J2EE vs .NET

Discussions

EJB design: J2EE vs .NET

  1. J2EE vs .NET (123 messages)

    I am working with my company's IT Division to decide the future direction of the Architecture for our company.

    There is a debate between choosing between J2EE and .NET

    Any thoughts on these lines would be helpful. Please answer keeping following issues in mind.

    1) Which architecture is more robust for applications which require extensive search mechanisms and documents sharing and transfer and collaboration ?

    2) Which options will reduce time to market and cost to market or develop and cost to maintain applications ?

    3) Which architecture will useful for future flexibility of moving towards an ASP ?

    4) Does size of IT shop and in house vs. out house development make a difference of what you choose ?

    Thanks
    Abhi




    Threaded Messages (123)

  2. J2EE vs .NET[ Go to top ]

    I have spent a lot of time working with both technologies and feel relatively unbiased.

    I feel as though both technologies can satisfy your need for facilities which will provide robust support for document collaboration and searching.

    The second consideration you raise is difficult to address without a fundamental understanding of your organization. If your company already has a slant one direction or another then re-education and staffing issues may become an important issue in determining time-to-market and development expense.

    All things being equal there are a few factors to consider. If you choose to go J2EE then you must factor in time and cost in evaluating what implementation is going to be best for you; a small overhead but one that shouldn't be overlooked. I have found from the last three middleware projects I have been on that J2EE tends to incur more time costs from staff education due to the separation of specification and implementation. The developers must have understanding of both the J2EE API and the underlying implementation that you end up using. This is not entirely obviated in the Microsoft platform but information seems to be more concentrated. If you are able to come up with a full staff of seasoned developers in your chosen platform it won't matter... good luck!

    I have found that, in general, speed to market has been faster in the Microsoft world due to the simplicity of some of the development tools and the use of VB. The more complex the system, however, the better J2EE becomes. VB and COM will become overburdened by their simplicity and C++ COM is more time consuming then J2EE to begin with. Cost of maintenance is almost always in the favour of J2EE for reasons I will touch on in response to your third question.

    One thing I feel positive about: J2EE is much more successful in promoting code re-use. When I say code re-use I'm not referring to (in Microsoft parlance) component aggregation and extension. I'm talking about old fashion OO fundamentals. Microsoft is always touting binary (component) re-use and this is obviously good practice. I have read a myriad of articles and books where Microsoftonians say that traditional OO code re-use just never worked out that well; they obviously haven't implemented to many projects or architectures using a quality OO language such as SmallTalk or Java! I have found that J2EE projects almost always have a healthy share of both component reuse AND general OO reuse. Not every piece of code you write is going to end up as a component/EJB. My current project has a wonderful framework of Java classes that is constantly being extended through normal OO inheritance. Obviously this is possible in the Microsoft world if you choose the right language and are careful to enforce the right project policies but this happens far less frequently than in the Java world where it just seems to occur more naturally. For this reason alone I prefer J2EE for companies that are using an ASP model and require great flexibility.

    For your fourth question I think it is pretty clear that J2EE lends itself to heterogeneous environments much more than Microsoft's technologies. This can help if you need to out-source pieces of development to third parties.

    Well I've rambled on long enough and I'm sure there are a bazillion points I failed to bring up. Hopefully this helped some!

         -Will
  3. J2EE vs .NET[ Go to top ]

    I would be very interested in hearing about any specific reuse you have realized in J2EE or COM for that matter. Types of reused assets, framework description, time saved, etc. would be great.
    Thanks,
    Eric
  4. J2EE vs .NET[ Go to top ]


    would be very interested in hearing about any specific reuse you have realized in J2EE or COM for that matter. Types of reused assets, framework description, time saved, etc. would be great.
    >

    reused assests: in the J2EE architecture, you can build
    up beans(software components). And the beans can be reused
    in the same project many times, or reused in the future
    project.

    That's something similar to Microsoft controls/objects.

    Yes, Microsft have their controls/objects already and
    don't have to build them up. But I doubt that if the developers can alter their properties/characteric, or to
    transform the controls to a similar controls and rebuild
    it by reusing just part of the codes from controls.
    I doubt and don't know. Anyone who can verify this feasiblity would be great.

    If the reader have read though this page, people can conclude that J (J2EE or Java) have it's comparative advantage in its flexibity, more complex system, more
    favorable to J.

    So if something can be reuse, that's time saved.


    And the asking about
    -reuse
    -time save

    can be extended to another topics, the platform portability,
    and the stability.

    If the client really want one thing -- TRUST, to be stable.

    that thing -- TRUST

    is not ONLY built on PRESENT,

    it's also built on the PAST, the HISTORY.

    Do you think NASA will use Microsoft Windows to be
    its core platform ? and the visual developing tools
    for its development ?
    that's the point of TRUST.

    Right that commercial client might not consider this
    issue very much, but seems that they WILL NOT NOT needing
    that thing -- TRUST, trust that the system is stable, and hard to encounter common world tech crashing, like code red , their data still need privacy ....


    And if you have developed a very stable system with investment on a considerable of time (surely very related to $$), and can be deloyed to different clients with similar core business nature, they might not be just using 1 kind of platform, the MS platform. So ....
    are you asking them to change all their platform for your
    products ??
    If they don't wanna have their platform change and you don't wanna missed that client, then .... is it to REBUILD the
    system again from other development tools ??

    That's the point of portability, and stability, and
    thus make it related to the time saved.

    It can also be extended to -- the QUALITY.

    In management strategy, it's known that better quality
    means more time saved, and thus money saved.

    It seems to be contradicting to common view that
    for building better quality, more time and money must have to be spent, so how can both of them be saved ?

    It's true that this view can be fit on the investment on the first few periods of the whole cycle.
    But better quality means that less time has to be spent on the after sales support, and better customer loyality.
    That is also revenue brought.




    Right that the development cycle of using MS visual tools
    would be shorter, but that's a short term issue.
    For any long term system development, the system must grow
    more and more complex, so ..... more complex the system,
    more favorable to J.

    So if it's to put the sight on long term, J can make developer save time.
    Longer the the sight, the vision, more and more time saved
    (also considering the deploying of your product/apps/system to different OS).



    The articles on this page can make the reader conclude
    some points.

    -In short term, microsoft development tools make the development cycle shorter.

    -More complex system, more favorable to J.

    -J is platform independent and language specific.
    .NET is paltform specifc and language independent.
    (don't know if it's released and how's the reality of the words about language independence.)


    Any fans on .NET here ??

    gameboy0
    @: >
  5. J2EE vs .NET[ Go to top ]

    A bit OT, but I don't like the "MS tools have faster build cycles than Java tools" topic: Look at JBuilder:
    *) Want to build like with C++ tools? Sure, only use the editor and compiler!
    *) Want to develop in a VB/Delphi paradigm? Sure, use the databound controls (something I really cannot recommend, neither with VB nor JBuilder)

    This is simply not an issue of the language but the developent tools and the frameworks (I guess we all agree on this, though someone may say "yes, but in VB I have somne operators..." blah blah, simply build a method or use ony "Object" references... the differences are really minor). And as I said, both Borland and IBM, and probably other vendors as well, provide VB-like databound components.
    This is even easier to see with Java vs. C# as the languages are quite the same (and no sane developer will use features like unsafe code "just for fun"; Don't tell me the get/set (property) paradigm of C# is time-saving, its just clumsy).
    Even C++ can be used to develop "VB-like", use Borlands C++ builder (though you really have some differences there, but if you are good with pointers and the like it is really like VB).
    So this really isn't the point, and the Java libraries are one of the most powerful yet easiest to learn I have ever seen (try getting started with MFC and Swing *gg* Still, complex UI elements are lots easier with Swing, e.g. TreeTable).
    I also wouldn't compare J2EE with .NET. .NET is more like "Java + JavaBeans + JTA + RMI + JSP", there's a lot more to J2EE.
    You will also miss some really important things in .NET (one of my favourites: Container Managed Persistence. Slow? Don't think so, even Borland used CMP for their ECPerf results, Orion recommends CMP for performance... Only with LOTS of tuning BMP will be faster. BTW: You could use Assembler if you like coding and are paid by LOCs... and a custom database. and custom protocols. Why not write your own OS? Or implement your e-App in Hardware?).

    regards

    Messi
  6. J2EE vs .NET[ Go to top ]

    First of all Com is pretty much dead in the .net world. I hava been working with(or should I say learning) as much as I can concurrently with both technologies. In the .net world, com is no Longer(Only for backwards compatibility). Look at Web services for a platform independent solution and If you are looking at an Intranet or Extranet situation .Net Remoting, which scales better than web services and out performs them as well is the way to go. VB is now VB.Net which is really a cross between VB and Java while maintaining a VB like syntax(very easy to use). If you are looking at using .Net you must explore the remoting
    functionalities that are in the .net Platform. J2EE has it's strengths as well, but the more I work with .Net the more I see it as a viable option.
  7. J2EE vs .NET[ Go to top ]

    I guess there is not enough time to point out all differences and points, and I think I don't have the knowledge to do this.

    4) Yes, when your size is big your option currently is only J2EE I guess... ever saw a big company settling on Win32 Servers?

    Well, I now just wanted to point out some general points:
    *) Many vendors (Sun, IBM, Oracle, Borland...) have always tried to get something they can fight MS with. And now they have the chance of doing this with J2EE. Additionally billions of dollars have been invested in J2EE... Do you think they'll just throw this away in favor of .NET? So I guess vendor participation will be better.
    *) If you are unhappy with MS .NET implementation (bugs ec., I cannot say that MS delivers bug-free software) what will you do? The same thing as now, wait for the next release/SP. In contrast, you somewhat freely choose your J2EE vendor. If coding carefully you can often switch within some hours (we even had applications being transferred with some clicks).
    *) .NET is currently not available. Big and small companies need enterprise solutions _now_, so they buy J2EE solutions, and buy application servers. I think this will be the same kind of situation (but much "worse") as with databases today: if they have Oracle they'll refuse to buy DB/2, although they are quite similar, compared to .NET vs. J2EE. So I guess companies will refuse to buy your products if they are .NET, who wants to maintain 2 complete architectures? However, the same can happen to you with J2EE: If one customer say "WebSphere and nothing else" you may have no chance because you always used BEA.

    Alltogether I think that MS is too late at the moment, their architecture is two years behind and needs some time of real-world use to mature. I also think J2EE has gained too much industry momentum for .NET to have a chance. Consider that, at the moment, you have no option other than J2EE for server side programs (besides plain Java)... we have a monopoly like MSs in this area.
    However, I also agree with Gartner Group, who say that .NET will be adopted by small shops/vendors while J2EE will rule in the big shops, but the small ones with .NET will cease with time. Again consider: Here (and with Linux) is the chance of the industry to beat MS. Ever wondered why Java and J2EE were adopted so fast (well, they are good, but many things were good, CP/M was, GEM was, Netscape was...)?

    just my 2 cents

    Messi
  8. J2EE vs .NET[ Go to top ]

    Good reasonably unbiased comments. It's good to see less of the religious positions that some people take (unless you're independently wealthy, most of us are in this to get paid and support the kids :-).

    I would just add the following point about maturity of the technologies.

    .Net is based on COM+, which is based on COM/MTS. This means that many of the technologies are deployed to the mass market, given they are a tightly integrated part of the operating system. Yes, .Net is "NotYet" and is not likely to be until later this year, but when it arrives it will have the advantage of being a mass market release and like the CORBA/COM competition, is likely to be the most widespread use of these technologies. So the "2 years behind" approach is naive, most of the base .Net technologies were in place with Windows 2000 (COM+, MSMQ, queued components, IIS5/ASP+, SQL2000, now BizTalk for EAI) - .Net is a neat layer of additional services which makes life even more interesting. The same component-orientated infrastructure that is running the Windows 2000 operating system, is also used to run the applications (where as OS's like Un*x, given their 70's heritage, can't be re-engineered to be component based operating systems and thus this support is always vertical and not as tightly integrated).

    So, no matter which set of technologies are "better", the real challenge to J2EE vendors will be the mass-market coverage of the Microsoft platforms. Un*x platforms still have a big edge on single box performance, but are getting squeezed below from clustered Windows 2000 deployments (see www.tpc.org) and above from mainframes who are interested in getting more into these e-times.

    Lee.
  9. J2EE vs .NET[ Go to top ]

    First: I'm always trying to be unbiased, simply because my job is to develop good applications... not to evangelize. Its true, I favor J2EE... but I think I have good reasons to do so, and if I see .NET is cooler I'll switch.

    That .NET is based on components already there is true... BUT: The best "component" MS produces is (IMO) MS SQL Server. COM+, MTS etc. cannot really compete with products in the J2EE area; simply because J2EE is open I think: Everyone can integrate their products, the things they know best, like IBM MQSeries, BEA Tuxedo, Inprise VisiBroker... specialized products which do their job very well, integrated all together. This is also the case with SQL Server: With this database I think MS has created something that really stands out: as easy to manage as Sybase ASA and as powerful as Oracle and DB/2 together. But the really nice thing is "Using it with J2EE?" No problem.
    I know that most products will also be able to be integrated into MSs platform (and in fact they partially are), but I want to see this before I believe it (this is mainly again "J2EE is here now")).

    I also wanted to point out that the "multi language support" is IMO something only few companies will use. Simply because yes, it is cheao to get a development team, because you have the know how there (well, not really, see below): You make up a team of 1 COBOL, 2 VB, 2 C++, 1 C# and 1 Eiffel programmer. Cheap to develop the application. But who will maintain it? The COBOL programmer may leave... getting another one? Oops, quite expensive.
    In general, I think companies will choose to switch language instead of going into this maintainance nightmare. Because learning another language is quite "easy", the fundamentals are normally the same. And Microsoft will also enforce this by preferring C# to develop for .NET, the same way C++ became favored above C (although you could still use it).
    Well, and now to the "you already have the know how"... I guess e.g. a VB programmer will have a hard time learning .NET, because it is a different world. The problem is: Learning a language takes you a week, even a complex one like C++, if you are familiar with the concepts (VB programmers will have to switch to OO, so they don't know the concepts); From C++ to Java is 2-5 days. But try to learn MFC, Swing, OWL, VCL, Qt, STL, EJB, JDBC and so on (... .NET). This will take you a lot longer. I think what counts is the _framework_ you have to learn. And the framework, your class libraries, are completely different with .NET compared to Win32, MTS... So I guess the time it takes you to switch to Java/J2EE will be equal to switch to .NET (both from Win32, ODBC, COM+...).
    However, one thing I really like about .NET: The WebForms. This is simply great (but there is a Swing-like library that does the same for the Web).

    cu and hope to recieve many nice comments

    Messi
  10. J2EE vs .NET[ Go to top ]

    What is the Swing library you're referring to? I'd love to take a look at it...

    Thanks,

    Damian
  11. J2EE vs .NET[ Go to top ]

    Swing is a library for GUI widgets included in JDK 1.2 and above (spearately available for 1.1); It is really nice to use and far superior to AWT.
    But I think you know, guess you'd like to know of the "swing-like" library for Web-Apps (like WebForms); There are two of them, SPFC which is member of the Jakarta (Apache) project and thus open source(http://jakarta.apache.org/cvsweb/index.cgi/java-spfc/), but not much to see yet, and Swinglet (www.swinglet.com) which is commercial.
    I have not used any of it.

    Hope this helps a bit

    Messi
  12. J2EE vs .NET - Swinglets[ Go to top ]

    Swinglets was mentioned earlier in this thread. It's a toolkit for creating JSP/HTML/XML/WML pages from Swing-like components and data models.

    IMHO this is one of Java's best kept secrets. I've used Swinglets successfully in a couple of projects. I found the author to be very receptive to feedback, requests for enhancement and bug reports (the product was in beta at the time, but still very stable).

    The URL given was slightly incorrect. The correct URL is:
    http://www.swinglets.com

    Regards,
    Matt.
  13. J2EE vs .NET[ Go to top ]

    SPFC was not more than a concept. They just dropped it. Have a look at wingS (www.wings.to.com). It uses Swing Models, Swing/AWT Events and Listeners, a powerful Template-Mechanism and has an API that is very similar to that of Swing. It's already in a stable state, the concept has proven in a number of commercial projects. The License is LGPL.

    Holger Engels
  14. J2EE vs .NET[ Go to top ]

    Your points about the framework makes a lot of sense. But it is also important whether the 'framework' can take you from A-Z consistently. While j2ee is a framework, it also comes with a 'native' language for that framework. This means that regardless of how small or big your projects are, if you are following good practices, you will want to start with UML, choose an OO language that reflect well the domain models, use standard patterns for your architecture, and find a wealth of implementation that follows those standards. From that perspective, j2ee is a one-stop
    shopping center.

    In the MS world, my impression is that this is not there.
    There is a rich set of classes with the C++ world, but if you started that way, you can't end up with .asp which is closer to the VB world.

    While entry into the j2ee framework may seem high, the completeness and the consistency of the framework is a significant advantage when you are there.
  15. J2EE vs .NET[ Go to top ]

    My 2 cents. Lets forget for a moment about .NET and J2EE. In the long time what matters a lot is the expertise. When we choose Java centric technologies we just spend our time once to learn the complex enterprise API’s and that is it. Example would be JTA. Just imagine a MFC programmer goes for MOTIF and vice versa. We don’t have to spend our time learning libraries again and again. It is trivial. What is more important is to understand the complexity of the domain and how to design a good Enterprise level application. Why I would learn win32 API, UNIX system calls or any other platform dependant API if I can almost always code against one interface (Java API). BTW so far in platform wars no one is winner and being a developer we all have to write code for lot of platforms. I bet lot of us already did that. And honestly I don’t want to learn new libraries again and again for different platforms. So my vote goes to J2EE.

    RJ
  16. Wait wait wait...

    I don´t understand. why are you so concentrated in the underlying bussiness platform? Why not to concentrate in the platform bussiness logic?. don´t we ear that the mission of the underlying middleware is to isolate the programmer form the complexities (the plumbing) of transactions, storage and so on?. If this is the case, J2EE, .NET or a third, will be part of the OS for the next generation of developers.

    If we can not concentrate in the particular bussiness logic, the two platforms do not keep up with their promises. Something is failing in the concepts behing them.

    from my point of view a _third_ platform, application layer or OS will achieve finally this goal.

    maybe the simplicity is a heavier and better use of XML, and the intoduction of web concepts inside the application server and OS´s, the use of logic programming, messages instead of function calls, queues (pipes) instead of treads.

    exotic, itsn´t?


  17. Perhaps we need a co-operating system. Ab Initio.
  18. J2EE vs .NET[ Go to top ]

    .Net is based on COM+, which is based on COM/MTS.


     Bosh!!!
    Lee, I think you better now what are you talking about before posting here such a statement.

     Well, I had worked with MS COM/MTS and COM+ before I moved to J2EE, and here is my comment on it:

     .Net is completely different and is not based on COM. Even people from Microsoft claim it. That means that .Net isn't mature enough to consider it as J2EE competitor, and what is more it hasn't been realesed yet at all and is still in beta. I think .Net will be mature enough only when a few years will pass, and MS will need really BIG investments to make .Net mature.

     IMHO it's no good that MS suddenly dropped COM and created completely new middleware platform without leveraging COM. There are already enough component technologies and the world doesn't need a new one. In the beginning MS with COM was 2 years behind CORBA and (in spite of the fact that COM+ was near to catch up with it(CORBA, not J2EE)) I think, having minimum 3 years of gap in the moment, MS's .Net came too late...

    >> Un*x platforms still have a big edge on single box performance,
    >> but are getting squeezed below from clustered Windows 2000 deployments
    >> (see www.tpc.org)

     TPC is not a gauge! In fact some vendors (try to guess who) are even changing their DBMS's kernel appropriately to match TPC specification, while the specification itself is too simple and straightforward to reflect real business applications.(Do you really think that W2K/COM+/SQL Server will be 10 times faster than AS400/Tuxedo/DB2 in a REAL scenario?)

    >>.NET supports a large number of languages.
    >>.NET is the winner in this area.

    And what about CORBA? It supports really all languages including Java and has almost 10 years of maturity.

    Microsoft held a presentation is our University, and the speaker said that the coolest advantage that you will have when using .Net "multilanguaging" feature is that "you can have VB class that you can inherit in your C# class for example" =)))(Just imagine such a nightmare)
    All the time during this presentation I asked myself a question: "Why Microsoft couldn't just buy Sun's licences? It could be much more cheaper than create and propagate it's own."


  19. J2EE vs .NET[ Go to top ]

    Ho hum. Enter the religious comments. Shame, coz the previous posts were so mature. :-(

    Anyway, just to quickly give you something to think about with regard to your post. Yes, you are right that Net is a new set of technologies, but the point I was making (which you chose to miss) is that the underlying layers are still COM+ (because Windows 2000 is fundamentally so - for example, MSMQ, ActiveDirectory, Index Server, COM+ transactions - yes, still underneath there). And yes, CORBA is more advanced than COM+, I think no-one can argue that - the point is that of mass market distribution and that is how MS have managed to swamp often better options.

    Drink less coffee and _try_ to be a bit more agnostic!

    Lee.
  20. J2EE vs .NET[ Go to top ]

    IMHO ...
    .Net Is Based on COM+ ... download the .Net framework and the beta release of visual studio, build a web service and connect to it.... check out the objects that get built behind the scenes

    .Net is not COM+ though and you don't need to know COM to do .Net, but the infrastructure is using all kinds of COM-like object interactions. The big difference is the CLR.
  21. J2EE vs .NET[ Go to top ]

    IMHO ...

    >.Net Is Based on COM+ ... download the .Net framework and
    >the beta release of visual studio, build a web service and
    >connect to it.... check out the objects that get built
    >behind the scenes
    >
    >.Net is not COM+ though and you don't need to know COM to
    >do .Net, but the infrastructure is using all kinds of COM-
    >like object interactions. The big difference is the CLR.

    I agree, .Net applications will use (existing) COM+ runtime services for e.g. transaction management, role based security etc. .Net (the CLR) will bring another (a far better) component/object model to Windows development, a model which has IMHO features already found in the Java model (some remarks on this new programming model see: Don Box column House of COM on http://msdn.microsoft.com/msdnmag/issues/1200/com/com1200.asp) A quote from Tim Ewald's book Transactional COM+ , appendix A: "The CLR is a replacement for COM, but not for COM+. CLR classes can be configured to use COM+ runtime services the same way COM classes do. In fact you can mix both classes using both component technologies in the same COM+ based system. This works because the CLR is backwards compatible with COM. The .Net Framework SDK simplifies COM+ programming...."
  22. J2EE vs .NET[ Go to top ]

    .NET is not based on COM/DCOM/COM+, it is a completely new model very different from these but suspectly similar to the Java in some ways (e.g. namespaces for packages).

    It's obvious why M$ is moving away from COM, DLL hell.
    In large applications, having x number of different versions of the same component becomes very tedious to manage.

    An important point to consider is that .NET is mostly vapour ware and hoping it will be as successfull as IE is not realistic as IE is only a desktop application while .NET is supposed to be an enterprise devlopment platform.

    my 0.02c anyway...
  23. J2EE vs .NET[ Go to top ]

    Since we've heard from the unbiased folks on this subject, I thought I throw my biased $.02 in.

    First, my biases:

    I am biased towards type safe languages that promote good programming practices and maintainable code (I feel C# and java both meet this criteria, C++ and VB do not)

    I am biased toward IDE/Tools that allow me to develop fast, native l&f type apps for my user's OS of choice and that means Windows. I haven't talked with 1 customer (insurance industry) who uses anything else.

    I am biased toward whatever architecture will allow me to reach my target audience in the quickest most flexable manner. This means distributed via the internet overwhelmingly via http.

    Next my j2ee/.net take:

    I am currently heavily invested in the application of the j2ee technologies. The architecture is good, the tools suck. My productivity in the tools sucks. I'm currently paying the price for choosing SilverStream (heavy indeed), but I don't see the other app server / tools combinations being that much better. There is a reason that the java programmers keep falling back on emacs/textpad. What choice do they have. Really.

    I saw the architecture, ide, SOAP web services in the visual studio .net demo and for a change I got excited about the tools. An IDE that will let me write code, compile quickly and debug, within the SAME tool, without going to lunch every 10 minutes for 5 minutes at a clip. (Excuse me while I restart my SilverStream server and ide for the 10th time today...)

    On top of that, when I am ready, I can deploy my web services in a manner that will allow other VS.net developers to call, see, and code to (with intellesense) my web services, all from within the IDE. I know this will get beat up on the security side of things, but, java or C#, SOAP is SOAP. The security issues are there. The difference as I see it is that VS.net appears to make this so easy that the security issues must be dealt with now rather than later.

    Java or C#, whatever. They are the same language. I like them both. The idea of multi language projects is laughable for the maintenance nightmare it presents. Eiffel would be cool, but, well, every time I champion it the other developers start looking at me funny.

    My concerns about .net platform independance loom in the backround married to my scalability concerns.

    I will cautiously plod down the j2ee path, but keep a sharp eye on .net hoping to to recover some of the productivity I use to have in C (DOS days) and PowerBuilder (C/S days). Who knows, I may even be able to build a distributed app in a real GUI again, instead of banging my and my users heads against these mindless, stateless browser interfaces.

    </rambling>

    Bill Pfeiffer
  24. J2EE vs .NET[ Go to top ]

    First: I'm sorry that your J2EE experience was signed by bad tools. I don't know SilverStream, but had a good impression from what I read; But I know very well what you mean... I restart WebLogic every 15 minutes.
    However, you may want to try JBuilder (which isn't exactly cheap either because/but it comes with Borland AppServer) for a really fast development environment (of course, some things must be done, e.g. generating stubs, but three times faster than for e.g. BEA); you could exchange AppServer against Orion or jBoss, which do not require stubs to be present and the deploy cycle is even faster then.
    Still, I think the best IDEs you can get are Java IDEs. For file based I like JBuilder best (this one is really great, especially the newer versions), but if you prefer a "modern" repository... well, VAJava does the job. And you can still look at Forte, WebGain, Kawa, AnyJ, IDEA... I think compared to VisualStudio they are all superior (and all with IntelliSense :-). And the AppServers... well, some vendors will have to learn, others got it.

    But you are exactly right, many things should be done for the tools to be better... but your conclusion is 100% correct, let's always keep .NET in mind.

    greetings

    Messi
  25. J2EE vs .NET[ Go to top ]

    The problem with IDEs is that those tools cater to the non-technical developers. I view them as crutches/training wheels. If you don't know how to ride a bike, training wheels come in very handy. They prevent you from falling down. But, if you already know how to ride, they are a hindrance.

    Same is with those VB-like IDEs -- if you know how to design your classes properly, the IDE will just be in your way (as it doesn't really support proper design practices). But if you don't know much about the OO design, the IDE may help by assuring some crude level of coherence in your code.

    The fact remains that about 70% of the software development effort goes into thinking. The remaining 30% is spent on implementing/deploying. IDEs cannot help you think better. They may assist you in some areas of implementation/deployment, but if you really know what you're doing, an IDE is simply an annoyance. There are many people, however, who are choosing to switch from being office workers (i.e. sales reps, accounting staff, etc.) to building software. They lack proper formal training in software development, but some vendors give them the encouragement by offering them an IDE. VB championed that, and millions were convinced that they could easily switch to developing apps. Unfortunately, people who really know how to build apps are often called in to clean up the mess after these guys have had their go with IDEs.

    In the long run, I don't think it's worth the trouble.

    Alex
  26. J2EE vs .NET[ Go to top ]

    This may be true to a certain degree, however... I think I _can_ design properly, and I can work with and editor and the JDK. I also agree with the useless VB like IDEs.
    I however, disagree completely that IDEs just hinder you. I'm sorry, but I like the syntax highlighting, I like a project structure pane, i like to see the structure of my classes, I don't want to edit DDs by hand, and neither do I want to code remote interfaces by hand. I enjoy UML diagrams, I like the graphical debugger, I like building UIs with some mouseclicks and so on.
    So I think if you would take the time, and get used to the IDE it can save you a lot of time and will not be in your way. It will help you a lot.

    Messi
  27. J2EE vs .NET[ Go to top ]

    I however, disagree completely that IDEs just hinder you. >I'm sorry, but I like the syntax highlighting


    You can have the syntax highlighting in a simple editor. You can download plugins for the editor which will cover the syntax for the variety of languages.

    >i like to see the structure of my classes

    The best way to do that is to build UML diagrams. You don't need IDE in order to do that.

    >I like the graphical debugger

    Here I must agree with you... Although, I found out that if I do my OO design properly, there's basically nothing to debug.

    >I like building UIs with some mouseclicks

    Ugh... I was never able to find a decent EDI that would do the legwork for me when it comes to deciding on the Layout Managers. If I have to do the layout coding by hand, then IDE is just a charade. Many IDEs would simply generate the code by imposing the absolute coordinates on the components, and that's not what we want to do in Java, is it?

    >So I think if you would take the time, and get used to the >IDE it can save you a lot of time and will not be in your >way. It will help you a lot.

    Believe me, I've tried. I've even completed IBM's courses on Visual Age for Java. I wasn't impressed. I mean, even such a powerful tool as VAJ would interpret my layout of the components by setting the layout to null and then using the absolute pixels. That's ridiculous! I would expect that such an expensive tool would take my layout and would apply some rules to decide which layout managers to deploy. But, no such luck.

    And also, I've noticed that it tends to generate a lot of crappy, redundant code. So, looks like I'll have to wait for a more advanced tool.

    Alex
  28. J2EE vs .NET[ Go to top ]

    Syntax Highlighting: Agreed
    Structure: Well, and I guess you build the UML diagrams by hand? Don't get me wrong, I like UML and some IDEs offer exactly this... switching from UML to code seamlessly (Together, may not mention the other, NDA). And with "structure" I meant I don not need to search for a file or method, but it is shown in the structure pane and I can simply click on it... some overview is always good I think.
    Debugger: Sorry, disagreed... although I believe in proper OO design I'm no Uerberhacker, so there is always something to debug (and may it be just Java's or the appserver's fault).
    UI: Again, don't get me wrong here, but with JBuilder I can quickly design a UI prototype and "fine tune" it then. BTW: I know it is not the "correct" way, but I use null (or Borland's XYLayout) often for dialogs... as it is so much faster designing the UI then, especially for simple "Input dialogs"... and I can live with that if I translate the app I'll have to correct some button sizes. I however know that I would never design the main UI with a GUI editor... it is simply not good enough.
    VAJava: Okay, agreed... although I'll probably start a flame war now I wasn't impressed by VAJava either. It was more in my way than it helped me, so I believe you tried and abandoned it absolutely. I think VAJ is far too "sophisticated"... the developer no more has control over the files and the code, and I do not see an advantage in the "pure OO" and "repository" approach.
    You should probably take a look at JBuilder Enterprise, I like this one very much. But it is even more expensive ;-)

    regards and hope you have luck with JBuilder

    Messi
  29. J2EE vs .NET[ Go to top ]

    JBuilder is considered as one of the best :
    try
    www.borland.fr/presse/2001/cmq_JavaOne.asp
  30. J2EE vs .NET[ Go to top ]

    Regarding the Layout vs IDE discussion, I personnally like the JBuilder approach a whole lot...

    Design what you want in XY coordinates then convert to GridBagLayout. I find it extremely elegant.

    Add items, align things as they should look, apply GridBag, make a few minor adjustments and effortlessly have a form which is resolution independant.

    my .02 :)
  31. J2EE vs .NET[ Go to top ]

    I´m biased towards Microsoft
    Let me tell you about the irrational part which biases my trough, because if we have to be honests, we have also to explain ourselves deeply at the full level.

    There are good things and bad things in this fight.

    The right part is the right fears of middleware developpers and companies about the Microsoft dominance. this is good for the market as long as competition is.

    The not so right part is to hate microsoft because MS does make a real effort into simplifying things, making today system administrators and programers tasks a matter of secretaries and application programmers in the future.

    Such a thing is good for ordinary people life and for the society as a whole, but, at the same time, plays against some people and companies thay does the opposite game: "keep it complicate to maintain power and money".
    And the last strategy is a dangerous option in the long term.

    for that reason, I like java and I don´t like J2EE
    I rather prefef HTML over GUI(people too, you developers!) and I like XML over relational overbudget-overhead.(and so on)

    Do SUN, ORACLE etc plays the "complicate and win" strategy? My answer is: maybe, at least this is what I smell in the atmosphere"

    What I reconmmend for J2EE companies is a good usability laboratory.


    Alberto
  32. J2EE vs .NET[ Go to top ]

    Believe me, I've tried. I've even completed IBM's courses >on Visual Age for Java. I wasn't impressed. I mean, even >such a powerful tool as VAJ would interpret my layout of >the components by setting the layout to null and then >using the absolute pixels. That's ridiculous! I would >expect that such an expensive tool would take my layout >and would apply some rules to decide which layout managers >to deploy. But, no such luck.


    Have you tried Visaj from IST? It's a pure Java GUI builder that doesn't enforce its own layout model. Take a look before giving up on GUI IDEs. (I've only built a small Java GUI with the product, but have built several large-scale Motif GUIs with their X-Designer product on which Visaj is based)
  33. IDE's[ Go to top ]

    Alex,

    I would like to highlight a few issues regarding IDE's. there is no doubt that devlopment is about thinking. An IDE job is however to free time resources for the developer to concentrate more on development logic, than using them to worry about Standards compliance and other such issues. An IDE helps the developer by generating infrastructure code, implementation code and allowing him to focus more time on the actual business logic code.
    If you looka the IDe's market, new IDe's that are very focussed like say on J2EE are coming yp so fast. They offer a standards compliant development environmnet which helps increse productivity. These no way make the job of the Developer less important, in fact they increase his importance by providing him better ways of doing it.
    Hope i made my point.
  34. J2EE vs .NET[ Go to top ]

    Bill --

    Found your comments very enlightening...thanks for sharing your experience with us regarding Silverstream. I've actually been considering Silverstream for a new project and would love to hear more details about your difficulties with the product? Is it that the tools are weak and undependable? How is the server itself?

    No one has written a review of Silverstream in the Reviews section of this site - it would be great if you could take a minute to post your thoughts to help other developers before "we make the wrong choice"....

    Thanks in advance!

    Damian
  35. J2EE vs .NET[ Go to top ]

    I just wanted to make a quick reply to Bill Pfeiffer tools problem. I also am sorry to see that you have had problems with the tools. I've had a much better experience. I've used both WebGain Visual Cafe 4.x Expert and Borland JBuilder Enterprise (expensive) with JRun, Weblogic(just a test run), Tomcat, and JBoss with excellent success.

    I especially like JBuilder since it is written in all java and will also run on Linux/Solaris. Debugging EJB's is also really easy with both with Inprise, JRun and JBoss. On the other hand, JBuilder is expensive, and it wouldn't run well it on any 'puter that has less than 256 meg RAM (I personally would get impatient with anything less than 512 meg RAM).

    anyway, I know this is sort of an off-topic aside, but I just wanted to let Bill know that there is hope for the tools.

    -Jason
  36. J2EE vs .NET[ Go to top ]

    Bill,

    >>... in a real GUI again, instead of banging my and my users heads against these mindless, stateless browser interfaces.<How nice and soothing to see that some people have the same opinion as oneself about _that_ subject...

    Bernard
  37. J2EE vs .NET[ Go to top ]

    Bernard Thouin wrote:

    Bill,

    >>... in a real GUI again, instead of banging my and my users heads against these mindless, stateless browser interfaces.<How nice and soothing to see that some people have the same opinion as oneself about _that_ subject...

    ---------

    Here, here! What is it going to take to move everybody to real internet applications?

    God bless,
    -Toby Reyelts
  38. J2EE vs .NET[ Go to top ]

    Toby,

    >>What is it going to take to move everybody to real internet applications<Probably a guy with a brilliant idea, but mostly with a superhuman charisma, not Bill Gates, NOT Sun, and who will convince the whole world that a browser is the worst invention on the surface of this planet when it comes to GUI capabilities, and that we should finally either go back to completely centralized solutions with graphic clients, or to "half-fat" (compiled, flexible) clients....:-)

    Cheers
    Bernard
  39. J2EE vs .NET[ Go to top ]

    Toby,

    >>What is it going to take to move everybody to real internet applications?<BTW, I thought afterwards, there already IS a good solution to that problem (and I already hear the horror shouts): Citrix.

    That allows a complete, singing and dancing, good-old-times, GUI-rich Windows application to be used from within a browser or on its own, but it runs on a central server, and you only see the presentation part of it, so to say, on your PC.

    So think about it that way:
    - no redevelopment AT ALL
    - full Windows GUI capabilities (multiple windows, all mouse actions, "block mode" or not with highest speed
    - no fiddling with app servers, Java, JVM, EJBs, CPM, session/entity beans, persistence, ... you name it
    - no stupid, slow browsers
    actually, only advantages: take you good old client/server, 2 or 3-tier app (be it a VB, PowerBuilder, Delphi, Centura, whatever one)... and put it on the Net !

    OK, it's clearly not "web-native", but so what ? Most mortals (our typical users...) wouldn't notice the difference anyway, except that they will LOVE a real GUI...

    :-)
    Bernard
  40. J2EE vs .NET[ Go to top ]

    BTW, I thought afterwards, there already IS a good

    >solution to that problem (and I already hear the horror
    >shouts): Citrix
    How well does it scale? Can it support all applications without any modification?

    >actually, only advantages: take you good old client/server,
    >2 or 3-tier app (be it a VB, PowerBuilder, Delphi, Centura,
    >whatever one)... and put it on the Net !
    Taking your 2-tier client/server as is means inheriting the same scalibity problems, even worse you need at the Citrix level high capacity hardware to support a lot of concurrent sessions and at the database level you also need high capacity hardware to support the amount of concurrent database connections (which equals the number of concurrent Citrix seesions because there is nog resource pooling like on a appl server). What would be the sollution to the 2tier scaling problem? Move to a 3- or n-tier architecture, meaning "- no fiddling with app servers, Java, JVM, EJBs, CPM, session/entity beans, persistence" is not entirely true.
  41. J2EE vs .NET[ Go to top ]

    Regarding the discussion about the use of remoting software for decent client-side graphical performance. One vendor mentioned is Citrix. However, there's a company called GraphOn that I used to work for that has an equally excellent solution. Last I heard, GraphOn's performance
    numbers for their remoting technology either meets or exceeds those of Citrix. There may be a whitepaper on their site, I'm not sure. GraphOn has been in the remoting graphical front-end business, one shape or another for the last 15 years or so. They've gone public only recently.

    However, they have excellent solutions for remoting both
    Windows and Unix/Linux applications into a browser. The browser support is in the shape of an ActiveX, Netscape Plug-in and Java applet versions. If a company needs to make an application available over the internet, or to it's
    employees at home and it's a not a priority to re-engineer
    it into being a web-based application, then I would seriously check GraphOn's stuff out. From what I've heard, it's an easier solution to implement then dealing with Citrix requires a developer/admin to jump through.

    In fact, GraphOn has online demos you can try out. I was
    impressed with the latest performance enhancements and
    functionality. Just go to http://www.graphOn.com and
    look for the online demo links...

    Andy K.
  42. J2EE vs .NET[ Go to top ]

    .NET is released. Any people has tested it, or
    some feedback, sounds about it ??
  43. J2EE vs .NET[ Go to top ]

    This is my 6th time installing it WITH rebuilding the box.
    And for what it's worth, when it crashes it bring down the whole damn os.

  44. J2EE vs .NET[ Go to top ]

    Guess what? It IS the OS *gg*
    Well, part of it, together with Windows, Office, Outlook, Internet Explorer, some components you are getting via AutoUpdate, ... ;-)

    have fun

    Messi
  45. J2EE vs .NET[ Go to top ]

    what's the meaning of 'gg' ??

    : >
  46. J2EE vs .NET[ Go to top ]

    n n, I have been reading ur mails in other threads also. what are u - neurotic neophyte? I hate getting personal on a tech thread, but ur comments are juvenile, hostile, inaccurate, and way biased to say the least - extremely irritating and very provocing.

    Far from what u have described, most people's experience with .NET has been more than satisfactory. I have now been using a machine where VS.NET Beta co-exists with my other apps like SQL Server, Office and VS 6.0 for the last 6 months or so, without a single hitch. Have a look at the tonnes of source code available on sites like codeproject, dotnetjunkies, gotdotnet, etc., A number of commercial web-sites are already running on .NET. Just do a search on google to see the amazing amount of work being done on .NET. Do u think it would have been possible with such instability u speak of?

    Stick to tech discussions man - do not create FUD.
  47. J2EE vs .NET[ Go to top ]

    Bill,

    I too had worked with Silverstream and i can understand when u crib about its features. I have had my share of problems with other tools like Kawa's IDE also.
    Recently we had evalauted Pramati's IDE , which they call Pramati Studio and it turned out to be a very good tool for server side development. The tool has excellent support for EJB's and it comes with a test server.
    Its deploy tool is very easy to use.

    Miguel
  48. J2EE vs .NET[ Go to top ]

    .> The architecture is good, the tools suck.

    We must have different evaluation criteria: My opinion is: 'The architecture is good and there are lots of tools, some of which are excellent.

    I particularly like, JBuilder, Ant, and TogetherJ (UML, round-trip modelling for Java). I haven't tried it yet, but Embarcadero's Describe UML modeller looks great too. Also, XMLSpy is a neat xml editor that knows how to Work with a DTD. Perhaps you're looking for a universal IDE - i'm not. I like best of breed java tools that synchronise on the source code, leaving me the freedom to pick and choose the deployment options. That's why I like ant (an xml coded make script interpreter).
  49. J2EE vs .NET[ Go to top ]

    Hi.
    Bernhard mentioned a report by the Gartner Group where it analyzes the future of J2EE/.NET. I would very much like to see the specific analysis (although Bernhard already mentioned the conclusion).
    Does anyone know of an online copy of that report? If you do, please post the address here or email me at nerlin at inter dot net dot il.

    Thanks in advance.
    Gal
  50. J2EE vs .NET[ Go to top ]

    I guess I would agree with most of the points made here. Kuodos to everyone for not makeing this a religious war! Some thoughts:

    Main Points
    --------------------------

    - Size of project definitely makes a difference. EJB is a lot of work for small projects (also can be very expensive). For these, MS products make a lot of sense (as do JSP/servlet products). EJB is far, far better for larger, more complex, more "true application" work. I totally agree with the comments about code reuse and the benefit to an ASP model.

    - J2EE is obviously the choice if you are working across multiple platforms.

    - .NET supports a large number of languages. If you already have developers who know C++, you may want to continue to develop in that rather than change languages. .NET is the winner in this area.

    - Microsoft definitely has a strong edge in the general developer community, with a large number of people who already use and know the basic tools that will be extended for .NET (with the acknowledgement that many will have to update their skills to use the .NET platform). As such, it will probably be easier to recruit a development team at a lower cost. This is, however, changing all the time. Microsoft just does a fantastic job of wooing developers to all kinds of trade shows.

    - Microsoft has excellent 3rd party application support. This is true in terms of development tools and add-on products. EJB has, in my opinion, just come of age with the J2EE platform and development tools like TogetherJ that start to create a good environment for the developer.

    - Microsoft's .NET is not here, but you can never underestimate the power of Microsoft to charge back (ala taking out Netscape). With over $100 billion in the bank, they can pour money into the project.

    Secondary Points
    ---------------------

    - The XML support in .NET, in my opinion, is better than the support in J2EE. J2EE is rapidly catching up, but MS has definitely gotten into the web services/SOAP model and the tools are pretty impressive.

    - J2EE clearly has a strong head of steam behind it.

    - .NET is mostly unproven in large scale deployments. I say mostly because SQL2000 is being used in large deployments quite successfully.

    - 3rd party software vendors (companies like Broadvision, etc.) have already shown that they will probably support both .NET and J2EE. I think this will continue.

    All that being said about Microsoft, I'm spending a lot of time right now on J2EE development because I think that long term it's going to have a huge impact on development. I think we are heading to more of a programming assembly approach with J2EE and the web services model associated with .NET.

    I know it's a little silly, but right now I'm playing around with J2EE products under Win2k. Why? I like a lot of the stuff of Win2k, but I want the flexibility that J2EE provides.

    That's why I'm excited mostly about SOAP and web services because interopt becomes a lot easier with these technologies.

    Anyway, just my 2 cents. I'm not a religious zeolot on either side...I happen to think both platforms are pretty darn good and hold a large amount of promise.

    Damian
  51. J2EE vs .NET[ Go to top ]

    I have a slightly different take on J2EE vs. .NET.

    Component architectures are a means to an end. The end is a scalable, maintainable application that is also quick to develop and deploy. When looked at this way, there are actually 3 alternatives.

    1. "LAMP", or Linux-Apache-MySQL/PostgreSQL-PHP (good for 80-90% of web applications, the low to mid-range)
    2. .NET (not clear what the range will be, but definitely something for existing Microsoft shops to consider)
    3. J2EE (IMHO, overkill for all but the most complex applications)

    Many companies will stay with Microsoft, but there is a basic contradiction in .NET. If it is open, then you should be able to bypass Microsoft and get your products from any vendor or even Open Source project. It's doubtful if Microsoft will let that happen. On the other hand, if .NET is not open, then you should be extremely wary of going down that path, or you'll end up getting locked in.

    I think the low cost and extremely fast time-to-market that PHP offers will offset any advantages that a full-fledged component architecture may have. Implementors are generally not religious. They may like the elegance of an architecture, but if they get a 2- to 3-fold improvement in productivity with a tool like PHP, that's not something they can sneeze at. I'm a Java programmer myself and I should be expected to support J2EE against all comers, but I've been pleasantly surprised by the productivity advantages of PHP. For small to medium applications, you can get something running much faster with PHP than with JSP (my experience).

    So in my view, your company should also evaluate the PHP option in addition to .NET and J2EE. True, it doesn't pretend to be a component architecture like the other two, but the productivity you get should be experienced at least once. Who doesn't want to improve time-to-market?

    Regards,

    Ganesh Prasad
  52. J2EE vs .NET[ Go to top ]

    This is indeed turning out to be a healthy discussion. thought about chiipin in.

    Well for me the most important factor for picking a development platform is code maintainability and extensibility and obviously scalability and stuff come next. i have devoted considerable amount of time developing for both Microsoft platform and now the Java enterprise platform. and the latter definitely has the edge in that respect. i think microsoft is a great platform for writing applications real quick but if you have an application which consistently changes then Java as a language is the one to pick. in terms of performance i think microsoft can scale as well as J2EE platform with the inroduction of windows clustering services etc.

    i know someone mentioned up there that .NET is based on COM+. this is in total contrast to what i know and have read. and to me .NET is totally a new platform for writin your components. in fact lately read a Don Box article(COM guru) and the bottom line of that article was .Net is not COM at all. its entirely different with the introduction of CLR platform and programmers are encouraged to program and leverage the benefit of CLR. though microsoft is not stopping us to write COM libraries but surely they wont encourage you to code them. also .Net is being talked bout the platform for Web Services. What is .Net? .Net is all about Set of languages which get compiled to one comman binary language and then run within the CLR. and then u have the big piece XML/SOAP/UDDI which binds all these languages for RPC communication. also you also have ADO.NET which can import n export RDBMS data in XML format.

    I really dont know how can all these things can make .Net the platform for Web services. in fact java enterprise has all these things. the big binding piece namely XML/SOAP is all about how you implement it in your applications.

    i know we are all talkin about web service but no one knows how they look and how will they work and collobrate together. so its to early pick one platform as the one platform for web services.

    though i can say definitely say one thing seeing the kinda of work i am doin at the moment, J2EE clearly is "the enterprise platform" for now and i m eagerly awaiting to see what .Net has in store for the enterprise.

    kapil
  53. J2EE vs .NET[ Go to top ]

    IMHO...

    The two enterprise architectures at the moment are WinDNA and J2EE. The .Net platform will only be realized when things like ASP.net, C#, and ADO.net become a reality (i.e. SQL Server 2000 is a great product but it seems silly to call this part of the development Architecture). The timing of this makes me think of the MS PDC 3+ years ago when we were first got a taste of COM+, and then couldn't really use it until Win2000 Server SP1.... don't expect to really build .Net applications until at least mid-2002.

    In my experience developing several applications on both WinDNA and J2EE platforms, the two architectures break down this way in terms of advantages / disadvantages they offer...

    J2EE
    * JSP / Servlets: (+)Multi-threaded, (+)Faster than ASP, (+)Extensible Tags, (+)full Java language for tag classes and scriptlets
    * Language Support: (-)Java only, (+)fully object-oriented
    * Security Model: (-/+)not integrated with platform
    * Service-Oriented System Objects: (+)Stateless session beans are automatically pooled objects, (+)Stateful session beans automatically tied to user context, (+)Stateful session beans support persistence, (-) vendor implementation of persistence mechanisms are inconsistent and not part of the current EJB 1.1 specification
    * Service-Oriented Data Objects: (+) Container Managed Persistence promotes rapid development, (-) Container Managed Persistence deployments have limited portability for migrating between J2EE vendors which can lead to re-work, (+)Bean-managed persistence allows for legacy integration and manual transaction demarcation, (-)Bean Managed persistence promotes long and complex development tasks, (-)Entity Beans do not handle data retrieval at high volumes of data as well as a stateless data access layer, (+)Entity Beans are an elegant design metaphor for representing relational data, (-)Entity Beans must be used with care since the concept of an single enterprise data representation can create artificial concurrency bottlenecks that might otherwise not exist using a stateless DAL (i.e. composed of stateless session beans) against an RDBMS

    WinDNA
    * ASP / ISAPI: (+)ASP has reasonably good performance, (-)language features are limited, (+/-)loose typing
    * Language Support: (+)ASP supports a pluggable script interpreter so you can use VBScript, Javascript, Perl, etc.,
    (+)COM objects can be built in a number of languages, (-)VB, the most popular WinDNA language, does not support parameterized initialization or polymorphism
    * Security Model: (+/-)Integrated with platform
    * Service-Oriented System Objects: (+)COM+ deployment simpler than EJB deployment, (+)VB components can be built more quickly than EJBs, (-)no support for automatic object pooling unless you use C++, (-)limited support for stateful components, (-)difficult to achieve loosely coupled vb components across system and data tiers since ASP to COM+ interface cannot bind to secondary interfaces except through named calls (i.e. interfacename_methodName) which actually creates additional overhead (IDispatch is the only way for VB components to bind)
    * Service-Oriented Data Objects: (+)WinDNA maximizes scalability through advocating stateless components (i.e. application DAL components, base DAL components), (-)WinDNA lacks any convenient representation of persisted data as an object

    To me, the PHP route is valid only for applications where horizontal scalability is not a concern, and maintainability is not a priority.
  54. J2EE vs .NET[ Go to top ]

    Steve,

    Very good comparison, compliments!

    2 questions:

    > don't expect to really build .Net applications until at
    > least mid-2002
    do you really mean build or do u mean deploy? in your opiniun can larger projects start designing and building using beta1/beta2/RTM and deploying when the VS.NET product is stable enough (SP1)?

    (projects where development starts Q3/Q4 this year and where delpoyment will be Q3/Q4 next year, and projects where .Net as outlined by MS now may for several reasons be a valid choice)

    >In my experience developing several applications on both
    >WinDNA and J2EE platforms, the two architectures break
    >down this way in terms of advantages / disadvantages they
    >offer...
    supose we are at mid-2002 and .Net is reasonably stable, how does your J2EE vs .Net advantages/disadvantages list look like? I see in the J2EE advantages list and WinDNA disadvantages list some issues which will be addressed in .Net (a quick scan of your list Faster than ASP, Extensible Tags, full Java language for tag classes and scriptlets, fully object-oriented, loose typing, no support for automatic object pooling unless you use C++, limited support for stateful components, (-)difficult to achieve loosely coupled vb components across system and data tiers since ASP to COM+ interface cannot bind to secondary interfaces except through named calls (i.e. interfacename_methodName) which actually creates additional overhead (IDispatch is the only way for VB components to bind), WinDNA lacks any convenient representation of persisted data as an object)
  55. J2EE vs .NET[ Go to top ]

    supose we are at mid-2002 and .Net is reasonably stable,

    >how does your J2EE vs .Net advantages/disadvantages list
    >look like?

    The decision of J2EE vs. .NET becomes a lot more difficult and a lot clearer. Basically, .NET will be easier to use and deploy (I've been using Beta 1 and it's pretty good) when compared with J2EE (particularly with EJBs)...and J2EE will have the advantage of being able to deploy on a large variety of platforms.

    So, leaving out scalability (a huge question which neither J2EE nor .NET can answer right now), your decision will come down to whether or not you need to support multiple platforms.

    If the answer is yes, J2EE is for you. If the answer is no, then .NET will provide, imho, a significant productivity boost over J2EE.

    Some additional thoughts:
    - J2EE server prices need to come down. $10k a CPU makes the "deploy anywhere" idea a waste of time in terms of horizontal scaling. Who cares if I can put my J2EE on a $2k Linux pizza box if the license costs me $20k for 2 CPUs. Then add in the cost of a good development tools and you're spending a fortune.

    - J2EE will continue to be the fav of those looking for large scale enterprise applications. .NET will hit at the low and mid-market developers that make up a large percentage of all development done today.

    /d/
  56. - J2EE server prices need to come down. $10k a CPU makes

    > the "deploy anywhere" idea a waste of time in terms of
    > horizontal scaling. Who cares if I can put my J2EE on a
    > $2k Linux pizza box if the license costs me $20k for 2
    > CPUs. Then add in the cost of a good development tools
    > and you're spending a fortune.

    Absolutely. And I think this reduction in prices is bound to happen. The Open Source JBoss (www.jboss.org) is getting better and better, and also comes integrated with Tomcat for JSP support. It will be a very good migration path for those currently doing "J2EE-lite" with Tomcat. The only thing currently lacking in JBoss (as of mid-2001) is clustering support, but that should arrive in a few months, judging by the pace of development so far. Enhydra Enterprise is another Open Source J2EE server. It is currently in beta.

    With Open Source implementations entering the market, the cost argument against J2EE goes away. Of course, users will have apprehensions about support as with all Open Source products, but we know today that these issues are not show-stoppers and the risks can be managed by making arrangements for support, either in-house or third-party.

    The JBoss Group does commercial consulting and support for JBoss, just as Lutris Technologies does for Enhydra. Independent consultants seem to be mushrooming, too.

    In short, there are ways to go J2EE while keeping costs at reasonable levels. Open Source is a good alternative to "Open Wallet" solutions.

    Ganesh Prasad
  57. "Reboot the server every 15 minutes..." (I'm there. So much for deploying session beans into a test environment. I mean, to change one line of code you have to compile, jar once, jar twice, copy to server, restart server --- is this some kind of "choose your own adventure" loop?)
    I don't know of any good environment that integrates web client side development with the server side development. Just testing is a nightmare compared to C/S and application development. Reminds me of using a kernal debugger! Printing out lines seems the state of the art! VJ++/Visual Interdev did make efforts to allow server side and client side development to be run from one UI. Useful but, as a site includes SSL and becomes bound to other network config, becomes almost impossible to use.

    What did this VJ++ programmer do? Moved to J2EE. Well, the writing was on the wall for C++ anyway. I mean, C# is like VB. It's a no-brainer. What professional programmer wants to limit job options to the companies that use a specific vendor? You see, I think MSFT is making a huge mistake with not making a cross-platform solution. Stiffled innovation not only hurts the consumer but leaves a corporation with a niche product. Well, lots of mainframes out there still. Heck, I worked with a family of companies that has 10x mainframe programmers to modern programmers. (flame fodder:-) Not sure any of them actually write code though.

    >>What is it going to take to move everybody to real internet applications<My viewpoint is that the real limitation is bandwidth. Who wants to wait on Citrix much less a real application to download. (Or just use VNC or remote X, even:-) Of course, one can foresee Java app.s that reside on a computer and use peer-to-peer networking. Of course, the proxy server tricks like AOL IM employs are then necessary. A web browser is like vector versus raster graphics as far as I'm concerned. Sure, the protocol is not too efficient or elegant but it is simple. I still cannot stand to use PCAnywhere over a cable modem because of the delay; so, graphic intensive applications are probably a long way off; yet the Napster software does beg the question of why we don't just go back to distributed (P/P, C/S) app.s and shove everything through port 80 is some queer HTTP tunnel.

    IMHO, the main benefit of J2EE and COM+ is the ability to make distributed systems. Calling remote methods is easier. The idea of a data or object pool using entity beans is interesting too. The rest of the stuff (session, context, database access) is formulation of a specific design from a more general networked-software design field. In the web world, the design of GUI's has gone to artists and ad agencies. The "design" is now limited to nice graphics and colors and basic design principles are gone (functional grouping, simple user/advanced user, hot keys, integration, etc).

    As to features, I hope .NET will be on a higher plane than J2EE; yet, for now, the best tack seems to be to use open source. As to cost and future, maybe you can get really cheap by using Perl but this will require senior programmers and a real design process. For open source, consider the need to fix bugs is probably better addressed by open source than by paying $100/hr to some application vendor and talking to some new grad. Some are investigating using JBoss, Jonas, or another solution rather than WebLogic. And consider using Borlands Interbase rather than Oracle. (Of course, planning to buy power as demand increases rahter than spending x100k up front sort of admits the exponential growth numbers in the business plan.:-)

    As for me, I liked C++ but have never liked having ever changing technologies that are really no better than what I was already using.

    Back to topic. I don't foresee any real technical benefits of going one way or the other. I mean, ASP/COM/SQL Server development is basically the same as JSP/J2EE/Oracle development. I guess the Oracle tools are a little touger to learn. The real issue is probably that of learning. I don't see where the learning curve consideration favors .NET unless you have a boat load of VB programmers sitting around. Still, Microsoft's idea that you can write a object in any language is a good one and should be adopted by all. Anybody from J2EE listening?

    Well, want to know what I really want? I really want a nice wizard for creating the new user web site sign up. How many projects will I have to work on before someone creates template code? Maybe I've been using the wrong tools? What else I want is more generic web/server side components. (I've got a list so maybe I ought to quit and work on them.:-)

    Tim Jowers
  58. Well, want to know what I really want? I really want a >nice wizard for creating the new user web site sign up. >How many projects will I have to work on before someone >creates template code? Maybe I've been using the wrong >tools? What else I want is more generic web/server side >components. (I've got a list so maybe I ought to quit and >work on them.:-)



    This whole posting was very rewarding to read. But the finish really cuts to the real issue for me. I don't care if you are developing with JSP, ASP, or PHP (perhaps 99.9% of all web apps), you are probably rewriting the same "application" again and again - not even counting the thousands of others who have written this same "application". I know that I am - and I have sworn not to write another web application without first writing a meta-application which will generate the code for the real application(Tim Jowers if you are indeed pursuing this, please contact me: chrisNO at SPAMactiveinterface dot com). Imagine XML-based templates and configurators which would allow you to express the logic of three-tiered applications. The build-time program could generate to whatever ASP, JSP, or PHP design pattern which you have specified. Now I don't care about the choice of deployment platform - I'll leave that to the operators/administrators.

    I have heard of some work in this area in some Open Source projects. Perhaps people with more direct experience could fill us in on the status.

    Why are not any technology vendors filling this seeminly lucrative void?
  59. J2EE vs .NET[ Go to top ]

    .Net applications exist actually (june 2001) , yes there are still in beta version but they exit.
    You will see on September the power of .Net and the java killer : C#


  60. J2EE vs .NET[ Go to top ]

    Hi
    it is not just about speed of development. It is about maintainability before everything. You never stay with small application. It grows, grows, grows, and then? What then? It is a mess.

    For that reason we have chosen J2EE. It is not about Internet time anymore - forget about it. Internet development shall be considered as complex distributed OO project.

  61. J2EE vs .NET[ Go to top ]

    Working on designing a new IT architecture I am considering using an application server as a major component.
    After considering the technical aspects of J2EE and .NET I considered the application server market. On over 50 Application Servers, only 1 used the Microsoft Technology and over 10 the J2EE technology. Moreover at least 40 out of 50 used Java Like technology.
    Another interesting aspect on Java Like technology is that it has triggered OSS projects that are becoming more and more interesting. The features proposed are almost equevalent to those proposed by propriatary solutions such as websphere or weblogic. Even if they are not quite as effecient today, there are good reasons to believe that they will become technically as good !
    Without any expert skills on these technology I must admet that it will be difficult for me to defend the Microsoft Position.
    Looking closer at J2EE I am now looking for information on the development tools that are available on the market. I found benchmarks on JBuilder, JDevelopper, Visual Age, Kawa, Forte for Java, CodeWarrior for Java, PowerJ and Visual Café. If you have any information on other products I would be quite interesting on having details on their features.
    Sorry for the english mistakes

    JeanLuc
  62. J2EE vs .NET[ Go to top ]

    I think there is no difference between .NET CLI and the JVM as regards supporting additional languages.

    There are very few add on languages for .NET (Perl Python and Cobol), wheras JVM versions of just about every language (except VB and Foxpro) are under the JVM.

    The real difference lies in the GUI builder tools perhaps?

    I have never properly understood what is and what is not .NET, multiple language support must surely relate to Visual Studio .NET, not an API?
  63. J2EE vs .NET[ Go to top ]

    I think the discussion on this issue has been very balanced and it is refreshing for a change. However, I think both .NET and J2EE are architecture-first approaches to business
    application development and this tends makes the development costs even for both J2EE and .NET for similar project sizes.
    Therefore, in the short run the key determining factor in picking one over the other would be legacy integration, availability of strong development staff and the learning curves for skill set and tool upgrades. Going by past history of Visual Studio and Microsoft's emphasis on tool ease of use and knowledge base support, it appears that Microsoft may have an edge in the short term in this area.
    In the long run if J2EE tool development vendors catch up,
    this factor may not be as important.
  64. J2EE vs .NET[ Go to top ]

    It really depends on your timeframe for implementation of the chosen architecture, and your chosen deployment platform. At the moment, J2EE is maturing and as such its uptake at the moment is huge; it is also (to an extent) portable across platforms. .NET is all vapourware and marketing hype at the moment, and your deployment platform is limited to Windows. An easy decision, I would have thought ...
  65. J2EE vs .NET[ Go to top ]

    .Net will definately allow you to get to market much sooner with a smaller learning curge (for much less $) than J2EE. M$ has always had the best integrated, most productive IDEs. Creating an application for MTS is as simple as dragging and dropping a DLL onto a window (.Net will be even simpler). BEA Weblogic 6.0, on the other hand, makes you edit XML files by hand and build JAR/EAR files using command line utilities. While this is great once you have got it all working, it really slows down your initial development effort. Its hard to believe that a $16K/processor piece of software is so "user-angry".

    That being said M$ products tend to make simple projects easier and difficult projects more difficult. J2EE makes complex applications much easier to implement. M$ makes the first 90% easy but the last 10% can be incredibly difficult. This is mostly because of their closed nature. .Net is more "open" than COM but you still can't look at the source code! I don't know how may times I got to a problem with a M$ product and hit a brick wall.

    .Net is a big step forward. It's simpler than J2EE but I prefer J2EE for large complex projects. I just wish the App Server market would mature more quicly and give a user friendly product that has the high availability/scalability of Weblogic.
  66. We need better UI

    I don't know why companies like Oracle, BEA, iPlanet can't make a good UI tool to manage and deploy Apps in their App Servers.

    It's weird but I have been using J2EE RI and it has a great UI, the best that I have seen to work with App Server, maybe javasoft could make it an open deploy tool source code to be followed by the others App Server in the market.

    I prefer to use an application to do my pointless work (creating the XML deployment descriptor, packing jar, Mapping fields, etc) and work more in the code.

    Regards.
  67. Well, try Sybase PowerJ for J2EE development. IMO, it has the coolest tools for developing and deploying EJBs, as opposed to handwritten XML DDs, et.al of BEA Weblogic.
  68. J2EE vs .NET[ Go to top ]


    People seem to imply that .NET brings down the time-to-market.

    Which "market" are we referring to? Cause the software consumer market (both for internal applications in enterprise and for packaged products) has past the phase where it could afford getting bogged down by limitations over a particular "thing" including an operating system, transmission protocol, hardware etc.

    And although MS claims that it will support "anything", "over anything" through .NET, how does it intend to do that without letting vendors and customers have custom hooks into its core?

    The "only" reason J2EE is successful is because it is not a being architected or built by a "monolithic" monarch. It is "mostly" a consistent, methodical and cumulative effort of everyone in the industry, from a novice programmer to the giants like IBM and Motorola. And this will be the reason which makes sure that J2EE stays that way.

    MS has the advantage of being the biggest dog in the pound. But it has been and is throwing that away by locking itself down. Even the most loyal MS shops are giving more than a cynical look at .NET as they are not willing to lose the J2EE advantage of "open"ness.

    Well.. I dont think these discussions will stop until a very long time but I sincerely appreciate all the cool heads that have participated in this thread. :-)

    KR
  69. J2EE vs .NET[ Go to top ]

    I just had to add this.

    I heard a top executive of Motorola say this about .NET and J2EE:

    "[Microsoft] not making .NET open is like building a beautiful house but asking you to stay in the backyard and not letting you in. When you are already living 'IN' a very nice house, what is your incentive to shift?"

    And everyone here already knows about Bill Joy's statement regarding SOAP.

    So, MS has a tough battle ahead if it intends on pushing .NET really hard.
  70. J2EE vs .NET[ Go to top ]

    Right. Microsoft is short of concentration and resources.
    They have too many investment and too much trouble issues
    to deal with, like the judgement about its monopoly,
    Xbox, Windows XP, and .NET, games (like AOE) ....
    many many.

    It make itself generalized in
    the software field after its booming from Windows 95 HIT, but doesn't concentrate on the enterprise e-business market like now Sun J2EE provided.

    The software market can be an IMAGINARY market, what you think can be archieved. That means it's hard for just
    ONE company to keep to be HEAD in ALL of the fields
    about software. The booming of internet make this fact
    ome clear and clear.

    Microsoft can make a lot of money from it,
    but it's hard for it to rule the whole world anymore like what it did in the PC age. Since now a standalone PC is ... just a very tiny and limited resource.


    The boundaries are just resolved and resolved, and the world
    is connecting up and up .... it's not just issues about IT.

    It's also about politics and economy, and human.

    The globalization ....
  71. J2EE vs .NET[ Go to top ]

    Never underestimate the power of a powerful marketing team and a bunch of lawyers.

    This is not just a technical debate, it's a major play for a very lucrative middleware market -- and a major offensive against companies like IBM and Sun.

    Dontuse.net
  72. J2EE vs .NET[ Go to top ]

    Another thing I'd want to mention about .NET: It again clearly shows you cannot rely on Microsoft and how bad vendor lock-in can get.
    Because all Visual C++ and Visual Basic developers are now at a dead end (both will have to learn new languages, although C++ developers may have an easier time) which are the vast majority of people developing for Microsoft platforms.

    Messi
  73. J2EE vs .NET[ Go to top ]

    Everyone has his own biases.

    But please consider what you are saying. MS is introducing new technology, new architecture, new language syntax, all rolled into .NET. Somehow this illustrates the horrors of vendor lock, and indicates that developers cannot count on M$. But didn't IBM do the same thing 2 yrs ago when they introduced their product called WebSphere? Didn't Spyder/Kiva/Netscape/iPlanet do the same thing several times over the course of 2-3 years? Didn't BEA do the same thing to their Tux customers when they bought WebLogic? While Microsoft may be forcing their developers to learn new tricks, M$ has no "monopoly" :) on _that_.
  74. J2EE vs .NET[ Go to top ]

    Yes, that is exactly what I'm saying: BEA, IBM, all others do the same thing. Apparently a vendor will always try to "Lock you in"... they don't want you to move to another vendor, although some put more and some less emphasis on this.
    But what I mean is: YOu should never lock into one vendor, and that is exactly possible with Java. While there is J2EE now and to some degree we must admit that Sun wants to see "normal" Java Enterprise Apps as "legacy code" the J2SE is still developed and maintained, and you can still use your Enterprise Apps. Even better, you can do a slow migration.
    This will probably be the same with MS, but do you know? They may cancel "traditional" VC++ soon (and it seems they will do this, abandon the "traditional variants" of VC++ and VB, and now try to migrate an old VB program to .NET... will be very funny I guess), and no one will hinder them. In contrast, as long as there is need, you'll always get a JVM I guess, if not from Sun, from IBM or any other vendor. Will you get a .NET implementation of someone else but MS?
    So while iPlanet, Netscape and others may do the same thing you are always free to switch vendors, as long as you didn't let you "lock in".
    The situation shows that things might come you'd never expect. WHo thought that MS would abandon MFC, VB, Win32 API and all those things so soon? After they used it for years although much better technlologies/methodologies were available?

    Messi
  75. J2EE vs .NET[ Go to top ]

    I agree - everyone wants to paint M$ as the bad guy from the standpoint of seeking a monopoly, but Sun is seeking the exact same thing with Java - a key foothold that cannot be eroded. I find this aspect of discussing the difference to bring out biases more than logic. The fact is that every one of these major companies wants to own the market -that's the business goal. I personally would not want to see M$ just lose out to Sun and have Sun be doing the same things that M$ has done. What I would like to see Sun do is open-source Java or spin it off...but hey, that's just me.

    After three months of playing around with J2EE, I've now started playing around with .Net...and I have to say, I thought someone earlier had the best breakdown: MS has the tools, J2EE has the scalability and better tools for the large projects...

    Now, what's it going to take to get a review of Silverstream on here? Seems like a great product, but hey, I'm not buying the hype right now....btw, Orion is a great product!

    Seeya!

    Damian
  76. J2EE vs .NET[ Go to top ]

    I don't agree that the M$ and $UN are equal evils at least from a developers perspective.To me key issue is that in the final analysis MS makes its $ from software Sun makes it moolah from hardware. Eventually when you apply Darwinian logic to it MS can only grow(when you consider its market cap) by taking marketshare from other people in the software market place by pushing as many commoditized solutions into the market(including niche but highly profitable markets like telecoms, embedded systems, games etc). As long as they own the platform there is nothing people can do about it. I think every one on this forum knows the history(and the bodies :-)). The ISVs already know and understand this(I work for one), hence their enthusiasm for J2EE. Personally speaking as a individual developer I know I'm at the bottom of the food chain but eventually MS is going come for me and charge me gatekeeper fees. I know this is paranoid but I'm great fan of Andy Grove's dictum that "Only the paranoid survive." So I'd rather play with my enemy's enemy right now.

    Other than that I agree with most of the forum that J2EE is really great architecure with poor GUI development(esp. integration) tools.

    MS really kicks ass when it comes to providing an out of the box integrated development environment.
  77. J2EE vs .NET[ Go to top ]

    I am very surprised to see that so skilled guys as you continue to confuse about who is who in this computer and IT business. Is M$ more evil than $un or Oracle$ or Ibm$ ? That is a lost-time debate... But I want to share my today conviction:
    1- All are Devils until they realize you have understood their real intend !!! (to sell shit at the highest price). And non one is better in that game.
    2- M$ have been ejected from the JAVA wagon by Sun/IBM/Oracle gang using the doubtful strategy, and the fact that M$ was at that time looked as an Evil for many middleware developers. But If you today developers are honest, you will agree that the Java solution, which started War against M$, is no more than a COM+ solution with a better packaging and a OO adapted langage for integration and worked to resolve the weakness of VB/C++ link to COM+. M$ tried to introduce some rules in java, to take his Windows/COM/MTS investment into account, but the java gang refused pushing him into licence violation and creating all necessary lawsuit. They won to put him out but they also won another real concurrent platform .NET as microsoft intend to reply by the same way.

    3- I have not already look at the .NET solution, because of lake of time, but I am barely sure the M$ approach will be to do the same thing the java gang have done in the past: Look at the weakness of the J2EE solution and trying to leverage his current investment in COM+/MTS.

    4- Some one ask why M$ don't simply licence $un Java ? He have to remember that the goal is to reuse his past investment(COM+/MTS/DNA) ... And those who want to believe that M$ is creating a new product have to quit right now. The new challenge is how to package the same old stuff in a new shining manner. I know they use developers to advocate that their solutions are new.

    5- Java by itsel is not technically new ... But the rules, policies and environment are newly specified to give a more open shared opportunities and solutions. With C#, Microsoft will try to copy all good stuff of Java and give the opportunity to his loyal developers to not regret Java environment... And I will not be surprise that the 2 languages will share the same concepts each copying the best of the other.

    6- For my conclusion, I think the issue is held by the developers ... If Java Developers continue to be more vigilant as in the past, the J2EE platform will enhance and acquire more tools and be more easy to use. J2EE will continue to keep the distance as long as developers continue to share knowledge and there are more shareware source for java solutions. In the J2EE environement, it is difficult to cheat ... You can look at the change of course at Oracle: They have losted tme trying to sell anything but Application Server solution in Java environment. Today the have to admit they were wrong trying to call a cat a dog. They have good tools, but they too want to sell old investment products. IBM too is not so famous and they are not realy J2EE compliant ... By doing so, they give opportunity to Bea to take the lead. You can realize tha also Sun have some difficulties to be the leader in the aera. I think they also have internal fight with those who want to sell old stuffs.

    7- M$ can win if they continue to pretend in that sense. Developers in M$ area are somewhat confused by what is really going on, and they usually compete with M$ as main developer. Dear Developers, do'nt worry with M$ .NET solution, because it will be a one supplier solution, for lowest integrator. I think Java gang will not give him the chance to come back and your J2EE succefull solutions are the only way to measure that come back.

    cheers.

    The only issue for .NET solution is

  78. J2EE vs .NET[ Go to top ]

    What are the development tools you are using and what are the problems you have with them ?

    Have you got any to recommend ?

    regards

    JeanLuc
  79. J2EE vs .NET[ Go to top ]

    I havn't seen anyone point out that .NET is (will be) a product whereas J2EE is an architecture. I think that this is an importand distinction. In a sense it's like comparing apples and oranges. There are many vendors producing products based on the J2EE architecture, however, .NET is the only product based on it's architecture.
  80. J2EE vs .NET[ Go to top ]

    .NET is a product that contains many of the features of the J2EE architecture:

    Distributed computing model
    Componant development
    Transaction/Message based infrastructure
    Web and/or Fat Client deployment choices
    etc, etc,

    That orange sure looks and smells like an apple!
  81. J2EE vs .NET[ Go to top ]

    That Orange and Apple example was a very good one from the developer's point of view.

    One has to assume .NET as the orange and J2EE (or Sun ONE?) as the apple. The apple (not connected to Apple computers :) is a lot easier to eat without extra tools you must buy (a knife,...) Especially if a problem arises - e.g. it contains a bug or worm - you can get a lot faster to its core. The orange is covered and protected by a strong skin and without any extra tool one can usually not get into it.

    That seems to me like a great metaphor for the Open Source approach of Java and some J2EE components (Apache, etc.) vs. the propriatory (even with XML, SOAP, etc.) technologies of M$

    Of course if you are used to orange juice (e.g because it's advertized better than the apples) and you can afford a big juice squeezing machine one might still be doing better with
    the "orange".
  82. J2EE vs .NET[ Go to top ]

    Good point there. But remember the core of .net is CLI and SOAP/webservice, both of which have been submitted to the standard body. So it is possible .net can be platform independent just like java. CLI is comparable to JVM in terms of platform dependency. I think MS is moving in the right direction. In my opinion, .net and j2ee based technologies are very similar. But .net has been designed with more internet-centric (xml, soap etc.) than j2ee, although j2ee has all basic ingredients. my 2 cents.
  83. J2EE vs .NET[ Go to top ]

    NET is a product that contains many of the features of the J2EE architecture:


    True. But it is only one product. There will be one platform for .Net. There will be one implimentation of .Net.

    I haven't been doing J2EE for more than six months and I've already used three different implimentations of J2EE. Four if you could the WebSphere test environment that comes with VAJ.

    The J2EE implimentation I'm using currently (Orion) runs on just about any platform with a JVM.

    Having been free of vendor lockin for the last four years, I'd rather fight than switch. :-)
  84. J2EE vs .NET[ Go to top ]

    Is this a fair consideration?

        .NET is probably ".NoWhere" if the U.S. Justice
        Dept. prevails in its anti-trust trial

    I think so! A breakup of Microsoft, as unlikely as it may seem, would undoubtedly have far reaching consequences for .NET.
  85. J2EE vs .NET[ Go to top ]

    Hi

    As I haveing the exp. in C/C++ and after that I moved to the java technologies and have experinece the maturity of technologies vs evolution of tech. , java is now going to be stable according to the bussiness need of today and MS is still in the process of development of new tech. , all the accessary is available for the sun , as for MS is for vb and COM. anway java tech is better then new one ??


    deo
  86. J2EE vs .NET[ Go to top ]

    I would rather use J2EE and stay away from .NET. I wonder what happened to former VJ++ developers.

    With J2EE you have the option of deploying your application on linux/Solaris.
  87. J2EE vs .NET[ Go to top ]

    Abhi

    It has been over 2 months worth of "discussions" regarding the questions you posed.

    I for one am curious if you and your company have come closer to a decision or reached a decision out-right?

    If you have made one, what sort of pros and cons counted most? If you haven't yet made a decision -- are you leaning one way and why that way?

    TIA
    Roger
  88. J2EE vs .NET[ Go to top ]

    IMHO
    I would think that we would all benefit if M$ and $un, etc. would all work together on these new frontiers.
    All have great ideas to bring to the table.
    "We need to get to a state of harmonic agreement on such structures as they will become a very important part of our future society."
    Technologies are going into the making of our history and an influence on the Evolution of Mankind, MAN...... :-)

    But how do you bring all these these mind-sets together?
    Looks like only with the $, a symbol of our current state.
    How about a Project whose goal is to Globalise the whole Technology Sector. We have some good Foundations and Objectives to get such a project started :-) There are loads of minnie efforts.....

    BTW I would be interested in getting someones opinion, (insiders would be great)
    How will Micro$oft operate when Bill Gates kicks the bucket, Crashes?
    What will life be without this Central Serving Computer Chip?
    When someone presses the big M$ reset buttonm will we wake up to a Brave New World?
  89. J2EE vs .NET[ Go to top ]

    There has been some research done for you already and infact its done by The Middleware company itself. Yesterday at JavaOne one of their guys handed me a white paper on J2EE Vs. .NET. You might want to get your hands on it and read it. It describes the pros and cons of both these technologies pretty well.
  90. J2EE vs .NET[ Go to top ]

    Here's another comparison (DotNet Vs J2EE) from the DotNet Side !!
    http://www.objectwatch.com/FinalJ2EEandDotNet.doc
    http://www.microsoft.com is going live with DotNet Beta 2.
    No Software would ever be Bug Free !! But Considering MS History in S/W Stability and Bugs, this one already (Beta 1) is a huge step forward in the Right Direction. Sure, by next year DotNet will speak for ItSelf !!

    Varadhg
  91. J2EE vs .NET[ Go to top ]

    Here's another comparison (DotNet Vs J2EE) from the DotNet Side !!

    >http://www.objectwatch.com/FinalJ2EEandDotNet.doc

    Why can't the 'Scalability' section of this document, also include the other possible 'real world' hardware/software selections when considering J2EE?

      -> WINTEL
      -> LINTEL

    I would consider this section of the document incomplete for the purposes of serious comparison.

    If there are glaring performance/scalability issues with J2EE on either of these platforms, Roger Sessions could have played that to the hilt...

    Are there any benchmarks or 'educated' guessworks comparing J2EE and MS/.NET that use the same hardware (and maybe even using NT/2000)?
  92. J2EE vs .NET[ Go to top ]

    The URL for the Middleware Company's analysis is at http://www.theserverside.com/resources/article.jsp?l=J2EE-vs-DOTNET

    The conclusion is the expected one, considering that Sun paid for the paper and The Middleware Company wrote it: they say that J2EE is better for building web services. But there is now a rebuttal from Microsoft, which says that the original paper is pretty inaccurate when it comes to .NET.


  93. J2EE vs .NET[ Go to top ]

    Now that Sun has released a COM bridge to access J2EE,
    (http://developer.java.sun.com/developer/earlyAccess/j2eecas/doc/index.html)
    Windows developers could write VB,C++, etc code that works with J2EE. This could be the key for J2EE succeeded in front of .NET.
  94. J2EE vs .NET[ Go to top ]

    You're always talking about development, but don't forget that an application needs to be deployed, managed and maintained !
    General statistics are saying you'll spend 20% of your time on development and 80% on maintenance. So, you need an environment with which you'll be able to work on you application for many time and when you'll find that something goes wrong in your deployed application you'll have to check what, where and what to correct. The 1st thing you'll see about "going wrong" is slower response times, after...

    So, my opinion on .NET vs J2EE is exactly the same than about Windows vs Unix : How can you administrate your environment ? In other words, how can you know when something is going wrong, how can you know where the things are going wrong and what can you do when something is going wrong ?

    With Microsoft environment : you don't have any information and you can't do anything. This is true with Windows as well as the different technologies envolved : MTS... And please don't answser that the systems are managing themselves automatically...
    The difference with J2EE (thanks to Sun) is they defined the inner and outer APIs of the containers but not HOW to implement these APIs and to administrate these containers. So the different editors can compete not only on performance but (mainly ?) on the management aspects. And I think it's better to have a slower (in theory) but completely manageable application server than the opposite ! And if you have problems with your editor, you can use the products of an other one. The same with the development environments : Java is still Java and your ".java" files can be passed from one environment to an other one. The most important point is to NEVER code something specific for an application server in your programs, you'll loose all the benefits of Java !

    With Microsoft, everything is coming from one editor. So even if you don't like, you'll continue with him !

    You can see Java and J2EE as a processor : You can choose the environment as you can choose your compiler : you'll always generated code for the physical or virtual machine. And how many of todays developers are knowing something about the instruction set of the Pentium 4 ? Not a lot even if you're programming every days ! With Java, it's the same : you'll learn the language, the APIs, your development environment. That's all, never think on the deployment when you're developing, so forget to code your application server specificities in your programs.

    The conclude : Java is an open world, you can choose what you want on top of this technology. The hardware layer is now the JVM / J2EE specifications.

    Daniel Jallais

    remark : excuse me if I made some mistakes, english is not my main language.
  95. J2EE vs .NET[ Go to top ]

    You're always talking about development, but don't forget that an application needs to be deployed, managed and maintained !
    General statistics are saying you'll spend 20% of your time on development and 80% on maintenance. So, you need an environment with which you'll be able to work on you application for many time and when you'll find that something goes wrong in your deployed application you'll have to check what, where and what to correct. The 1st thing you'll see about "going wrong" is slower response times, after...

    So, my opinion on .NET vs J2EE is exactly the same than about Windows vs Unix : How can you administrate your environment ? In other words, how can you know when something is going wrong, how can you know where the things are going wrong and what can you do when something is going wrong ?

    With Microsoft environment : you don't have any information and you can't do anything. This is true with Windows as well as the different technologies envolved : MTS... And please don't answser that the systems are managing themselves automatically...
    The difference with J2EE (thanks to Sun) is they defined the inner and outer APIs of the containers but not HOW to implement these APIs and to administrate these containers. So the different editors can compete not only on performance but (mainly ?) on the management aspects. And I think it's better to have a slower (in theory) but completely manageable application server than the opposite ! And if you have problems with your editor, you can use the products of an other one. The same with the development environments : Java is still Java and your ".java" files can be passed from one environment to an other one. The most important point is to NEVER code something specific for an application server in your programs, you'll loose all the benefits of Java !

    With Microsoft, everything is coming from one editor. So even if you don't like, you'll continue with him !

    You can see Java and J2EE as a processor : You can choose the environment as you can choose your compiler : you'll always generated code for the physical or virtual machine. And how many of todays developers are knowing something about the instruction set of the Pentium 4 ? Not a lot even if you're programming every days ! With Java, it's the same : you'll learn the language, the APIs, your development environment. That's all, never think on the deployment when you're developing, so forget to code your application server specificities in your programs.

    The conclude : Java is an open world, you can choose what you want on top of this technology. The hardware layer is now the JVM / J2EE specifications.

    Daniel Jallais

    remark : excuse me if I made some mistakes, english is not my main language.
  96. J2EE vs .NET[ Go to top ]

    I have been playing with the VS.NET Beta 2 version for a
    few weeks. I must say I am very impressed. When it releases later this year, .NET will be the best platform for developing "Web services", period. The level of XML and SOAP integration that Microsoft has achieved in the .NET framework is remarkable. You don't need to know any SOAP or XML to write a web service; that way you can concentrate on your application needs and not worry about the transport. I do not think J2EE will acheive this level of integration- may be some J2EE development environments will in the future.

    So, granted, .NET Tools are great for building Web services.
    Having said this, web services are still very much an unknown; it remains to be seen how widely they will be used. For building back-end applications for Web servers, it remains to be seen how well .NET will compare with J2EE.

    Anyone wanting to choose .NET vs. J2EE today will need to consider that .NET is still somewhat unproven today; although that may change in 6 months. Do not underestimate the costs associated with going with J2EE too- the training of developers and so on. SO from this point of view, companies more familiar with Microsoft will probably want to transition to .NET rather than learn a whole new development environment.
  97. J2EE vs .NET[ Go to top ]

    Hi Abhinav,

    Why not developing in both environments? In my opinion it actually takes less time to make two implementations than to do it only once in java.

    1) First decide your java architecture. For me I use Tomcat/Poolman/Velocity but you could also use Orion, jBoss or anything execept the "big" application servers.

    2) Split your project in small milestones. Develop each milstone in C# first with all superior tools, debugger etc. After there is a breeze to port to java.

    It sound crazy but try it! (As long as dotNet is not available in other plattform than windows..)

    Regards
    Rolf Tollerud
  98. J2EE vs .NET[ Go to top ]

    J2EE vs .NET
  99. J2EE vs .NET[ Go to top ]

    I read through 90% of the comments here... some good information, so religous rants... but that is to be expected when we have a MS vs. anybody else dicussion.

    I have been thinking about J2EE vs. .NET thing for a while, and have some general thoughts... would love some feedback.

    <li>I am just getting started with J2EE, and agree that it is not for simple projects... way overkill.
    <li>J2EE seems to have one advantage, if you design a project/application with it, it can "force" you to use at least some "good" design patterns, simply by its nature. Never had anyone talk patterns of any sort with MS.
    <li>MS has by far, the best IDEs out there.... they just throw so much effort at them.
    <Br><br>
    One last thought... IMHO, i sometimes thing that the only reason a MS vs. everyone else debate exists is for one reason, pride. I know that using a MS product, that I can click a few buttons and have a huge chunk of code created for me, makes me feel a little useless.... i mean, anyone can click the mouse a few times (obviously it is not quite that simple). But using something that requires configuration and tweaking and testing and more configuration and investigation, when you get that built and running, I know I have a sense of pride that I was able to figure it all out and get it running... and that anyone off the street couldn't do it. I just remember that when i was in college using SQL 6.5, and finally getting it to work and configured, and then SQL 7 came out with all these wizards.... i felt betrayed. Strange... but just a random thought.
  100. J2EE vs .NET[ Go to top ]

    re: your final thought... I think you miss the major point.

    It is not an issue of ease of use or point-and-click vs. hand editting XML descriptor files, etc., it boils down to a proprietary vs. an open standards debate.

    Microsoft wants to hold most, if not all the "keys to the kingdom" -- without sharing them (or sharing sparingly).

    Microsoft makes damn good products and developement tools! I have spent much of my professional career in their sandbox.

    But I am now drawn to Java/J2EE. I find the J2EE architecture brilliant and the standards process (ie, product evolution) refreshingly open.
  101. J2EE vs .NET[ Go to top ]

    Open standards are OK gurus, but are we really in need of them ? i mean to ask.. if its that there are so many heterogenous and OPEN enviroments out there, then why and how is it that MS products are still in domination? If open standards are the way for the future... why dont people switch to them ? why do we only see people who are doing things-from-the-scratch going for Open standards based technologies ?

    Please correct me if i am wrong.

    thanks

  102. J2EE vs .NET[ Go to top ]

    Any thoughts ? Roger ?
  103. J2EE vs .NET[ Go to top ]

    Microsoft's dominance didn't happen overnight. They dilligently and methodically pursued their course during the nascient PC years when other companies were to lethargic or distracted -- much to their credit.

    Why don't we see more people going for open standards -- as you suggest?

    Linux has and continues to grow in popularity. eBay recently chose IBM's WebSphere to power its next generation offering. According to NetCraft (www.netcraft.com/survey) Appache enjoys a 2 to 1 margin of usage over Microsoft as far as webserver deployments. Based on the strong participation at this site and others I visit, there seems to be very strong interest and use of open-standards based products.

    Cheers!
  104. J2EE vs .NET[ Go to top ]

    hey roger!!

    i saw you are very optimistic about open-standards based technologies. i have been working with J2EE for a while now only to find that its a JAVA-ONLY solution. microsoft now claims that dotnet is not a single language solution unlike java and that would probably fetch them good market share, once again in the application framework segment. remarks ? comments?

    what do you think about the much talked about c# language, i understand its mostly like java. so, maybe the open-source community guys need to work a little bit more in the areas where microsoft is strong. for instance, microsoft has some very good development tools like visual studio, which is full of great features and at the same time its FAST. and microsoft has always talked about language-independence, starting all the way with COM, and now with .NET
     
    donno how many applications have really found the language-indep. feature useful. most of the COM/COM+ dev. is done with visual studio though.

    hope i did not drift away from what i actually wanted to ask and convey.

    Regards


  105. J2EE vs .NET[ Go to top ]

    i have been working with J2EE for a while now

    >>only to find that its a JAVA-ONLY solution

    IMHO ... I do not have a problem with Java-only. But that is me. ;)

    And the fact that C# is so Java-like ... well, copying is the best form of flattery.

    >>microsoft now claims that dotnet is not a single language
    >>solution unlike java and that would probably fetch them
    >>good market share, once again in the application
    >>framework segment

    Maybe this is true, but developers, architects and planners will consider the "entire picture".

    But [I'm moving to stand on my box] this sort of debate boils down to choosing those things that fit within the constraints and budget of a project and the comfort zones of its developers. Most often there is no right or wrong just a best fit. For some that best fit is open source or some mixture, for others it is Microsoft. [I'm off my box.]

    I'm not a hardcore zealot of either... just a recent user of Java/J2EE. And I'm enjoying it! I haven't been this happy since I stopped diddling with COM! ;}

    Cheers

  106. J2EE vs .NET[ Go to top ]

    J2EE seems to have one advantage, if you design a >project/application with it, it can "force" you to use at >least some "good" design patterns, simply by its nature. >Never had anyone talk patterns of any sort with MS.


    Isn't this to do with Java rather than J2EE.
    Or, more precisely, Java programmers, the Java community and what I would call 'the Java style of programming'.
    Small classes which do one thing well, designed to be re-used from the start.

    Java started after design patterns became popular, and most Java programmers tend to design with OO re-use in mind.
    C++ inherited a lot from C, and started before design patterns were popular.

    I've see old style C programmers mangle Java, and Java programmers develop good OO designs in C++.

    I would find it dificult to write a non-OO design in J2EE, but I like J2EE _because_ I like Java style OO design. It would be dificult to write in without OO design patterns because I naturally think in Java OO style.

  107. J2EE vs .NET[ Go to top ]

    In J2EE Design Patterns apply to architecture (application partitioning in particular) more than to coding specific objects. There are a number of ways to skin a cat and the patterns illustrate this in detail.
  108. J2EE vs .NET[ Go to top ]

    Java is language-specific and platform-independent, and dot-NET is language-independent and platform-specific."


    http://www.sdmagazine.com/documents/s=733/sdm0103a/0103a.htm
  109. J2EE vs .NET[ Go to top ]

    Java is language-specific and platform-independent, and dot-NET is language-independent and platform-specific.


    So if a company wanna develop something and have quite long
    long sight and visions, they wanna maintain JUST 1 version
    for their applications/systems and can deploy in ALMOST ALL
    of this world, any platform -- MAC, linux, unix, novell, windows, VMS, BSD .... just to maintain and upgrade ONE version, with a little bit speed sacrificying ....

    then bet on JAVA.

    How do you think ??
  110. J2EE vs .NET[ Go to top ]

    I have watched this battle for a long time. I have also re-read all these posts about J2EE vs Microsoft.

    Here is my take:

    To the people who think that a J2EE server must come from BEA or IBM and cost 16k per processor. Orion webserver http://www.orionserver.com is a phenominal J2EE server. The cost is 1,500 per box (no matter how many processors) it is easy to configure and also has a free developers license. I loaded it up on my linux machine and had it serving pages in less than 30 minutes.

    I refuse to be locked into the Microsoft line. Several reason:
    1) I have seen them change things in their programming APIs for no aparent reason. No announcement of the change. No documentation on it.

    2) Security. Period. Each month this year another crippling vulnerability has been exploited in IIS and the windows line of server software. Yeah all software has bugs, but it seems like M$ has more than their share of exploits. In the medical insurance industry our clients privacy and medical information are protected by federal law. A break in to our systems and the release of patient information could topple our company.

    3) Platform dependence. Yeah it is a rare day indeed that a company changes platforms. But it can happen. I wouldn't want to throw away years of development costs because our company switched from win2k to solaris, linus or freeBSD.
    Arguments are made that in a clustered environment Win2k is reliable. However I have a linux server I paid 700 for the pieces. It has a mail server, J2ee server, 2 databases, a web server, an FTP server and more. On this architecture I can expect uptimes of well over 9 months to a year and the total cost for all the above pieces of software, fully licensed, and the hardware is 2200$. I can't even get MS exchange (for mail) and win2k on a server for that. Much less expect uptimes of longer that 30 days. (before I get incinerated by flames, my authority for this is www.netcraft.org find the uptime for www.microsoft.com (they should be best able to maintain their own servers) and for www.ford.com (another large shop on the MS framework)

    4) In the early 80's when IBM came out with the PC they also were the main supplier of the operating system DOS. A copy sold for 400$. The whole reason MS even came about is they pieced together a version of DOS and sold it for 40$. IBM was on top of the software market and was extremely arrogant. The eventually fell and by their own admission, almost lost the company in the middle to late 80's.

    Microsoft is exibiting the same sort of behavior. The very act of them making our department spend 10's of thousands of unbudgeted money just to verify that we are being honest with our licensing of their product (guilty before proven innocent) is absurd. Bundling of products. Forbidding the removal of the IE icon from the desktop by OEM's. Attempting to make the licensing of XP like an ASP. They are exibiting signs of an arrogant company and IT departments won't stand for it. Similar to the early 80's there is another player in the OS market, linux and as soon as someone perfects WINE or provides a way to run ALL windows programs on a linux machine...The MS goose is cooked.

    5) Java was one of the first attempts to bring together architecture. Dissimilar operating systems could be developed for, allowing for niche markets to get software targeted for them regardless of the operating system. I see none of this in the MS plan.

    6) Java is one language and it is centrally controlled, although many can develop tools and parts to it.
    In the MS plan (.net) many languages are supported on one platform using a commmon binary. I don't know how the rest of the IT shops are run but I have enough of a time getting people in my shop to write code legible for the other members to follow. There are some parts of the code that are a pain in the MULE to maintain even though we have several senior people familiar with this task. This will OBVIOUSLY compound with the introduction of several languages. VB looks nothing line c# and c# looks nothing like ASP. I have a tough enough time finding qualified people in our area of expertise and getting them teched and their skills examined as it is. If I had to maintain a diverse staff and people (or materials) to quantify their skills before hire, just to maintain code, I would incurr dramatic additional costs.


    In closing,

    The very needs of our business, security, reliability, scalability and maintainability are not as equally represented in the MS (.NET) plan. It rather seems like the same old tale of upgrading to fix the OLD BUGS and to get NEW BUGS. I cast my lot with J2EE.

    dwilson
  111. J2EE vs .NET[ Go to top ]

    Don't underestimate the market power of the big ones. What is "best" is not necessarily what gets adopted. Did you ever use FORTE 4GL? J2EE is still years behind what FORTE 4GL is; but SUN decided to can it.
  112. J2EE vs .NET[ Go to top ]

    How big is your organization?

    From my point of view, in a medium to big-size company,
    the existing application landscape has to be kept in mind.

    And whether there are ressources and "knowledge" networks for Microsoft technology or J2EE.

    Don't try a mixture (like one technology for medium project / one technology for large project), it is hard to have capable staff for development and maintenance of more than one technology.

    What other pieces of standard technology are there and are they linked to either J2EE or .NET (e.g. databases?)

    Regards, Boris
  113. J2EE vs .NET[ Go to top ]

    1) Which architecture is more robust for applications which require extensive search mechanisms and documents sharing and transfer and collaboration ?
    J2EE and UNIX

    2) Which options will reduce time to market and cost to market or develop and cost to maintain applications ?
    The cost of a UNIX environment (unless you go the Linux route) will be higher, but in the long-run, which should be your concern if this is a critical application for your enterprise, will be less in maintenance as well as scalability is easier.
    We have been developing applications in Java for the last year, and we have seen a reduction in the development time, versus that of similar applications written in C, C++, or Visual Basic.
    Also, the advantage of developing in a Windows workstation, testing there to later move it into a UNIX environment without the need of recompiling has proven to be a incredible benefit.
    If you follow sound Project Management techiques, following the logical Software Phases (requirement gathering, analysis/design, etc.), you should be OK in any language that you choose, but development time can be cut by 15 to 20% using Java. We have experienced it ourselves.

    3) Which architecture will useful for future flexibility of moving towards an ASP ?
    Java, and the J2EE provide you with multiplatform capabilities that Windows cannot.

    4) Does size of IT shop and in house vs. out house development make a difference of what you choose ?
    Yes, if you are dealing with a small shop, and they already have a Windows environment, I would use that environment while still planning on a J2EE architecture as when the shop's needs/demands for more power become an issue, you can always move the code to a more robust UNIX environment.

    The beauty of J2EE is that you can actually have people outside of your organization work on pieces that can later be deployed, and if they need to, they can access beans/JSPs, etc., running on your app. server from a remote location. You don't need to deploy parts of the system to them so that they can test.

    We have worked with the Windows environment for many years, and when the demands became too much for Windows, we moved to a Sun-Solaris environment, and it has proven to be a very good decision.

    The problem you face with .NET and other MicroSoft architectures is that you are stuck with Windows as they don't intend to support other platforms, and if the time comes when you need to make the move, you will most probably be faced with extensive re-writes.

    I have nothing against Windows/MicroSoft, I believe there is a place for their systems, it all depends on the budget, the environment, and the technical needs. For a small shop, Windows may be the ideal solution, for them UNIX would be an overhead.

    Good Luck,

    Manny
  114. J2EE vs .NET[ Go to top ]

    Here's a question I have not seen addressed here yet...

    Many shops are dealing with an established M$ shop for web-side development, which was put in place to get on the web quickly. Now that the time and money are available to handle architectural projects like middle-tier development, you have to deal with your selection between J2EE and .NET in the context of having already developed a lot of M$-based web side code.

    Does anyone have feedback on integrating a highly-sophisticated mission-critical business tier with ASP/VB for the web side? Bear in mind that a primary objective is that, when done, your component-based middle-tier logic can be shared by your internal applications as well as your ASP code...

    I'm a fan of the massive uptime and power of the Unix-based systems, but like the ease of development in the M$ world, especially when I have a lot of people to train. These two issues always seem to be juxtaposed when making the J2EE vs. .NET decision!
  115. J2EE vs .NET[ Go to top ]

    Eric,

    I think Sun has something which runs ASP pages on iPlanet. This may be ok for a transition period, i.e. you "slowly" remove ASP code and put the business logic into your EJBs.
    Nonetheless, you will have to port/rewrite the majority of your code. I however think it is a good investment, as ASP pages on IIS are not really well suited for "mission critical" applications.
    But if you only intend sharing EJB code among your internal applications and ASP, this should bhe possible with the right AppServer (Borland for example) because it will give you CORBA access to your beans, and Borland VisiBroker (the CORBA ORB) even delivers a COM-CORBA bridge as far as I know.
    It will not be the easiest task however, guess in the long run it is cheaper adopting one platform (where I think J2EE is the better choice) and rewriting the code (to JSP, if deciding for J2EE).

    kind regards

    Messi
  116. J2EE vs .NET[ Go to top ]

    blabla
  117. J2EE vs .NET[ Go to top ]

    You said that development time in Java is shorter that in MS languages, how can you explain this?
    I can't imagine it's simply because you don't need to recompile to move your application to your Unix production server.
  118. J2EE vs .NET[ Go to top ]

    I'm just an average developer with experience in both tech's, but primarily MS due to business requirements. I complain tons about MS tech, but apart from being almost forced to use it, .NET landed pretty well with me.
    The talk about language interop is sales really, since in practice your code needs to be really clean for that to work flawlessly. Cross-language debugging is not in Visual Studio.NET (methinks).
    The real downsides are really the same as before .NET:
    1. Stuck with the win32 environment
    2. IIS
    3. MS proprietary rubble like the SqlClient classes which are only 1mm above binary.
  119. J2EE vs .NET[ Go to top ]

    We are looking for hard core J2EE developer with 2-5 yr experience.Salary is not a issue for right candidate.
    This urgent requirement for India(Mumbai) based company.

    Please send ur resume to

    varmaproject at yahoo dot co dot in
  120. J2EE vs .NET[ Go to top ]

    hi fellow programmers,
       I have just started working in Java. I find it cool!!! but now iam hearing about this .net thing.....i also attended a seminar on c#...but i found it very similar to java. What u guyz suggest for me? i should follow j2ee or .net... hope to hear from any one of u soon
    izzy
  121. J2EE vs .NET[ Go to top ]

    ALL of the above are suggestions.
  122. J2EE vs .NET[ Go to top ]

    Sorry abhinav
    but if you are from MIT do call me at 510-792-8069

    Thanks
    Ajay Jindal
  123. J2EE vs .NET - Innovation[ Go to top ]

    * Multi-vendor has led to a maze of tools/products that prevents me from finding a clear starting point like with C#/Visual Studio.

    * Microsoft is leading the way now in several area's: bringing multiple languages on one platform means everyone can choose his/her own.

    * Domain specific languages is a topic I do not find on J2EE discussion boards. This is really a revolution in bringing extra abstraction to the programming environment.
  124. JAVA vs. .NET[ Go to top ]

    I work with j2ee, j2me, .net and various other tools for client server and web (php, powerbuilder, etc) And after 12 years of being a developer, still bothers me the question "its tool A better than tool B"? (used word tool, but i can think a lot of most convenient words. use one instead) I think there is NO BETTER TOOL FOR ALL TYPES OF DEVELOPMENT. The right choice depends in all cases on need. Need to write a small-to-middle desktop application with data access in client-server form? USE .NET (any flavor, me=C#) Don't even try swing. Is pathetic. Need to write an enterprise-wide application, with thousands of distributed components, load balancing and grid clustering? Thing about J2EE, that is very mature in this field (this would be my answer to your question) Need to write a small-device application? USE J2ME. Need to write a dynamic site, with very few transactions? USE PHP. don't try to use a Application Server for that. Apache will do the trick. Obviously, you actually can write a full enterprise application in C++. See ya in 2017. Or you can use Java to develop a complex calculator in 3D. Waste of time. Please, stop asking what is better. A submarine can't be better than a plane. For the right job, there is no ONE TOOL. There are many. You can't marry to a specific vendor.