open license, no money was saved.
  21. Reduced costs...[ Go to top ]

    It seems to me that the separate licence fees are incorporated within one licence now, so at the end of the day at all boils down to the same thing. I remember in one of my previous jobs we built a pretty decent web application that uses Struts, JBoss and mvnForum and it had cost us almost nothing...

      I just wonder whether the current costs you saved using .NET will still be something to consider in a couple of years time.
  22. Reduced costs...[ Go to top ]

    It seems to me that the separate licence fees are incorporated within one licence now, so at the end of the day at all boils down to the same thing. I remember in one of my previous jobs we built a pretty decent web application that uses Struts, JBoss and mvnForum and it had cost us almost nothing...

    >
    >   I just wonder whether the current costs you saved using .NET will still be something to consider in a couple of years time.

    Yes and no. The fees we pay for CodeNotes:

    1. Windows 2K/2K3 - Same as before (The JSP version runs on Win2K now, the ASP.NET on Win2K3)
    2. SQLServer - Same as before
    3. License for VS.NET - As a company, we already pay for this anyway, so it's a wash as far as CodeNotes goes.
    4. License for Jive - Gone
    5. License for JChat - Gone
    6. License for JRun - Gone

    Because the license for Windows covers IIS, we can eliminate the JRun fee. We also drop the Jive and JChat fees. As for other non-cost licenses, we don't save any money and possibly pick up cost for internal development and maintenance. However, we replaced Jive and Lucerne with the open or "community source" Starter Kit. We replaced Struts with a simple http handler and basically kept all our actions intact at no significat cost. In the long run, we trade three separate (small) payments for one payment we already made anyway (to Microsoft). Does any of this apply to a pure Java/Linux shop? Probably not. But in a mixed environment such as ours, it makes sense.

    CodeNotes is a pretty low cost website to begin with. Any cost savings are going to be pretty small one way or another. Although CodeNotes makes an interesting technical case study, it's not really a good website to use (for either side) to compare cost or performance or hard numbers. We're talking about a website that has constant but fairly low traffic and is basically a content portal. Also, because the JLCA does a fairly conservative conversion and tries to maintain your existing code structure, our perliminary counts for number of objects and number of lines of code are pretty similar between the ASP.NET and Java versions. You'll only start to see changes when you get to the phase 2 conversion and start optimizing for ASP.NET (e.g. ditching Javascript for web controls). I'm working on getting these numbers together and they will eventually go on the migration website (no promise as to when).

    Purely IMO, the best reason we did the conversion was just to see if we could do it. We're a training and consulting company so any chance we have to play with a new product or idea is worth taking. Although "management" needs cost justification, most of us "developers" just wanted a real project to use as a learning experience. The CodeNotes website worked well in Java and it works well in ASP.NET. We started with a non-trivial codebase that incorporated several common J2EE patterns. We wanted to see whether the patterns could apply in ASP.NET and how much work it would take to do a conversion. We definitely found the answers. Would we have done a conversion without the tool? Probably not. No reason to reinvent the wheel (or website).

    Finally, I'm most definitely not recommending that everyone rush out and convert all of their Java to .NET. What I will say is that, if you have a website that meets the right criteria, the JLCA provides a rather useful means of conversion. Conversion isn't for everyone, but the fact that you can get most of the way with a tool makes it a viable option in some circumstances.
  23. Reduced costs...[ Go to top ]

    4. License for Jive - Gone

    >5. License for JChat - Gone
    >6. License for JRun - Gone

      I still can't see why you would need a license for those at first time. There are several completely free and open source java forums, there are several completely free J2EE Servers (JBosss, Jonas, Sun ONE to name but a few). As for JChat, I never used so I can't say anything but I am sure you could find something similar that's free.

    > As for other non-cost licenses, we don't save any money and possibly pick up
    >cost for internal development and maintenance.

      this is not a valid argument I believe since you still need internal maintenance and development for a .NET application.

    > In the long run, we trade three separate (small) payments for one payment we
    > already made anyway (to Microsoft).

      You hit the nail on the head here. That's exactly what happened. You already paid the costs for using Microsoft products so any incurring costs seemed like overkill to the overal budget. However I *believe* if you didn't already pay for .NET license the costs would have been far lower using Strust and JSP.

      I am just wondering. Did you also have to pay JLCA license fees or was it also included in the .NET license?
  24. Reduced costs...[ Go to top ]


    >   I still can't see why you would need a license for those at first time. There are several completely free and open source java forums, there are several completely free J2EE Servers (JBosss, Jonas, Sun ONE to name but a few). As for JChat, I never used so I can't say anything but I am sure you could find something similar that's free.

    Remember that we built the CodeNotes site two years ago. Most of the free alternatives we examined weren't production quality back then. We were happier paying a small fee and using a supported third party product than we were with writing our own or waiting for open source versions. Jonas and Sun ONE weren't around then, Tomcat (version 3.0) was ok, but not great and neither was JBoss (not sure of version back then). If we rebuilt the site today aiming for "totally free" we could definitely find matching products in the J2EE space.

    >   You hit the nail on the head here. That's exactly what happened. You already paid the costs for using Microsoft products so any incurring costs seemed like overkill to the overal budget. However I *believe* if you didn't already pay for .NET license the costs would have been far lower using Strust and JSP.

    Entirely possible. As I said, in our particular circumstances, conversion made sense. For a small site built on a completely free stack, license cost would obviously not be your primary reason for migration. You still might convert if, for example, your website was bought out by a company that exclusively used Microsoft Servers. Rather than rewriting your site from scratch, or maintaining a unique system, the JLCA gives a means of reusing as much as possible. In the case of CodeNotes, a fully Microsoft stack *is* essentially free because we already account for that cost as a part of our regular (mixed environment) training and consulting business.
     
    >   I am just wondering. Did you also have to pay JLCA license fees or was it also included in the .NET license?

    The JLCA is built into Visual Studio .NET (version 2.0 is an extra install). You don't need to pay for an extra license for it. Although the JLCA has a command line version (instead of the wizard), I think you still need VS.NET (license) as opposed to just the .NET framework (free). I'll have to look into that some more.
  25. Reduced costs...[ Go to top ]

    Rob,

    For larger sits, lincese and profit are an issue.

    You said
    1. Windows 2K/2K3 - Same as before (The JSP version runs on Win2K now, the ASP.NET on Win2K3)
    * Linux is free, it is illegal to charge for Linux, you can download RH9 for free, as is oppen office.org. You can save a fortune by just using OpenOffice.org and linux.
    2. SQLServer - Same as before
    *postgresql.org is free
    3. License for VS.NET - As a company, we already pay for this anyway, so it's a wash as far as CodeNotes goes.
    * Struts is free
    4. License for Jive - Gone
    * Open souce
    5. License for JChat - Gone
    * easy to write
    6. License for JRun - Gone
    * Tomcat is free.

    So if instead of 100's of concurent users, you had 10,000 or more concurent users, it would cost you. Large sites that have to pay say a $250,000 license, usualy turn to theri R&D and say, can you guys write this for $150,000 so I do not have to pay a vendor license?
    Currently, there is a lot of competition/choices in Java, so I know I will have great stuff for a long time.

    If you look at licenseing costs.... I would avoid .NET.

    Lets say you want to re-sell a war application. Instead of sending out Tomcat/PostgreSQL/Eclipse/Struts/JSTL/iBatis I have to sell each of my clients MS licnese for .NET, and MS SQL and IIS, etc. Also some large clients ban Windows for internet, becuase of security. (So issue of cross platform is valid, larger sites tend to be Unix/Linux, etc.)
    So what is my profit?
    For profit, it makes more sense Java.
    First of all it is a larger market than MS:
    http://news.netcraft.com/archives/web_server_survey.html

    If you could save your organization money, I would do it.

    .V
  26. Reduced costs...[ Go to top ]

    Because the license for Windows covers IIS, we can eliminate the JRun fee. We also drop the Jive and JChat fees. As for other non-cost licenses, we don't save any money and possibly pick up cost for internal development and maintenance.


    It is true, nobody of us work to reduce your costs, you must do it yourself.
    If you can do it with autoconverter, do it.
  27. Reduced costs...[ Go to top ]

    Rob: Moving from Struts didn't eliminate any 3rd party license. Moving CodeNotes from Java to ASP.NET resulted in the elimination of third party licenses for Jive Fourms, JRun, the JChat module, and the Lucerne Search engine and I may be forgetting one or two more.

    It didn't eliminate those licenses. You had already (hopefully) paid for them. You don't get any money back. And considering you were running on a total of one server, you would have been better off with a hosted admin'd server running an OOTB PHP package for a total cost for $20/month.

    Frankly, I just don't get it. I don't understand why you hacked together such a complex single-server (or more probably a Dell desktop stuck in a DMZ?) system in Java, and now why it was hackiformed into a .NET world. Perhaps it would have been better served on a mixed main-frame / Tru64 environment using EDI, CORBA, web services, 3270 terminal emulation, NEC Teradata databases, VB and Tcl.

    Was the company paid to do this? Were there any incentives given? Or was it just a training exercise?

    Peace,

    Cameron Purdy
    Tangosol, Inc.
    Coherence: Easily share live data across a cluster!
  28. Reduced costs...[ Go to top ]

    Cameron:

    > It didn't eliminate those licenses. You had already (hopefully) paid for them. You don't get any money back. And considering you were running on a total of one server, you would have been better off with a hosted admin'd server running an OOTB PHP package for a total cost for $20/month.

    No, we didn't get any money back, but we also stopped paying for upgrades.

     
    > Frankly, I just don't get it. I don't understand why you hacked together such a complex single-server (or more probably a Dell desktop stuck in a DMZ?) system in Java, and now why it was hackiformed into a .NET world. Perhaps it would have been better served on a mixed main-frame / Tru64 environment using EDI, CORBA, web services, 3270 terminal emulation, NEC Teradata databases, VB and Tcl.

    Struts and DAO is hardly hacked together. And the emphasis placed on this particular point in the case study is not how I would have chosen to describe the project. We're running on an actual server class machine in a co-location facility (in both versions). And I'll freely admit that the original site was a bit of a frankenstien in terms of the various pieces we added in. However, when we built it two years ago, we had to go with (as most companies do):

    1. The systems we had available
    2. The skill sets we had available
    3. An agressive development schedule

    Was it perfect? No. Is this a "best practices" example for either Java or .NET? No. Did I ever claim it was? No. It's a pretty common example of the types of Java development I've seen in the companies I've worked with. A decently designed site that has a few rough edges.

    >
    > Was the company paid to do this? Were there any incentives given? Or was it just a training exercise?

    We (Infusion Development) were paid to write the JSP to ASP.NET migration guide (independent of the conversion). We also received incentives in the form of being part of the Early Adopter program, which gave us greater access to the JLCA development team. Beyond that, it's just your typical case study.

    Once again, I'm not holding up CodeNotes as a shining example of how to do things in either Java or .NET, much less as a comparison of which is better. We had valid reasons for conversion, we had access to the JLCA, and Microsoft wrote a case study about it.

    At this point, I am going to stop answering questions about why we did it or whether we should or shouldn't have done it. The J2EE vs .NET, single vs multiple vendor, free vs fee, and ejb or not horses have been beat to death way too many times on the ServerSide. If anyone wants to know more about how we did it, or what does and doesn't convert well, I'll still answer those questions.

    Rob
  29. Reduced costs...[ Go to top ]

    Thanks for the clarifications. (I wasn't trying to start a mud war either.) I haven't looked at the JLCA since it was pre-1.0. I'm actually kind of curious now if I could test it on some "large piece of software" that I have lying around ;-).

    Peace,

    Cameron Purdy
    Tangosol, Inc.
    Coherence: Easily share live data across a cluster!
  30. Reduced costs...[ Go to top ]

    Thanks for the clarifications. (I wasn't trying to start a mud war either.) I haven't looked at the JLCA since it was pre-1.0. I'm actually kind of curious now if I could test it on some "large piece of software" that I have lying around ;-).


    No problem Cameron. As for testing it on 'some "large piece of software" you have lying around', I'd say go for it. If you've got VS.NET 2003, you can get the JLCA off of the link at the top of the thread. If you run into any issues, post to the www.CodeNotes.com forum on migration (it'll be up tomorrow).
  31. Reduced costs...[ Go to top ]

    At this point, I am going to stop answering questions about why we did it or whether we should or shouldn't have done it. The J2EE vs .NET, single vs multiple vendor, free vs fee, and ejb or not horses have been beat to death way too many times on the ServerSide. If anyone wants to know more about how we did it, or what does and doesn't convert well, I'll still answer those questions.

    >
    > Rob

    Rob, the problem is, this is a Java forum. No one (well, almost) wants to know how to port their java system to .Net. If it were the other way around, it would be ok for us. This is not a .Net forum, so basicly most of us would want to understand WHY someone would want to do it, not HOW.

    It´s like selling a Ferrari-to-Lamborghini converter in a Ferrari Forum...

    Regards,
    Henrique Steckelberg
  32. Reduced costs...[ Go to top ]

    I agree...this is a Java forum...I'm sick of the microsoft croud lurking around here telling everyone they should try out all these .NET tools. Maybe microsoft wants a little of the same?

    SHOULD WE ALL GO OVER TO THE .NET WEBSITES AND TELL THEM HOW THEY SHOULD TRY OUR WONDERFUL NEW JAVA UTILITIES? IS THAT WHAT YOU WANT MICROSOFT?

    Keep it up and that's what you'll get.

    Maybe we can get Rick Ross to start a movement to mobilize 1000's of Java developers to go to all of the .NET websites and tell them how they can 'upgrade' to Java.

    Microsoft plays dirty. So can we.
  33. From Michael:

    > I agree...this is a Java forum...I'm sick of the microsoft croud lurking around here telling everyone they should try out all these .NET tools. Maybe microsoft wants a little of the same?

    Sorry Michael. I don't work for Microsoft and I didn't post the original news release. I joined this thread because I have been a ServerSide reader for a few years and thought that someone involved with the CodeNotes conversion could help clear up any misconceptions caused by the press release.

    If you'd rather talk exclusively about Java, then perhaps a thread on a Java Language Conversion Assistant is the wrong place (even in a Java-centric forum).

    If you don't like the thread in the first place, you should probably complain to Sandeep Deth (the OP) or to the staff at The Server Side.
  34. Reduced costs...[ Go to top ]

    No, we didn't get any money back, but we also stopped paying for upgrades.


      Out of curiousity, won't you keep paying for .NET upgrades?
  35. Reduced costs...[ Go to top ]

    No, we didn't get any money back, but we also stopped paying for upgrades.

    >
    >   Out of curiousity, won't you keep paying for .NET upgrades?

    Sure, but as I mentioned previously, our compnay pays for the Windows OS, SQL Server and Visual Studio .NET licenses as part of routine business. It's a cost we'd already be making regardless of whether CodeNotes was Java on a free stack or ASP.NET on Windows.

    Quite simply (and hopefully for the last time), we did the conversion because:
    1. We reduced costs - Admittedly, we could have done this in other ways, and the costs are pretty minimal. We didn't save a ton of money and any dollar comparison I make will immediately get an "apples and oranges" tag.
    2. We reduced interlocking third party product problems - Again, we could have done this other ways and our cost is picking up some development ourselves.
    3. We got to be an early adopter of a new technology - This is a huge benefit to our primary business (training and consulting). If Sun (more likely IBM, BEA or Oracle) came out with a C# to Java converter, we'd want to paly with that too.

    The CodeNotes case study is not a performance comparison or a price comparison nor does it provide any substantiative numbers about whether J2EE or .NET is better. In our particular circumstances, conversion made sense. It may not make sense for you. That's it. End of story.
  36. Reduced costs...[ Go to top ]

    The CodeNotes case study is not a performance comparison or a price comparison nor does it provide any substantiative numbers about whether J2EE or .NET is better. In our particular circumstances, conversion made sense. It may not make sense for you. That's it. End of story.


    I'd save my breath if I was you, mate. These people simply can't hear what you're saying. You say, "We believe we had good reasons to do it. It worked for us. It may or may not work for other people. The choice is yours." They hear "blaah blaah blaah Microsoft blaah blaah blaah .NET" and off they go. It's like a conditioned reaction, they just can't help themselves.

    I've never used .NET and probably never will, but I thought the article and your comments here were interesting. Thanks for giving us some sensible information in the midst of other people's invective and rhetoric. As far as .NET goes, I don't have an axe to grind or a bee in my bonnet or any other metaphors up my sleeve. Now I've stuck my head above the parapet, I daresay I'll get a stern telling off for not believing that .NET is the work of the Devil and that all who touch it should be burned at the stake, but there you go.
  37. Question on article[ Go to top ]

    It seems to me that the guy is completely confused and doesn't know much about thir web app architecture or/and Java server technologies. And if it's so it's good they ported to .NOT. MS products are well known as easy to use.

    Good job guys. You might want port the app back when EJB 3.0 comes in. Sun promisses EJBs will be much easyer to use :-))
  38. Question on article[ Go to top ]

    It seems to me that the guy is completely confused and doesn't know much about thir web app architecture or/and Java server technologies. And if it's so it's good they ported to .NOT. MS products are well known as easy to use.

    >
    > Good job guys. You might want port the app back when EJB 3.0 comes in. Sun promisses EJBs will be much easyer to use :-))

    It is very trivial to port portable application without any kind of tools and workarounds, just run it on Win if you want. Have you tried to port .NET application and run it on UX ? It is no meaning to convert and break portable applications, It must be mutch more usefull to convert win32,COM code. Do you have this kind of tool ( win32->.NET )? Why are you trieng to solve not existing problems first ?
  39. Question on article[ Go to top ]

    I wonder who will have AOP as standard first Java or .Net? If Java is first, I'd like to see Microsoft's little conversion tool convert that!
  40. Question on article[ Go to top ]

    I wonder how they have time to write a tool such as that.

    There is an AOP article at MSDN. They basically consider all the declarative services for the components to be aspects. Of course, you cannot define your own aspects but metadata is a way to add an aspect to a component. So we have them in Java too :).
  41. Question on article[ Go to top ]

    It seems to me that the guy is completely confused and doesn't know much about thir web app architecture or/and Java server technologies. And if it's so it's good they ported to .NOT. MS products are well known as easy to use.

    >
    > Good job guys. You might want port the app back when EJB 3.0 comes in. Sun promisses EJBs will be much easyer to use :-))


    I can't comment on the guy who wrote the article being confused. As the project manager, I can definitely tell you I am not confused as to what we did. Our web architecture did not use EJB's in the first place, and the only comment I recall directly making was that EJB's aren't converted by the JLCA (version 2 at least) so if your site depends on EJB's, it might not be a good candidate for conversion.

    The lead developer who performed the conversion was primarily a Java developer with little direct experience in .NET. The fact that he could use this convrsion engine to convert a reasonable two year old J2EE web application (Struts, JSP, Servlets, Taglibraries, but no JSF or Filters) says that yes, it is pretty easy to use. I don't see that as a negative.

    Our original application used an MVC approach with data access objects and was nicely structured. The JLCA helped us maintain all of that structure and we directly reused all of the JavaScript, CSS and look and feel. Once we were in C# and .NET, with a working application (Phase 1 conversion), we started using ASP.NET pieces that don't have direct matches in J2EE. Such as the validation controls (I know you can use tags to do this... but you can't easily drag and drop a tag and set two properties) and the web controls to help cut back on the JavaScript.
  42. I'm surprised no one has written a tool to create JNI wrappers around every single durn .dll on your Windows machine.

    That's basically what they're trying to do here.

    Steve
  43. I'm surprised no one has written a tool to create JNI wrappers around every

    > single durn .dll on your Windows machine.
    >

    That would take forever and it would always leave the JNI-wrapper writer playing catchup with whatever Microsoft does. A better approach would be to do something more general: write a utility that could import any DLL (or COM object or Unix shared library) and create a JNI wrapper for it. That should handle most of user's needs. (Microsoft has a similar utility for importing COM objects into .NET.)

    Once this is done, the static utility could be turned into a library and integrated into Javassist (see http://www.csg.is.titech.ac.jp/~chiba/javassist/) to allow users to dynamically link to DLLs/COM objects/Unix shared libraries. (.NET's version of native methods has this capability).

    But is there a real need for such a general utility or library? Java is very much a world on it's own. For nearly all applications I've written, I haven't found the need to write a single JNI call. Even if I did, chances are high that the SWT has already written a JNI interface for me, so I won't bother re-inventing a nonportable wheel.
  44. Such tools already exist.

    http://danadler.com/jacob/ (Open source, but you need to write wrappers for each of the com functions)

    http://j-integra.intrinsyc.com/
  45. Maybe M$ should leverage their knowledge of java and J2EE (they will be there soon if they will convert entire EJB apps) and implement a J2EE server and a cool slick IDE and sell them under competition price. I think they will make more money with this scenario.

    Regards,
    Horia

    P.S. And maybe after a while unveil them into the open source. :)
  46. As the project manager for the CodeNotes conversion, I'd like to add a few things before this thread gets too long. Before addressing the Java aspects, I'd like to say a few things about the JLCA.

    First, it worked really well and did a very nice job of converting JSP's, Servlets, Tag libraries and most of the back end Java code. The few areas where conversion failed relate to aspects of .NET that don't have direct correlations in Java (and vice versa). One particular challenge was replacing the struts controller functionality. The JLCA converts servlets to asp.net pages (with no graphical components). However, the struts controller is better modeled as an HttpHandler in ASP.NET. Also, we had to rewrite most of the database access as the JLCA converts JDBC to similar (but often non-ideal) ODBC calls. Because we isolated all the database calls in our original application, this conversion step was actually pretty easy.

    I personally don't think of the JLCA as just a toy for .NET afficianados. It's actually a really nice alternative when you have small areas of Java code in an otherwise .NET environment, or when you want to try out .NET without sacrficing your nicely designed Java application. All of our MVC and DAO patterns came through unscathed and the site works just as well in .NET as it did in Java.

    With regards to the "claim" that "maintaining the system was a nightmare", I can't tell you where that quote came from. However, we did include maintenance as a major reason for the conversion. Because of the time and budget constraints imposed during the original site development we used a variety of different vendor products, including JRun, Jive (forums), JChat (chat room), Lucerne (search engine) and several others. Although this choice made our initial development much faster, it did impose a serious maitenance burden. Several times we were forced to give up new features because of incompatibilies between the different products. We also had trouble when we tried to upgrade because the site was developed using J2SE 1.3 and J2EE 1.2, whereas the newer product versions used J2SE 1.4 and J2EE 1.3. Although it may not seem like a big issue, we had several problems becauase of incompatible method calls.

    The truth is, we could have gained the same advantage by going to a single company solution (such as Oracle or BEA or IBM). However, our database was already in SQL Server (we had the license available), our host machine was already Windows (again, we had the equipment available at the time), and we were interested in seeing how well the JLCA would work. By performing the conversion, we removed almost all of our interlocking product problems at the possible cost of vendor lock in.

    In our primary business (consulting and training) having more options for solving integration problems is better than having fewer options. I see the JLCA as another alternative for the all too common scenario where you have mixed environments. Sometimes, especially with the JLCA, the cost of converting is less than the cost of creating web service interface layers, messaging systems or other "plumbing" to make your Java talk to your .NET.

    Is it a one touch system to take JSP/Servlets and make it ASP.NET? No. Will it ever be a one touch system to make J2EE into .NET? No. But it is a pretty decent alternative and makes such conversions possible.

    I'll be happy to answer any technical questions about either our Java or .NET website, or how the JLCA works.
  47. MDA[ Go to top ]

    A question:

    if you want to transform J2EE to .NET why don't you just start with MDA?:
    - Build transformation cartridge from code to J2EE PSM (maybe this is already available?)
    - Build transformation cartridge from J2EE PSM to .NET PSM
    - Build transformation cartridge from .NET PSM to code

    I think if you take this way:
    - You use standard (OMG conform).
    - You can sell those cartriges to other people which are interesting to build .NET applications, so you can reach a broader market ;-)

    Check out the discussion about MDA at the main page of TheServerSide.

    Regards,
    Lofi.
    http://www.openuss.org
  48. Rob,

    Did you also convert Lucene et al. to .Net? I was wondering, are there any licensing issues? do you have to release the converted code back into the open source community?

    cheers,

    Joost
  49. Rob,

    >
    > Did you also convert Lucene et al. to .Net? I was wondering, are there any licensing issues? do you have to release the converted code back into the open source community?

    I is "safe" to convert and "close" Apache code,
     but you can not use original name. We do not see any problem and we will be happy, if somebody can make
     this code "better" in "secret" product.
    I do not trust autoconverted or autogenerated code. Try to convert Tomcat, who will support it, autocommunity ?

    >
    > cheers,
    >
    > Joost
  50. Although I have run some Apache code through the JLCA, we chose not to convert this code for several reasons. First, other people are actively porting some of these projects to .NET (e.g. nLucerne on sourceforge). Second, we had to replace other pieces that weren't open source (e.g. Jive Forums). Our replacement for Jive was the Forums Starter Kit from www.asp.net, which already included a search engine that we slightly modified to also search the articles (instead of just the forums).

    We are actually planning on releasing the CodeNotes source code on the JSP to ASP.NET Migration website on MSDN. That should happen in the next week or two.

    (Insert I am not a lawyer caveat) Reading the licenses, both the ASP.NET Starter Kits and most projects under the Apache license are open, but not GPL. That is, you can do almost anything you want to the code as long as you maintain the attribution and don't charge for the software (without first negotiating with the license owner).
  51. Please give the full disclosure[ Go to top ]

    Hi Rob,

    Please, before it's too late can you give the full disclosure?
    Ex: Relationship of the company with MS. Privileges regarding educational and consulting provided by codeNotes, proportion of business the company makes on MS related products vs J2EE, etc...

    I think theserverside readers will be very enlightened by this info.

    Thanks.
  52. well at least JCLA works[ Go to top ]

    It's an interesting read, if only because it shows JCLA works pretty well. Codenotes wasn't a trivial codebase.

    p.s. If anyone notices my name on codenotes.com, I left Infusion in May 2002. using Struts was partly my idea, back in the day. It was a good site. :sniff:

    Cheers
    Stu
  53. well at least JCLA works[ Go to top ]

    The worst part of J2EE is easily EJB, and it will keep embarassing Java developers. (as will JSF embaras us I think, Java Everywhere is a product strategy, not an architecture.)

    However, J2EE without EJB is great.

    Struts + JSTL + iBatis (Resin or Tomcat) is faster and cheaper than .net.
    It is also faster to develop a Struts app than .NET.

    But http://asp.net/webmatrix/guidedtour/getstarted/tooloverview.aspx !
    And Borland is shiping great C# IDE ($65!) so we have some great new alternatives, to presure JCP with.

    This is great for developers, more choices!

    In my opinion, JavaScript to XML-RPC and Resin/Tomcat is current best practice. But since I do want to support .NET and J2EE with my client... answer is (light weight) Soap.
    I plan to use new version of this:
    http://www.macromedia.com/devnet/mx/flash/articles/firefly_components.html
    so that I can talk via soap to Struts FormBeans(and iBatis) or .NET.

    Avoid EJB and JSF, use Struts & iBatis w/web services, and Java will look great.
    I user the people from CodeNotes, that I could show them then or now, that J2EE is faster and easier to develop with, or anyone, even if I have to do it via a web conference.

    .V
    self promo: baseBeans.com has some nice links on cheat sheet page, and I also teach Struts.
  54. well at least JCLA works[ Go to top ]

    The worst part of J2EE is easily EJB, and it will keep embarassing Java developers.


    They are not so bad for distributed systems if not overused for persistence (iBatis is a good choise for data access).
    There are a lot of good use cases for EJB too and I believe .NET style metadata
    will make this technology mutch better too.
  55. EJB's the worst part?[ Go to top ]

    I really cannot understand why EJB's should be the worst part of J2EE. To me they are the core along with servlets. For building web sites, ok, they're no use, but enterprise applications are a bit more than just a web site. I think EJB (stateless session, mdb) are just great for building an enterprise service architecture. In fact, I'd like to call stateless session EJB's Service Components or something. As Roger Sessions put it, Java has too much "terminal cuteness" that obscures things.

    BUT the spec is too complicated and there are too many bloody component types. Keep it simple, stupid.. Well, too late for that. I wish we had a spec like Tuxedo for Java. A REAL simple service interface, and that's it.

    My 2c,

    //ras
  56. well at least JCLA works[ Go to top ]

    I'd say not that EJB sux. just Entity. And Statefull session beans a little bit, but they have their uses tho.
  57. well at least JCLA works[ Go to top ]

    Avoid EJB and JSF, use Struts & iBatis w/web services, and Java will look


    JSP 2.0 makes web development much easier than struts. Just add 2 or 3 classes to create a framework much easier than struts to develop
  58. More advertising for .NET[ Go to top ]

    I just have to laugh everytime I read these so-called 'java' websites and see all of the blatent advertising for .NET. This reminds me of the recent JavaWorld article where they showed you how you could 'upgrade' to .NET. Are there none faithful left?

    Until Microsoft offers FREE UNRESTRICTED (JBoss) use of their .NET for production purposes on a FREE OS (Linux), with FREE kick ass development tools (Netbeans, Eclipse, ANT, etc.), microsoft can go home. And, oh yea, it's all got to be open source.

    I'm not saying Java is perfect, because it isn't. But it pretty much kicks ass for zero dollars and lets me have to freedom to choose appserver/OS, etc.

    Maybe somebody should write a .NET to Java conversion tool. Then we could go to the .NET websites and tell them how well the tool works, and how they can 'upgrade' to Java.
  59. Michael,

    Maybe you can educate me on free tool like this in J2EE:
    http://www.asp.net/webmatrix/guidedtour/getstarted/tooloverview.aspx

    or how to do this with free source in J2EE?
    http://www.datagridgirl.com/customcolumns.aspx

    Open Source should/will treat C# same as Java is my point.
    There are ports to .NET on the way:
    http://marc.theaimsgroup.com/?l=struts-dev&m=105567816628944&w=2

    .V
  60. More advertising for .NET[ Go to top ]

    If open source developers will find it usefull,
    they will implement it, we do not waste time for things we do not need.
    I do not think it is pragmatic to lock yourself in platfrom, OS, PL
     just because you prefer mouse as input device for development.


    C# is used by open source developers, but it looks like VB and Assembly are more popular than C#.

    As I understand the most of OSS developers prefer old good C/C++ or JAVA just because it is not platform specific, yes in practice C is more platfrom neutral than C# at this time. Assembly is more platfrom neutral too, is not it ?

    http://sourceforge.net/softwaremap/trove_list.php?form_cat=160
  61. Does that tool run on Linux? Mac OS X? I don't know of any open source java tool that does that, but there was SilverStream which had the exact thing you are showing there, and it was available YEARS ago. We are migrating AWAY from the SilverStream drag and drop because it locked us into their proprietary page architechure which limited us greatly.

    DataGrid, bound columns? Any programmer worth his salt can write one of these. We have written multiple implementations of this where I work. If you are not in the microsoft world you can easily figure out how to build these things. If you are used to 'drag and drop' everything, then I could see how this could be confusing.

    My point is that Java already has everything you need to build kick ass systems, runnning on a completely free stack. (OS, App Server, VM, frameworks, dev tools, etc). No it does not have drag and drop everything, but you are are addicted to that kind of stuff then you've got some serious issues as a developer.
  62. Why is it that Java costs zero dollars? Where do you people work? Our company is switching from Microsoft to Java and the costs are astronomical. As a fairly large company, you do not choose an open source app server, you go to IBM or BEA and spend 15-25k per cpu. And then you buy the development tools like Visual Age from IBM for 6k per seat. And my biggest gripe is all this nonsense about platform independence. Java performance is a joke unless you are running it on UNIX. Maybe if I was writing packaged commercial software, Java would be my choice so I could share code libraries across platforms, but the reality is most developers are solving business problems and integrating with existing sytems and there is no way a Java solution is cheaper, faster to develop, and easier to maintain than a Microsoft solution.
  63. Zero dollars? That is easy.

    Linux $0
    Sun VM (or blackdown) $0
    JBoss or Tomcat $0 (or even Sun ONE App server for 1) $0
    Struts or (framework of your choice) $0
    Netbeans IDE (or Eclipse) $0
    CVS $0
    ANT $0

    That is how it is zero dollars. Your company obviously does not know about open source software.

    repeat after me:

    Microsoft can't touch that, and never will.

    Platform independence a problem? I regularly develop on Mac OS X and deploy to Windows, and vice versa...no problems as long as you are not doing stupid things like hardcoding references to the D: drive, etc. It has worked quite will for us across platforms.

    Java a performance joke? The Sun VM on windows runs quite nicely (unless you are talking GUI, and that is another story). I would bet that 99% of the time performance 'problems' in Java are due to stupid things like bad string concatination and the like. I see this ALL THE TIME from C programmers who think they can just program in Java the way they did in C. Stupid things like that can absolutly kill the performance of a Java program (i.e. string concatination in a large loop using +=). My point is that Java is fast (tried 1.4.2 on windows?). Write good Java code and it is damn fast. Don't blame Java for stupid programmers.

    Java not faster, easier or cheaper? Get a grip bud. You could not be more wrong. You can go back to renting your software from Microsoft...I'll stick with the FREE java stuff.
  64. Agree,
    JAVA is not "slow", stupid code is "slow", and I do not believe stupid code becomes faster after transformation. It becomes more stupid: broken portabilty, bugs introduced by converter, unreadable generated garbage.
  65. More advertising for .NET[ Go to top ]

    Agree,

    > JAVA is not "slow", stupid code is "slow", and I do not believe stupid code becomes faster after transformation. It becomes more stupid: broken portabilty, bugs introduced by converter, unreadable generated garbage.

    I agree completely. That was one of our concerns about using the JLCA in the first place. We had a reasonably well architected website and were afraid that a "keyword tranlsator" would completely wreck the structure. What we found is that the JLCA actually maintained all of our code organization and isolation, and that made the rest of the migration even easier. If we hadn't already had an MVC pattern, DAO for data access isolation, specific XML helper classes, etc. then the conversion would have been very ugly.

    Clean Java code stays mostly clean (with exceptions for the parts you can't directly convert). Ugly code in Java will end up worse looking code in .NET after a JLCA conversion. Good code is good code, regardless. Bad code is bad code. You can write it in any language.

    And I will say that your code will probably become slower immediately after a conversion (or migration to any new platform). Why not? You have code optomized for one system that's being translated to another. You have to tweak all kinds of things to improve performance on your new platform. That's true regardless of whether you are talking about moving from BEA to Oracle or Java to .NET.

    That's one big reason why you won't ever see performance numbers out of the CodeNotes conversion. We're automatically talking apples and oranges if for no other reason than that our JSP server is a three year old computer and the ASP.NET server is brand new (fater processor, more memory, etc.). Not to mention the change in OS from Windows 2K to Win2K3.
  66. By: Juozas Baliuka in response to Message #89382.
    > The worst part of J2EE is easily EJB, and it will keep embarrassing Java developers<
    ”They are not so bad for distributed system”
     
    IMO, EJB are relatively not good at distributed.
    Compare to Web Services(Axis, Soap), XML-RPC or Live Connect, that are heterogeneous, allowing JavaScript. Etc. to talk to Java, for Rich UI.


    By: Ras Tafari
    “I really cannot understand why EJB's should be the worst part of J2EE. To me they are the core along with servlets. For building web sites, ok, they're no use, but enterprise applications are a bit more than just a web site”

    One of the reasons that EJBs are worst part of J2EE, (IMO), is that anyone objective and remotely experienced can conclude that anything on .NET is much better (faster, cheaper, more scaleable, easier, etc.) than EJB!
    (Also iBatis, Hibernate (and web services) are much better than EJB but that is OT to your point. That is why we have the Pet “Cemetery” alternatives and we flame EJB’s.
    ).

    So a reasonable client that is using or considering EJB *SHOULD* go to .NET because it is better than EJB!

    If you are a Java developer, EJB will and does erode our client base (not just because of this paper, but I have other anecdotal evidence that clients convert from EJB to .NET at a high percentage, and objectively… again, they should).

    Look at the Bitter EJB posts on here, and other EJB comments on Google. Look at Jboss looking for alternatives to EJB in 4.0 (also on TSS. IBM and BEA should also dump EJB, for the good of J2EE.)

    I am surprised, frankly, some of the failed projects and companies that did not file a lawsuit against Sun et al. since EJB was positioned to be “Enterprise” apps and technical papers and other evidence is out that they are relatively not scalable and to complex. I do not think that EJB JCP and BluePrints/Pet Cemetery JCP “Experts” ever built a production web app. The emperor has no clothes!

    I am not saying that you (or I) can’t build a site that uses EJB! I am saying that after you succeed in building it, it would have cost you to much and it would be not as scaleable as most reasonable alternatives.

    Of course, clients that use Struts+iBatis+JSTL find that .NET is worse than J2EE (.NET is slower, less productive, more expensive, less powerful than J2EE (w/o EJB.)

    As to Enterprise apps, let me try to define 2 types of web apps:

    There are corporate web applications that have 20-100 concurrent users doing CRUD. (is this enterprise since it is used by and organization/enterprise? This fits on a laptop for a server!)

    There are commercial web apps that have 10,000 + concurrent users doing CRUD. This is served out by Rack 1U or Blade (mostly Linux as far as clients I see) in 19” racks and racks.

    Clearly neither is EJB candidate, but first one, corporate could be .NET.

    And then lest define a web site.
    Apache! Static!

    Very few are using static web sites, because everyone want dynamic sites, portals and CMS. Therefore CRUD, therefore one of the 2.

    The old sales pitch by Sun the HW company, is you need to buy 3 servers:
    One for Apache, One for “Tomcat”, and One for EJB.
    This helps meet sales quotas. So if your architecture is done by a sales guy, OK. But if it does not work, consider legal action.

    An architect will tell you that it is more efficient and simpler to just have a “Resin” app server, since it can serve the occasional HTML page, but it also does JSP, Struts, DAO, etc.

    I take pride in what I do, as a Java Developer. And most clients/organizations are mixed, they have some .Net and some J2EE.
    When the C# people point at me an say “ Oh, you are the J2EE, EJB, BluePrints, PetStore guy. Yeah we tried that (with a I am naïve tone). So you like Java Server Faces for UI? Yeah, we read the Spec.”
    This I find embarrassing, because that is the perception of corporations out there. I do not use any of that!
    My DAO (iBatis) is better than .NET. My framework (Struts) is more elegant than anything on .NET. My rich UI (FlashStore+FirweFly) can talk to services (.NET or Sturts FormBean) is better than .NET. If you want to do it right, hire an engineer or an architect, to save you time/money (one at baseBeans.com, but lots of others)

    As an architect I have to try to be objective and I see the .Net pluses! Everyone should buy the $65 C# Borland IDE.

    I hope EJB gets deprecated, it is embarrassing.
    (I hope JSF gets deprecated for same reasons, complex, and not Java Everywhere).

    Of course, I do not see EJB deprecated in the future, unless people do file the lawsuits. So a lot of the corporate developers will go .NET. But large commercial apps that are run by Linux types will be J2EE w/o EJB.

    .V
  67. EJB Flame # 38,457 and counting[ Go to top ]

    MO, EJB are relatively not good at distributed.

    >Compare to Web Services(Axis, Soap), XML-RPC or Live Connect, that are >heterogeneous, allowing JavaScript. Etc. to talk to Java, for Rich UI.

    J2EE1.4 brings Soap bindings to EJB

    > One of the reasons that EJBs are worst part of J2EE, (IMO), is that anyone >objective and remotely experienced can conclude that anything on .NET is much >better (faster, cheaper, more scaleable, easier, etc.) than EJB!

    In what way is .Net better, in Transactions? in distributed environments? using Messaging systems? integration into legacy systems?

    >Look at the Bitter EJB posts on here, and other EJB comments on Google. Look at >Jboss looking for alternatives to EJB in 4.0 (also on TSS. IBM and BEA should >also dump EJB, for the good of J2EE.)

    Look I won't say that EJB is great but it has matured considerably, MDBs, Stateless Session Beans rule, mixed with JDO J2EE becomes is a killer of note!


    >Look at the Bitter EJB posts on here, and other EJB comments on Google. Look at >Jboss looking for alternatives to EJB in 4.0 (also on TSS. IBM and BEA should >also dump EJB, for the good of J2EE.)

    Please don't make me start saying what I think of Marc Fleury's personal vision

    >There are corporate web applications that have 20-100 concurrent users doing >CRUD.

    I know EJB applications doing 10 times that running on a shared mainframe.

    >There are commercial web apps that have 10,000 + concurrent users doing CRUD. >This is served out by Rack 1U or Blade (mostly Linux as far as clients I see) >in 19” racks and racks.

    Ok you got me there I I don't personally know any EJB sites doing that only know that our web site handles that amount and it's written in poor old Struts running on plain old Tomcat talking to Oracle all on
    on one rather old Sun box which never goes above 10% usage at peak, granted it's not EJB...


    >The old sales pitch by Sun the HW company, is you need to buy 3 servers:
    >One for Apache, One for “Tomcat”, and One for EJB.
    >This helps meet sales quotas. So if your architecture is done by a sales guy, >OK. But if it does not work, consider legal action.

    If your CIO fell for that one, then he deserves to have had his project fail

    >When the C# people point at me an say “ Oh, you are the J2EE, EJB, BluePrints, >PetStore guy. Yeah we tried that (with a I am naïve tone). So you like Java >Server Faces for UI? Yeah, we read the Spec.”
    >This I find embarrassing, because that is the perception of corporations out >there

    Petstore was designed as a starting point, I agree that JSF was initially dissapointing until I realised that alone is't not that great but combined with JSP2.0 and JSTL it will rock!

    >I hope EJB gets deprecated, it is embarrassing.

    AOP will eventually make EJB redundant (but we need a mature and standardised AOP framework) EJB3.0 is begginning to move in that direction anyway with meta data, and Even EJB2.1 is great and shows a move in the right direction


    >As an architect I have to try to be objective and I see the .Net pluses! >Everyone should buy the $65 C# Borland IDE.

    PS: www.eclipse.org --or-- www.netbeans.org they're free :-)

    Look most people agree that EJB is a good idea gone wrong, EJB is maybe a forerunner to AOP having said that, most of your points were made irrelevant after EJB2.0
  68. |
    |So a reasonable client that is using or considering EJB *SHOULD* go to .NET
    |because it is better than EJB!
    |

    Oh dear...

    So, can you please explain to me how writing your own server in .net is better than deploying a component in an EJB container? Can you show me .net's appserver?

    Opinions are like arseholes. Everyone has got one...

    -Nick
  69. So, can you please explain to me how writing your own server in .net is better than deploying a component in an EJB container? Can you show me .net's appserver?


    I disagree with the previous statement (that if you write EJBs you should switch to .NET), but .NET uses COM+ services). There is no need to write your own container, its there. And its much easier to code these distributed components in .NET (that is stateles session beans and message driven beans). As far as I can tell, there are no equivalents to entity beans and stateless session beans.
  70. There is no need to write your own container


    If you want to use .net remoting, you have to write your own server - much like you have to with RMI...
    Agreed, its better than RMI, because you can use attributes to mark your methods to use the COM+ transaction manager etc.
    BUT you still have to write your own server.

    -Nick
  71. There is no need to write your own container

    >
    > If you want to use .net remoting, you have to write your own server - much like you have to with RMI...
    > Agreed, its better than RMI, because you can use attributes to mark your methods to use the COM+ transaction manager etc.
    > BUT you still have to write your own server.

    It is about MTS, is not it ?
    As I understand MTS manages ActiveX components.
    .NET uses some kind of tool to port/convert .NET code to ActiveX controls, doe's not it ?
    Looks like very "interesting" technology JAVA->.NET->ActiveX ("Old good" win32 :)

    >
    > -Nick
  72. Exactly!

    In order to use MTS you have to revert back to using DCOM. (You dont want to go there)

    I say again, if you want to use .net remoting, you have to write your own server.

    Which, if you were to believe the anti-ejb offocionados like Vic, is a really cool thing to do.

    Its not.

    -Nick
  73. This may well be the best way to target .NET: develop it in J2EE, let Microsoft convert it. :-D

    I don't know if Microsoft is aware of what it's doing but it's actually encouraging people to go J2EE way until such a time that .NET makes sense, if ever, knowing a Microsoft tool will automagically convert the J2EE app to an assembly when the time arrives.

    I should be missing something...
  74. Automagically converting...[ Go to top ]

    I don't know if Microsoft is aware of what it's doing but it's actually encouraging people to go J2EE way until such a time that .NET makes sense, if ever, knowing a Microsoft tool will automagically convert the J2EE app to an assembly when the time arrives.

    >
    > I should be missing something...

    Having gone through a conversion process, I can say that the JLCA is not an automatic or automagic conversion. Sure, much of Java maps nicely with the .NET Framework and C#. But the JLCA does much more than simple keyword replacement. However, it has limits and you have to fill in the gaps. Finally, the JLCA is not a cross-compiler. You can't build a Java code base and instantly cross-compile to either a WAR or a .NET assembly.

    And to counter your argument...
    [ImaginaryMicrosoftRepSpeak]Sure, build in J2EE until .NET makes sense. You would anyway. Only with the JLCA the cost of moving to .NET is much less when or if you decide to. [/ImaginaryMicrosoftRepSpeak]

    Rob
  75. http://news.com.com/2100-1009_3-1026420.html?tag=fd_top

    Windows users should expect to have another update from Microsoft waiting for them on their computers.

    The software giant issued a patch Wednesday morning to plug a critical security hole that could allow an attacker to take control of computers running any version of Windows except for Windows ME.

    A group of Polish hackers and independent security consultants, known as the Last Stage of Delirium, discovered the flaw and worked with Microsoft to fix it.

    "It should be emphasized that this vulnerability poses an enormous threat, and appropriate patches provided by Microsoft should be immediately applied," the group said in an advisory posted to its Web site. The group said that programs designed to exploit the vulnerability will likely be available on the Internet soon.

    The flaw is in a component of the operating system that allows other computers to request the Windows system perform an action or service. The component, known as the remote procedure call (RPC) process, facilitates such activities such as sharing files and allowing others to use the computer's printer.

    By sending too much data to the RPC process, an attacker can cause the system to grant full access to the system.

    "This would give the attacker the ability to take any action on the server that they want," Microsoft stated in its advisory. "For example, an attacker could change Web pages, reformat the hard disk, or add new users to the local administrators group."

    Jeff Jones, senior director for Microsoft's Trustworthy Computing effort, said that, in addition to applying the patch, users and systems administrator should close down any unused communications channels, or ports.

    "Customers should protect their network with a firewall," he said. "Individual users should use the Internet Connection Firewall or some other personal firewall." The Internet Connection Firewall is a feature of Windows XP and Windows 2003 that limits the ways that a potential intruder could attack from the network.

    Ports are standardized software addresses that allow applications to exchange data. Firewalls routinely prevent access to such services from the Internet by blocking the specific port used by a computer to offer those services.

    Internet Security Systems, a network protection company based in Atlanta, warned its customers of the flaw on Wednesday. The company said in an advisory that it had raised its measure of the danger posed by threats on the Internet because of the vulnerability's seriousness.

    Microsoft is well into the second year of its Trustworthy Computing initiative. Aimed at boosting customers' trust in the company's products, the initiative has been both praised as a bold move to become a leader in security and criticized as largely ineffectual.

    Jones says the company is learning from its mistakes. In this case, Microsoft analyzed where the flaw crept in, and it developed plans to build in the expertise to detect it in the company's in-house development tools.

    "It was primarily a process issue," he said. "We will be updating our automated scanning tool to make sure this type of issue is detected in the future."
  76. Bill Gates: I will give you 1 mieellion dollars if you convert your site using .Net and 24 support to keep it up since its so buggy. Once convert you will tell the J2EE world that .Not is cheaper. Muhahaha

    CodeNotes CIO: Yes master !!
  77. Microsoft Aahhhhhh.
    Now-a-days, they don't work anymore on core ideas. They are busy with making utilities on converting everything in this world to Windows Compatible.