Discussions

News: .NET Dissed at Comdex, SD East 2002

  1. .NET Dissed at Comdex, SD East 2002 (92 messages)

    "J2EE has it all over .Net," said James Gosling at a panel yesterday at the SD East 2002 conference. "It's more feature-complete, more mature, and what .Net has going for it is a marketing budget only God can dream of. .Net is a product from a company. J2EE is a market." .NET platform support and the "embrace and extend" strategy were also debated in a panel at Comdex 2002.

    Read Father Of Java Takes Swipes At Microsoft's .Net (from SD East 2002), and Comdex: .Net remains point of contention (from Comdex).

    Threaded Messages (92)

  2. Once Java has the easy-to-use tools and fantastic out-of-the-box experience that Visual Studio.NET brings, I think the case for J2EE walking all over .NET will be cemented. Right now, the incredibly steep learning curve and rather substantial knowledge requirements for J2EE developers is what inhibits J2EE (and Java).

    Regards,
    Reuben.
  3. I always wondered, how do you become a MS programmer?

    I went to college (96-99) and studied Computer Science. The tools they tought me were UNIX (Sun and SGI) as the OS, Oracle as DB, ftp/telnet/rlogin for the Network (what's a computer browser?), rcs for version control.

    The programming languages used were Pascal, Fortran, C, C and C again, Assembler, Lisp, Ada, OpenGL, some C++ and some Java.

    Big surprise, people study and learn how to create solutions and really solve problem, they learn how to design a solution how to analyze it knowing strenghts and weaknesses.

    Java is hard only if you have never designed an algorithm. A true programmer should not care of the implementation language. Java is chosen because it makes life easier and faster.

    True that with IDE's and libraries you don't need to design an algorithm... it there already is the one you want. But what if you can't find it?
    I am in a project were one requirement is a fast RAM database that supports TX and is distributed among different machines. It is able to self-recover from crashes, and the TX must be distributed as well.
    Java gives me the fast Hashtable for holding data. But where is the 2-phase commit on the Hashtable?

    So I had to go back to basic algorithms, picked Reb-Black trees and build one in a couple of hours since I had pseudo-code for it
    (an IDE doesn't help here).
    Guess what, it still wasn't enough, TX were too big, I did research and found there are new algorithms doing almost what I want using a Relaxed Concurrent Chromatic Tree.

    Do you need to implement new data structures all the time, no.
    Is it nice if you can do all your work with a mouse, sure (I am waiting for an IDE who is a videogame, at least it would be fun).
    We are a long way from the moment that we will forget how to manually write code because IDE's are so much better.

    Should you be able to do both and choose the best solution for the problem? That is your first duty, that is why people pay you.

    Personally I like to use JBuilder, but it's a tool I use, it is my work.

    Marco Bonechi

    P.S. The fact that a lot of people learn and use MS technologies through bootleg copies of the OS and the tools is a shame for the entire IT industry.
    It damages the bottomline of us all.
    P.P.S. Out of ignorance, are there UML-to-.Net tools?
  4. .NET Dissed at Comdex, SD East 2002[ Go to top ]

    The Rational tool that plugs in to Eclipse is available for VS.Net.
  5. .NET Dissed at Comdex, SD East 2002[ Go to top ]

    The problem with most developers is that they can't see the forest from the trees. They focus so much on the technical issues that they forget that the whole purpose of IT is to serve their business client.

    How many projects have I seen where the requirements where very simplistic and developers would spend months building some system that is complete overkill.

    Don't get me wrong, I firmly believe that J2EE (minus entity beans) is the way to go for large scale enterprise systems. I have been developing in Java since 1996 and EJB since 1999. But I have finally come to the conclusion that for the majority of applications, EJB is completely unnecessary. This is where .Net comes in. It fills a niche for those systems that don't have allot of users and need to be developed rapidly using less experienced developers.

    Let's face it, most business clients don't give a hoot what language the application is running in. All they care about it that it fulfills the requirements, it's delivered on time and it performs well on low-end hardware. In most cases the application will never have thousands of users or be deployed across a cluster of enterprise UNIX systems. Low end systems is the market that Microsoft is targeting and they are doing a dam good job at it.

    .Net has been able to chip away at J2EE market's share in a very short time because of obvious shortcomings in J2EE. In particular the complexity of J2EE and the lack of rapid application tools. Sticking our heads in the sand and pretending these shortcomings don't exists will not help our cause. Until we stop pushing session beans and those god awful entity beans as the solution for every single distributed application, .Net is a technology that we will all have to learn to live with.

    Yvon Lavoie
  6. .NET Dissed at Comdex, SD East 2002[ Go to top ]

    "How many projects have I seen where the requirements where very simplistic and developers would spend months building some system that is complete overkill. " - Yvon

    User requirements very simplistic? Is that an oxymoron?
    Time is needed upfront to design applications that are not brittle, instead forward looking. It's called decomposing. I seriously doubt you have had solid J2EE experience, your succinct statement demonstrates this.

    availability, manageability, scalability, etc. shouldn't be sacrificed for drag and drop technology.
  7. .NET Dissed at Comdex, SD East 2002[ Go to top ]

    J2EE is part of Java. There is J2SE too. That with or without parts of J2EE is great for small projects.

    And I can do drag and drop programming in Java too.
  8. .NET Dissed at Comdex, SD East 2002[ Go to top ]

    "availability, manageability, scalability, etc. shouldn't be sacrificed for drag and drop technology. "

    You are missing the whole point of my argument. There's no question that J2EE is far superior than .Net. But for allot of applications, it's complete overkill. This is the market that Microsoft is targeting. Why drive a tank to work when a 4 cylinder car will do?

    BTW, I have more J2EE experience than you might think.

    Yvon
  9. Yvon,

    Did you know that it *is* possible to design a complex J2EE application without EJBs? Design overkill it not the fault of Java, but the fault of the designers.

    BTW - what evidence is there that proves ".Net has been able to chip away at J2EE market's share in a very short time"?

    Ryan
  10. Ryan,

    Almost, every EJB book I have pushes the whole Session-Entity bean solution. Even experienced designers can fail prey to overdesign if they are not careful. It's not until you really do some extensive performance tests that you realize just how poorly a solution like entity beans perform. It's no wonder that Sun's Pet Store got blown away by .NET. You should read "Expert One on One Java J2EE Design and Development" by Rod Johnson. It will really open your eyes.

    Let's talk market share. I used to program in Smalltalk back in the mid 90's. When Java came on the scene in 1995-96, every computer book store was flooded with Java books. I was lucky to find 1 or 2 Smalltalk books. It didn't take long to figure out that I should be learning Java. Well guess what? Walk into any computer bookstore today and what do you see? Learn .Net, ASP.NET, VB.NET, C#.NET, design WebServices with .NET, etc... The number of .NET books at my local store outnumber Java 3-1. I don't need hard numbers to tell me that .NET is here to stay.

    Yvon
  11. .NET Dissed at Comdex, SD East 2002[ Go to top ]

    Yvon,

    I am sure that .NET is here to stay. Somehow though, I don't see the book count as a quality measure of product strength and/or market share. Anecdotally interesting though.

    It seems that both products are here to stay.

    Jason McKerr
    Northwest Alliance for Computational Science and Engineering
  12. Yvon,

    I know that .NET is here to stay. I never denied that is wasn't. It's kind of hard to stop the Microsoft juggernaut. I am also not saying the .NET is a bad platform.

    I just think that a few anecdotal stories about over-designed J2EE projects does not prove that J2EE is two complex or that EJBs are without merit. It proves that those projects were poorly designed.

    Ryan

    P.S. On a side note, I keep hearing that MS tools lack refactoring. If this is indeed true, why is that?
  13. .NET Dissed at Comdex, SD East 2002[ Go to top ]

    <quote>
    Walk into any computer bookstore today and what do you see? Learn .Net, ASP.NET, VB.NET, C#.NET, design WebServices with .NET, etc... The number of .NET books at my local store outnumber Java 3-1. I don't need hard numbers to tell me that .NET is here to stay.
    </quote>
    Yvon,
    I use the same heuristic too to find out what thw trends are in the market. I too noticed the book shelves filled with .Net Stuff. What I also noticed was that more than 90% of the books were from Microsoft press. It's just a point to note.
    I am sure that the booksellers have a sweetheart deal to display those books. Nothing wrong with that and I am not about to go off on a rant about M$ is buying the market etc. They are free to do whatever they please.

    Cheers
    Ravi
  14. <
    <
    I agree. I am happy to know that .NET solves the above problems :)

    <
    You are contradicting yourself. First you quoted Entity Beans as a reason for failure??? in large scale enterprise applications. But when you defended .NET you are talking NOT A LOT OF USERS. For those NOT A LOT OF USERS you can use entity beans and write very little code, not to mention that you get transactions, concurrency, security and database independce.

    <
    I agree. But I also believe the clients need to care about other things like Initial Cost, Maintaince Cost, Security and Reliability. APACHE + TOMCAT + LINUX + JAVA may be a better bet in my opinion for small scale industries.

    <
    Figures please....
  15. .NET Dissed at Comdex, SD East 2002[ Go to top ]

    <Q>
    Once Java has the easy-to-use tools and fantastic out-of-the-box experience that Visual Studio.NET brings, I think the case for J2EE walking all over .NET will be cemented. Right now, the incredibly steep learning curve and rather substantial knowledge requirements for J2EE developers is what inhibits J2EE (and Java).
    </Q>


    Java does have very good out of the box tools. While they are not perfect and not everyone likes them, they do provide choice and, having used VS.Net, VS.Net is not perfect and is not liked by many.

    Also J2EE should not be compared to .Net. Java does. Personally, I don't think VS.Net compares to Java either.
  16. .NET Dissed at Comdex, SD East 2002[ Go to top ]

    Please, give me a break. The tools for java are quite primitive. The preformance of most of them is horrible.
    And the making the excuse "they are not perfect but...". They are not even good. I have been programming with java from the beginning, and recently I tried VS.net. The evironment is, in most cases better than eclips as well as jbuilder. Don't get me wrong, VS.net has its problems (the tool should GENERATE getters and setters, and the concpet of refactoring code is well, like missing. That's probably due to the fact there is no code around to refactor or MS programmers don't value code refactoring. I don't know, but anyway, Java programmers, be careful of the arrogance that is running wild in this site, or you may end up like OS/2 programmers ( in the OS/2 vs Windows debates ). Listen when people say this stuff is difficult to use. Learn from this, don't dismiss it. We want the best of all worlds for our tools. Tools that are POWERFUL, AND EASY to use.
  17. Yes some tools are primitive although if you look at some tools like IntelliJ IDE or Eclipse they sure are heading the right way. It's a nice thought to have powerful IDEs that do everything. Pbms start when these tools are used by bad programmers who think that they are J2EE experts because the IDE generated an EJB for them... but then again those people are spotted quickly in those post dot.com days :) where budgets are tight, aren't they?
    I use the Eclipse IDE or IntelliJ IDE or a basic editor in Linux (don't care). Ant is the heart of the project and can be trusted for repetitive tasks. I know exactly how my application is packaged. I am not dependent on any IDE. I can deploy this application on Linux, Windows & solaris, etc... At the end of the day, it gives the project total flexibility in terms of development environment, hardware, software and operating system. Priceless.
  18. .NET Dissed at Comdex, SD East 2002[ Go to top ]

    <Q>
    Please, give me a break. The tools for java are quite primitive. The preformance of most of them is horrible.
    And the making the excuse "they are not perfect but ... and recently I tried VS.net ...
    </Q>

    I'll give you a break because you've just recently tried VS.Net. I used VS(5/6) for years before doing Java and have used and am using VS.Net. WSAD is as good if not better. You should check out Carlos' (I think that is his name) viewlets.
  19. .NET Dissed at Comdex, SD East 2002[ Go to top ]

    <quote>
     Please, give me a break. The tools for java are quite primitive. The preformance of most of them is horrible.
    And the making the excuse "they are not perfect but...". They are not even good. I have been programming with java from the beginning, and recently I tried VS.net. The evironment is, in most cases better than eclips as well as jbuilder. Don't get me wrong, VS.net has its problems (the tool should GENERATE getters and setters, and the concpet of refactoring code is well, like missing.
    <quote>

    Then in what ways is it better?
  20. .NET Dissed at Comdex, SD East 2002[ Go to top ]

    <quote>
     Once Java has the easy-to-use tools and fantastic out-of-the-box experience that Visual Studio.NET brings, I think the case for J2EE walking all over .NET will be cemented. Right now, the incredibly steep learning curve and rather substantial knowledge requirements for J2EE developers is what inhibits J2EE (and Java).
    </quote>

    I think that's a feature, not a bug. Distributed systems are difficult to build correctly. Adding wizards so the clueless can build them just nets you bad systems.

    And anyone who thinks Java (as a language) is difficult to use needs to find another line of work.
  21. whereas everyone touts their futuristic products at COMDEX, Sun happens to be the only company that repeatedly harps on the tired-old beaten-to-death swipes at MS. Are they simply running out of anything productive to exhibit in trade conferences? James Gosling will probably make a career out of "dissing" .NET which only suggests he has nothing better to do at Sun.
  22. Its one thing to protect corporate intellectual property and rights, its another to create a profit center out of it.

    Stop the MS bashing and talk, and save Sun if that's what's meant to be. The industry through solid competition is rapidly creating a situation where Sun's costly proprietary platforms are losing to comodity product with better value propositions, its time to focus on business.

    Java is a great platform, .NET is more elegant and is designed with Java used as just one of the references in the biblio.

    These two platforms will compete. Java is IBM, .NET is MS and SUN is.......
  23. .NET Dissed at Comdex, SD East 2002[ Go to top ]

    Steven, you are so clueless.
  24. .NET Dissed at Comdex, SD East 2002[ Go to top ]

    I'm only going to say this once- anyone that I've ever worked with that complains about java(and usually they're VB guys trying to learn java)as being overally complex and that there aren't enough(visual studio like) "visual development tools" for java that basically allow them to hack code- have a common trait- they aren't good developers. I mean- maybe if you need a totally visual development environment and lots of help during the process-along with cookie cutter solutions that you don't really understand at a fundamental level- maybe you shouldn't be developing at all. As most likely- your app isn't going to scale,nor will it be easily modifiable, and its going to be full of unreadable, procedural like, hack code. But on second thought, its ok, as I'm always going to have a job, as I'm one of the guys that's gonna have to come in after and clean up the mess that these guys create with their MSDN subscriptions.
  25. The problem is that for many small jobs 'hack code' is fine. Scalability is a non issue for a VB app that is used for a small company of say 40 users. I have seen major banks work for years on so called crap VB software with never a complaint.

    I have seen project created in Java that took 3 months and a mod rewrite done in VB in 3 days! Yes blah blah they were crap Java coders but 3 months vs 3 days!! I would not want to use VB on big projects (actually would not want to use it at all) but I would guess that a huge part of the market can be found in small cheap apps. I think that MS cash would substantiate this.

    Stef
  26. I'm sorry, but I thought the context from which we are speaking is of enterprise apps- which is what .net and j2ee are touting themselves as ideal for- I really wasn't talking about use of vb on smaller projects- you're right- there are good reasons for using vb- smaller projects being one of them. I've used python on smaller apps as well- with good success- and its extremely fast to develop in- I didn't even consider java.
  27. .NET Dissed at Comdex, SD East 2002[ Go to top ]

    I do not know where you guys work or what programming skills you have. I have been an enterprise developer and consultant for over 10 years, and I do know that who ever says that a program developed in language A took 3 months and the same program written on language B took 3 days is FULL OF IT!!. I have developed in C/C++, VB, Delphi, and Java for big and small companies, and there was not one project I can honestly say that if it would have been coded in any other programming language it would have taken so much less time to complete. It is all about having the right people for the right job. Why would anyone take a Cobol programmer and have him or her code in C++? Same goes to VB programmers that are used to writing macro like snipped of code, if all of the sudden you have them do EJBs, you just selected the wrong programmers to successfully finish a project on budget. Never mind that you have also killed those programmers self esteem and set them in a course for failure. Tools will aid in the programming and design of a project, but tools will never teach or make a programmer code better. I know this for a fact!
  28. Sheesh! This superiority complex that many Java Developers has is quite nauseating! By the way, yes, I am a senior developer, and yes, I grew up on vi, notepad and then textpad.. but no, I'm not arrogant and stupid enough to lump all developers who like a decent IDE into the "retarded developer" category.

    You don't need to be a Java Developer to like patterns. I've developed great code in Java, and I've done it in C# as well. While I still love Java, I can't help but wonder how much more productive I would be if there were a Visual Studio for Java. Anyone who thinks that Java developers invented patterns have obviously not read a page in the vast collection of computer science literature of the past thirty years. But then again, what more can I expect from arrogant snobs who consider themselves superior for the most ridiculous reasons.

    I for one am not keen on showing off my programming prowess by writing the same inane code over and over, when an IDE would be just perfect for handling the mundane nonsense.. which by the way, "retarded developers" would be quite happy to do over and over again!

    What can I say -- Some people buy BMWs to compensate.. and some like typing in textpad and sneering at others. Ah well, 'tis the way of the world.
  29. It looks like you are too concerned about IDE. Please take a look at Eclipse. By the way it is Free.
  30. <Some people buy BMWs to compensate.. and some like typing in textpad and sneering at others. Ah well, 'tis the way of the world>

    That is exactly my point too. Many people cannot afford BMW and they are happy with Honda. It may not look as luxurious as BMW, but functionality and reliablity is as equal as BMW. By the way this comparision is only for IDE. On the backend Java is BMW that comes cheaper than BMW :)

    When it comes to MS even before any proven enterprise applications all of a sudden everybody is bashing Java. You are accusing Java Developers heavy headed. Tell me how is Java more complex than VC++. Did MS programmers complained like this to MS that VC++ is complex and MS responded with C#.

    People who are working in VB [When Java came they were around 40%] started complaing Java is too complex where as object oriented programmers embraced Java because of its ease of programming. In a week people were able to write a HTTP Server.
  31. Only rednecks drive BMW, at least in Germany.
  32. .NET Dissed at Comdex, SD East 2002[ Go to top ]

    While talking about Developers or software professionals, I have seen two types of people.

    One type, Who went to school, Got a graduated and trained in software technology and applied that in wide range of applications.

    Second type: Never been to school, may be a high school diploma and worked in projects where they mainly did drag drop colorful screen works.
     
       Most of the times both of these types try to establish themselves as software professionals. I don't think Microsoft is a company for enterprise applications or they do have a framework, which a client can apply for a growing business.
       You don't have to talk big enterprise applications with millions of transactions like Fidelity or citigroup to compare these technologies.
       Personally I have two middle level clients who really experienced the real idiotic side of Microsoft technologies and decided to throw that away.
       One is Universal studio's classical music division, second is Towers perrin. A person who have little common sense can understand who many classical CDs and musical items universal sell in a regular business days in there iclassics portal. They do have some business only during the holidays and special occasions. But universal do that business cuz of their prestige issues. BUT Microsoft technology fucked them very well during the 2000 Xmas holiday season. That is the shit Microsoft selling out.
    Later Universal kicked out MS sales man and deiced to move into Java and BEA Portal Technologies. One time investment, they got a scalable, enterprise, everlasting, understanding technology, which help them to run their business.
       Second case: Towers Perrin in Philadelphia, a Microsoft client in last 15 years. In their internal client applications, average load is 67 users, only have heavy load during the enrollment time. Microsoft technology again fucked them during the enrollment time. Finally they said goodbye to MS. Microsoft technologies never show their hidden problems until and unless you have to scale it or load it with real load. Soon or later a real technologist or businessman will realize that and kick them out.
     But guys who like dragging and dropping development, VB is very nice, colorful, easy, blahhh, blahhh.
  33. .NET Dissed at Comdex, SD East 2002[ Go to top ]

    I worked in VB and VC++ for 3 years. I've been working in java for 4 and recently have built a couple of c# apps even though our bread and butter project is J2EE.

    For those of you who still believe VS.NET is the killer IDE, it's not compared to others out there; specifically IBM's WSAD. No I'm not promoting it, but I'm extremely happy with its boat-load of editors, J2EE support, refactoring support, XML editing and transformation support, WebServices support, and so on.

    C#, I can take it or leave it. Although it is good for is building GUI's with JAVA/C like syntax using VB metaphors.
    It's still not as strongly typed as java. The collections API sucks. Form element data types are not consistent(MS Why does listbox.index return a double instead of an int?). Yada Yada Yada...

    MS will let you build the 80% solution within c# API. The remainder, as always, will need to be done with the Win32 API or C++.


    I'll use .NET if need be, I just don't prefer to do so until it matures quite a bit.
      
    Just my 2 cents..

    cheers
  34. <MS Why does listbox.index return a double instead of an int?>

    Because .NET is for enterprise application and MS expects a list box to be constructed out of an datawarehouse table which could have billions of rows in SQL Server :)

    One more reason could be Java returns int. We want to win Java, so let us return double :)

    NOTE FOR MANAGERS:)
    Eventhough I believe choosing .NET vs J2EE should not go to method level, I am not able to stop myself :) Sorry guys...
  35. Yeah, right..

    Well an int(32 bit) is 2+ billion.

    Hmm... Thanks MS, I want my users to wait for
    3+ billion records to load, run out of memory, reboot; and then when the list is populated, find the unique entry they want.. Yes, that is what I want instead of my programmers wondering WTF is up in Redmond....
    Although, my VB -> c# programmers will think it's cool to load that many records at one time within a control....

    Cheers....
  36. Jayakumar your ignorance argues against your point. (Assuming you ever had one.)

    Anyone with a very basic knowledge of computer science knows that numbers can be represented in a variety of ways.

    Two common formats are IEEE floating point (eg float and double) and your basic fixed point types (int long etc).

    Double is the wrong choice because it can suffer from a loss of precision. It is useful for protraying relative amounts over a large range. If a precise amount to the nearest unit is required fixed point arithmetic must be used. A double over a large range would not allow you to access all the records anyway!

    You do not understand basic computer science, so why should anyone listen to you? Let alone managers?
  37. anyone who puts a billion items in a list box should be shot
  38. .NET Dissed at Comdex, SD East 2002[ Go to top ]

    <quote> Form element data types are not consistent(MS Why does listbox.index return a double instead of an int?).
    </quote>
    There is not even a ListBox.Index property in ListBox control in Winform or Web Server control. What are you talking about?
    There is ListBox.SelectedIndex which returns int or
    ListBox.Items.IndexOf(Object value) which also returns int. I wonder where you get that information. Care to elaborate?

    Regards,
    Hun Boon Teo
  39. listbox.index[ Go to top ]

    Suez Zeus wrote:

    > (MS Why does listbox.index return a double
    > instead of an int?).

    Umm, what are you talking about? The ListBox class in .net does not have an index property. It _does_ have a SelectedIndex property, which returns an int.

    > MS will let you build the 80% solution within c# API.
    > The remainder, as always, will need to be done with
    > the Win32 API or C++.

    And Java will let do the same. After all, any project which uses JINI or CORBA or AWT is built on top of "native" code.

    Art
  40. listbox.index[ Go to top ]

    actually, this is grossly inaccurate. You can build a complete solution with c#. What do you think the entire asp.net framework was built with. guess what? C#. You can use whatever language you want to in the framework. In the .net framework, the laguage is really irrelavant and only serves as a means of a syntax checker.
  41. .NET Dissed at Comdex, SD East 2002[ Go to top ]

    I completely agree with you about the superiority of IBM WebSphere Studio over Visual Studio.NET. It's the real killer IDE, for those who like it. It has everything a real developer will ever need in a super simple and easy-to-use user interface. I especially like the concept of having different "perspectives" for the different tasks. Although my current project got me involved in Visual Studio.NET I have to admit that IBM kicks ass AGAIN. Sorry MS, you seem to put a lot of hours to make those nicely-shaped menus and "new-style" floating toolbars, but I wish you had spent this time on making the IDE more ROBUST.
  42. .NET Dissed at Comdex, SD East 2002[ Go to top ]

    That’s very funny; allow me one citation from another member of TSS:

    <Q>Is it ethical to let companies just blindly walk into the gas chambers of websphere?
    "Is it good for anyone or any company if they waste money on such an inefficient kludge as Websphere is ( 300 products and $11 spent on services for every dollar on product -per IBMs own study!)
    It is because of "products" like websphere that gives the whole industry a bad reputation and makes consulting a scam rather than a honorable profession.
    As far as ethics goes, "Intentional negligence" that leads to harm is no different than "Intent to harm"<Q/>

    The mystery goes on,
    Where are all the big J2EE projects?

    Regards
    ROlf Tollerud
  43. P.P.S. Out of ignorance, are there UML-to-.Net tools?


    I've just started using Visio for architects, which has UML, DDL etc tools in. Looks like they slow down for big models, but they are fine for drawing pretty pictures and generating some vanilla classes. Even used it to rip off a schema from Oracle and generate DDL for SQLServer.

    >Declared exceptions v's.....

    Good description of why they made their choices is at
    http://one.cs.washington.edu/csharp/response.html

    > IDE's religion
    They all play leap frog with each other. Every 6 months the comparisons change depending on who has the latest release. I prefer Intellij over visual studio, but I like knowing how to code rather than drag and dropping. It's hard to give up 'knowing', simply for 'doing'.

    Jonathan
  44. That´s the beauty of J2EE. It´s about choice. If you are not satisfied with an implementation/vendor, just change it!

    I wholeheartedly agree with this another member comments about Websphere (I've used it) and for that matter all other things coming from Big Blue (except Eclipse - it rocks!).

    Ok, I know it has been discussed thousand times before, but you Microsoft guys keeps using the same old (and failed) arguments! Don´t use factual (isolated) comments/events to generalize. Stop F.U.D.!

    To play your game, I could use the same arguments used by this another member about Microsoft and its products (I know, I've been there).

    Best regards,

    Andre
  45. .NET Dissed at Comdex, SD East 2002[ Go to top ]

    Microsoft does not have hidden problems. All problems are published on MSDN. It is just a matter of
    - running into the problem
    - finding out, if your are the problem or the Microsoftware
    - finding out, it is not you
    - Going thru extensive research on MSDN to find a workaround
    - Find either a workaround that leads to the next hidden problem published in another article (unlinked) or find out that there is no workaround (this is actually almost a relieve because your search is over and you have a MSKB to take to your boss who thinks you are an idiot because you are already a week over the project deadline)

    Just recently happened to me while working with the Outlook Object Model and CDO, Microsofts messaging "flagship".

    So my advise to be a successful MS developer: Don't believe what Microsoft is advertising - Lookup MSDN first, there always is an article ...

    Me bitter? Naahh
    -b.o.b.
  46. Try Intellij, for Java its quite good.
    You can find it at http://www.intellij.com

    I also come from a vi/unix background, and do appreciate IDE's that make my life easier but I dont need to rely on them to write maintainable code.

    I dont feel all Java developers are snobs but if you've ever had to recruit developers to join large development teams, even today I find it shocking how many people cant write a simple Hello World program without their favourite IDE.

    ( I remember the good old days when all we had was vi ( or axe ) and gcc or cxx and that was it! LOL )

    Anyway thats a criticism i belive is true of both Java and .NET developers to varying degrees, however unfortunatly in my own experience its far more prevelant amongst VB and .NET programmers where application coding is so dependant on the IDE.

    Anyway thats my 2 cents worth
  47. P.S. On a side note, I keep hearing that MS tools lack

    >refactoring. If this is indeed true, why is that?

    vs.Net doesn't have good refactoring. eg rename a class, and then you still have to go into the class to rename the references. I guess the reason was about scope :- they had it working enough to ship it so they did. Adding more complexity at that point was probably scarey. And decent refactoring in java has been about for 18 months (that I'm aware of).

    I'm starting to agree about the troll comments above. I must admit to mentioning the ms stuff alot myself here over the last 6 months. But I've got my mind arround enough of it now to find it less interesting. And I believe most comments are a genuine interest in comparing through discussion. Maybe we should have a seperate area of TSS for .net v's java. It IS a valid thing to talk about here, but it is currently swamping most discussions.

    The rest of tss can then be about the future of java. i.e. how it MUST evolve in terms of tools, and techs to slim down and gain power. To counter the last 2 years of bloat.

    The trends I want to see explored are 'to ejb or not ejb', 'IBM stealing the ground from Sun in java land', 'How closely tied is the future of j2ee to linux', 'can java complete for smaller projects hosted on ms platforms'.

    These are not really techie arguments, but will effect the future of java, and our investment in it - in terms of lines of code, and education.

    Jonathan
  48. Maybe we should have a seperate area of TSS for .net v's

    > java. It IS a valid thing to talk about here, but it is
    > currently swamping most discussions.
    [...]
    > The trends I want to see explored are 'to ejb or not
    > ejb', 'IBM stealing the ground from Sun in java
    > land', 'How closely tied is the future of j2ee to
    > linux', 'can java complete for smaller projects hosted
    > on ms platforms'.

    I wholeheartedly agree with those two points. And I would like to add the following point:

    It seems to me that the common point in TheServerSide´s public is *not* "Java/J2EE" or it´s advocacy. Our goal (well, at least mine :-) ) is to produce quality software systems using this platform. Period.

    To achieve this goal, comparsion *is* important. But not in the "us versus them" tone - instead, we should try to be more realistic, and ask ourselves "which are the good points in that other platform that are missing here ?", "is EJB/JNDI/checked exceptions/name your pain here making us less productive than those guys, or in the long run it pays off ?".

    I have a strong feeling that looking without prejudice at the other side of the fence is *much* better to increase the quality of our work than trying to feed the collective hallucination on the line "everything Java/UNIX is pefrect, everything MS sucks". Unfortunately, real life is a bit more complicated than that...
  49. I agree with the above statements. I played with .NET when in Rc1 and it gave me some ideas that I now use in my J2ee work. Now again I am playing with it (on the side) looking at how MS approached things and again some ideas are coming about.

    The way I look at it, developers should know both J2ee and .NET; makes you a better coder and a job may depend on it.

    Stef
  50. .NET Dissed at Comdex, SD East 2002[ Go to top ]

    IDE issue:

    I have been developing java using intelliJ for the last 6 months and Visual Studio is just horrible compared to it. Please give intelliJ a try before saying there is no good ide for java

    Major Sites using J2EE:

    In canada, almost ALL major banks run on Java

    www.tdcanadatrust.com
    www.cibc.com

    and many many more
  51. .NET Dissed at Comdex, SD East 2002[ Go to top ]

    I thought I said Intellij WAS good;)
  52. Aye, aye Matthew. I've been cleaning up crappy projects for the last three years.

    You make an excellent point. Java is not rocket science. And if VB developers don't understand Java they shouldn't even try to code with it.
  53. .NET is more elegant


    I am tired of seeing this.

    Yes it is newer. But more elegant?

    Is the lack of checked exceptions more elegant?
    Is the necesssity to declare "override" on the base class for polymorphism more elegent?
    Is the lack of integration with major messaging vendors more elegant?
    Is the fact that you have to write a server (for non-http) more elegant?

    Yes, there are some features in C# that I like too - but what makes you say it is more elegant?

    -Nick
  54. I am a Java programmer, but for the sake of truth:

    >> Is the lack of checked exceptions more elegant?

    Checked exceptions are nice feature of Java language, but are overused in the JDK libraries so that Java code is often bloated with something like:

    public void someMethod() throws YYYException {
      try {
        // ... some real code
      }
      catch (XXXException ex) {
        throw new YYYException(ex);
      }
    }

    >> Is the necesssity to declare "override" on the base class for polymorphism more elegent?

    Yes, it IS !!! Declaring method "override" when you realy intend to do so ensures that method you want to override really exists in the superclass and also prevents you to override some method that you do not want to. So, I think it is better language solution to have special keyword for declaring overrided methods.


    >>Is the lack of integration with major messaging vendors more elegant?

    You are right on this one, sure it is not more elegant.

    >> Is the fact that you have to write a server (for non-http) more elegant?

    This is false. NET has remoting, and you can also use COM+ through 'Enterprise Services'.

    Also, IMHO, although C# and Java are very simmilar languages with a lot of same/simmilar features, I think that C# is slightly better language than Java (it has stucts, enumeration types, properties, delegates and more, but lacks checked exceptions and non-static inner and anonymous classes).

    On the other hand, I found that Java as a platform (J2SE, J2EE and related technologies/libraries/APIs/communities/companies/...) is much more suitable for enterpise applications then .NET.

    Regards,
    Mileta
  55. I can see your confusion...You want to support C# as well as Java.

    1. Your arguments against checked exception is subjective. May be it will look like bloated at sometime. In 'Design by Contract' and B2B integration it makes more sense to respond to certain exceptions.

    2. Your argument towards 'override' is not object oriented. You should be aware that you can make a method 'final' in Java if you don't want to override. MS wanted a Java clone, but not an exact copy so they started reversing the idea [in their language innovation].
    To explain why it is not object oriented, let us say Car is the super classs and you have a run() method, Camry extends Car and can override run and declare it as final. Why should I explicity say in the base class as 'override' to run() method as that is the main purpose of having a base class and subclasses to create dynamic behaviour.

    3. Believe me MS had the option of studying Java and made a clone. So there could be something better and something worse. But those small choices does not make C# a good language when compared to Java. It will always be called Java Clone. Except MS if any other company had comeup with a C# language it would have died already. You know why because only MS has locked so many small to medium size companies to their platform.
  56. "I can see your confusion...You want to support C# as well as Java"

    Why do u think using both Java and C# as "confusion"???.. I ask this question as I'm one of those "confused" parties who use both at work. ;-) .
  57. <Jayakumar>

    1. Your arguments against checked exception is subjective. May be it will look like bloated at sometime. In 'Design by Contract' and B2B integration it makes more sense to respond to certain exceptions.

    </Jayakumar >

    Check this out : http://www.mindview.net/Etc/Discussions/CheckedExceptions
  58. <Jayakumar>
    Your argument towards 'override' is not object oriented. You should be aware that you can make a method 'final' in Java if you don't want to override
    </Jayakumar>

    You can do this in C# by using 'sealed' instead of 'final'.
    But this wasn't the point.

    Mileta's Cekovic point was that using 'override' you override explicitely, that is you can have a compile-time check on whether you really did override the base class.

    In Java if you mispell the method name in the derived class there ain't gonna be no warning at compile time.
    In C# you'll get a 'no suitable method found to override' compile-time error.

    P.S.
    By the way someone mentioned an index of a listbox as having the type double. Can you please give the fully qualified name of the class.
  59. For #1, I like Design by Contract. For that I commonly use Eiffel (which works in the .NET world although some things are missing but DBC is there).

    For #2, the major difference between C# and Java with regards to methods is this -

      Everything in Java is virtual by default.
      Everything in C# is not virtual by default.

    In your Car superclass example, for C# you would mark the method as virtual (not override). Override would be done in the subclass (Camry). The benefit of this (as someone else pointed out) is that if I do not mark "run" as virtual and someone tries to override it they get a compile time exception. This follows the C++ paradigm (to a degree). IMO this is better because designing a framework that can be overriden is hard.

    For #3, what modern language today isn't defined to some extent by features in another language? I like C# properties better than the get/set method of Java. I like C# indexers. I like the foreach statement (assuming you implement the interface IEnumerable).
  60. .NET Dissed at Comdex, SD East 2002[ Go to top ]

    "The benefit of this (as someone else pointed out) is that if I do not mark "run" as virtual and someone tries to override it they get a compile time exception."

    That should be compile time error. Sorry.
  61. .NET Dissed at Comdex, SD East 2002[ Go to top ]

    <quote>
    Yes, it IS !!! Declaring method "override" when you realy intend to do so ensures that method you want to override really exists in the superclass and also prevents you to override some method that you do not want to.
    </quote>

    You really do things like that? That sounds like a class of error that wouldn't make it past the first unit test run.
  62. <Nick>
    Is the lack of checked exceptions more elegant?
    </Nick>

    I still think checked exceptions are better. In IDEA whenver it encounters an unhandled exception while you're typing, a lightbulb appears and with an alt+enter the exceptions are catch()-ed. And printStackTrace is what you see in the catch block. This is like C# actually. Now if you want to handle an exception in a different way then just revisit it and do some custom stuff there. So it's not a big deal at all (the annoyance factor in Eckel's article). I think it encourages good programming style and btw later when you want to review your code and handle exceptions properly then it's a lot easier to find "catch" in source and review them.

    Something which is really a big deal is the fact that exceptions are a formal part of interfaces/virtual-methods too. So if you have an interface AnIntf and a method "f() throws AnException" then AnException is the only Exception you can throw from the implementing method of a class implementing that interface. Now what if you want to throw a ValidationException too? Well, if AnIntf is defined by a third party you can't add that exception to the signature, and btw it's not ok to do so, because then you should catch that exception anywhere else too! You just want to throw ValidationException for this specific overriden method and the caller (you probably) knows that it should catch it too. With C# it's possible, with Java it's not. I think neither Java nor C# came up with a good solution. Maybe a better solution is: checked exceptions, but allowing Exception to be always throws from methods of an interface or public/protected and non-final methods. Maybe....

    Ara.
  63. You know, I a little confused, If you have an intervace anItf that is supplied by a third party, this is a well known interface. This interface may be used by lots of other code. Now you want to CHANGE that interface (contract) by adding NEW exception. I believe that conceptually this is a problem. Exception are part of the contract. If you want to widen the contract, then its a different contract. AS you stated in your example. You wanted to widen the contract so that YOU could call it. In that case, you would implment multiple interfaces, and use a different method maybe ?
    ... Anyway if this is an example of why you don't want checked exception.... It is a pretty poor example. What about the other 99% of the time where checked exception help you write correct, predictable code ?
  64. .Net is not vaporware."

    Since i have never worked on big projects like Merril Lynch is there anything that compares in the J2EE market? (75 million transactions per day?? that seems very fast!) And what exactly would be a reason for .Net to be worse for enterprise apps than J2EE? cant the same rules/ideas be implemented in .NET? i cant see the language as a barrier. (Other than that you will have to stay with one platform) for all i care someone could make something like Enterprise C# Beans, "Jini" for C# or whatever, not that im waiting for that but anyway. Im a big J2EE fan myself i was just wondering.
  65. To All Managers....

    1. DESIGN and EXPERTISE: I have never seen anybody till now in MS world talk about GOF patterns. My friend who is just of school who is learning Java talks about Design patterns. Better programmers are on the Java side. ADV JAVA.

    2. STANDARDS and PORTABILITY: Forget about portability across different platforms, what about writing a dll that can run in all MS platforms. So All the concerned managers should start thinking that .NET is good, but do I have the assurance that Microsoft will make backward compatibility as an important issue. ADV JAVA

    3. CHOICE: As a person I want to go Lee and buy a Jean and I like to think about my freedom to select a shirt from some where else. Same applies to Software which is much more complex than buying a shirt. I want to run .NET server on linux because i like .NET the ease of programming and the stability of Linux on top of Oracle. Is there a way. Please don't tell me it is possible. I don't want to day dream. ADV JAVA
     
    4. HISTORY: C++ Vs VC++, JavaScript Vs JScript/VBScript, Java Vs C#. COM Vs CORBA. Except VC++ [only to some extent] every other counter technologies innovated??? by MS were aimed to solve the problem that were already solved. ADV JAVA

    5. PERFORMANCE: Recent JDK performance shows that Java is inching closer to C. And also I am aware during the C times I had 486 machines and for the same price I am getting a PIII with GHZ's of processors. Since C# is a Java clone and it is object oriented I don't except it will perform better than C. BOTH EQUAL

    6. TOOLS: May be MS can have better tools in the market. But here too, I have lot of choice. I like Eclipse, NetBeans, JEdit, Apache, SourceForge etc with lot of committed people around here. Since MS have lot of money and expertise to produce quality tools, eventhough it looks equal, MS has slight advantage here. ADV .NET

    7. TRAINING: People who are complaining about complexity of Java, J2EE and don't want to read ever, I recommend training. [Eventhough in most cases I found reading manuals and documentation is more than sufficient. But I don't believe a VB person can do program easily VB.NET without training and same applies to a person from C to Java. so BOTH EQUAL.

    8. IMPLEMENTATION: MS has released its 65 security bulletin yesterday for this year. Till 1999 [what about now???] , MS did not have a reliable server platform. They were just stand alone boxes. UNIX is a matured secured platform to implement enterprise level applications. ADV JAVA.

    So I request all the NEUTRAL managers to turn towards Java and J2EE. You are the people who can make a difference. You are like those middle class family who are really concerned what is good for the company but never vote. But believe me this time you can make a difference...
  66. 6. TOOLS: May be MS can have better tools in the market. But here too, I have lot of choice. I like Eclipse, NetBeans, JEdit, Apache, SourceForge etc with lot of committed people around here. Since MS have lot of money and expertise to produce quality tools, eventhough it looks equal, MS has slight advantage here. ADV .NET


    Well, this really is a myth. There are better (more productive, more options, better thought out ...) IDEs out there for Java then Visual Studio .NET for C# is. Look at to JBuilder and IDEA in addition to your list. For example, are there refactoring support in VS .NET, how about code formating or code metrics ?
    However, it is easier to write GUI's with VS .NET and Java IDEs do require more memory to run smoothly.

    Regards,
    Mileta
  67. 1. DESIGN and EXPERTISE

    I think this is an experience thing rather than a .NET vs. J2EE or C# vs. Java. I've worked with both and I worked with C++ before. I've been using patterns for a long time. I've met C++, C#, Java, and even VB programmers that know patterns. I've also met C++, C#, Java, and VB programmers that don't know anything about patterns. I do believe that Java programmers use more patterns because there are more resources out there. Until .NET came out, the Microsoft programmer community was large but only for specific problems because the glue (COM/COM+) between environments (C++/VB) couldn't support a lot of the major patterns.

    > 2. STANDARDS and PORTABILITY
    Microsoft is the king of backward compatibility. I cut my teeth coding in C on Windows 3.1. Those apps actually still run on Windows XP, and I know most of the APIs I've used are deprecated, but since they were publicly documented, they will always be supported. Also with .NET's side by side execution, it is even easier to get backward compatibility (although I haven't tried this myself the documentation sounds good :-))

    >3. CHOICE
    Today you have no choice for commercial development except Windows. In the future? I don't know. Project Mono is going pretty well. Rotor (the shared source CLI code) is working on most platforms now. Granted you can't use Rotor for commercial software. But it proves that they can get it to work. And since ASP.NET was written in C#, technically it should be able to work as well on the other platforms.

    >4. HISTORY
    I really don't have anything to say about this. To a certain extent, every language that comes out solves a problem that had already been solved. C came out. WHy do I need C? Assembly works fine. C++ came out. Why do I need C++? I can do it in C. Etc. Etc.

    >5. PERFORMANCE
    I agree with you on this. I think they will always be one-upping each other though. MS will come out with a better GC algorithm or better JIT compiling algorithm. Sun will respond. MS will respond.

    >6. TOOLS
    Back in the day, MS had the best debuggers in the business. And well they should. They owned the OS that they ran on :-) To this day, I still believe that MS has the best debuggers. I mean, stepping from C# code into my SQL Server Stored Procedure? That's pretty cool. For Java, I use IDEA. I personally like it better than VS.NET. I think that VS.NET is much snappier when redrawing the screen, compiling, and debugging, but I like the tools in IDEA (refactoring, automatic creation of methods, etc.) much better.

    >7. TRAINING
    I agree.

    >8. IMPLEMENTATION
    I believe that Windows will always have more REPORTED security bugs than any other OS. Why? The amount of people that want to see it fail. The amount of people using it everyday. This to me is a good thing. Because they are being brought out in the open and being fixed. I'd rather them be embarrassed for a couple of days and get a fix out, then to have it silently sit there being exploited by the cracker that was smart enough to keep his mouth shut.
  68. .NET Dissed at Comdex, SD East 2002[ Go to top ]

    I wish ... I wish I could talk that calmly on this topic. That was good!
  69. "2. STANDARDS and PORTABILITY: Forget about portability across different platforms, what about writing a dll that can run in all MS platforms. So All the concerned managers should start thinking that .NET is good, but do I have the assurance that Microsoft will make backward compatibility as an important issue. ADV JAVA
    "

    This really says it all. This group has no clue as to what .Net brings to the table. Have you heard of the concept of shared dll. The .net framework will acutomatically know what the right version of the dll to use. This is pretty embarrasing.
  70. <
    First of all I apologize if my comment on STANDARDS AND PORTABILITY is wrong w.r.t MS. This comment came from me largly due to the fact that our team in one of the previous project faced such a dll incompatiblity. But I am proud about what Java brings to me and my project in other areas.
  71. As an IT manager (but with admittedly strong personal bias towards Unix & Java) and maintaining professional detachment
    I agree with your points. However, as others have mentioned
    Microsoft has huge resources and substantial influence over
    the IT industry. Companies will decide based upon their
    particular circumstances and existing skills base.
  72. I do not believe better developers are developing in Java. Its really the individual developer's skills and descipline that makes a difference. I am a .NET developer myself and I pride myself in the quality of enterprise applications I develop. Software engineering principles apply to all languages; all real software engineers and architects know this.

    The other issue you did not address when it comes to J2EE vs .NET is the overall cost of a J2EE application. Its usually 2x a .NET implementation. Companies have limited budgets and it comes down to ROE.
  73. <shepard>
    The other issue you did not address when it comes to J2EE vs .NET is the overall cost of a J2EE application. Its usually 2x a .NET implementation.
    </shepard>

    How do you figure?
  74. The other issue you did not address when it comes to J2EE

    >> vs .NET is the overall cost of a J2EE application. Its
    >> usually 2x a .NET implementation. Companies have limited
    >> budgets and it comes down to ROE.

    I too would be interested to know how you arrived at this.

    The average project cost breakdown as I see it is (in order of magnitude):

    1) Maintenance (70%) (industry figures). Here the .net platform suffers from its coupling to the operating system. Not only is its ability to consolidate limited (can we run two versions of the CLR?) but OS-patches have an impact on the application (XML support dependant on IE - why?)
    Currently, the support of the pre-.net Microsoft platform costs us dearly (MTS is a peice of junk).

    2) Development (25%). Here again, the poor integration with messaging, non-sql server databases, lack of non-windows OS support, lack of automated persistance, lack of ability to deploy components (non http, .net remoting components), poor non-windows security integration, all leads to extra development. Also, the lack of implementation alternatives in .net means that you are forced into coding around product defects/shortcomings - and it makes it difficult to find the source of the problem.
    While VS is a good IDE, it doesnt compare to dev environments like IntelliJ or Eclipse (which are almost free and free)

    3) License/hardware costs.
    Given that J2EE will run on whatever hardware/OS your corp uses, its difficult to see how J2EE hardware could be more expensive than .net hardware. More often than not, you can leverage existing hardware (consolidation).
    Linux is becoming the intel OS of choice for the budget conscious.
    Even on win2k, choose JBoss or Tomcat on Win2K and you are even.
    Choosing Weblogic, Websphere or OracleAS admittedly costs more - but we are now arguing about the 5% slice. Who cares - 5% is much less than the average project over-run.

    -Nick
  75. Nick,

    If, for instance, a software engineer (eg you) were given 100k salary, for 6 months, do you thnk they could produce a really brilliant petstore in .Net.

    If the answer is yes, then all of you arguments basically fall down to the fact that you prefer java.

    If the answer is no, then the developer in question is too thick to be part of this discussion.

    AND, you could turn this on it's head, and make the exact same offer to a .net developer, asking them to do a java one.

    Both techs do the job. All it needs is motivation, in this case 100k for 6 months work!

    And I'm not being devils advocate here. This EXACT situation is happening in loads of projects all over the world. Apart from the salary - the motivation in the real projects is because some manager, some place, says so.

    Jonathan
  76. So where are the big sites?[ Go to top ]

    That’s a good question.

    The J2EE guys all the time condescend to admit that "MS Products at the best good for small, noncritical sites", and so on. Ok, so where is the big time movers and the shakers J2EE sites? Somebody mentioned eBay - I put www.ebay.com at netcraft, but the result was IIS and NT, and anyhow they are not in the near of 75 mill hits per day. (.NET at Merrill Lynch with 75 million transactions per day at 25 million users that’s 21,000 transactions per second..)

    So where are the big sites? Can somebody enlighten us? Even with the 70% failure J2EE server failure rate according to Gartner, the must be someone left? Maybe there are J2EE that are good for "small, noncritical sites" or as a hobby for impractical theorists?

    I actually look forward to read some of the answers. They are very entertaining; the community is very good at squirming excuses. At least they are innovative in one area!

    Regards
    Rolf Tollerud
  77. Rolf,

    I'm from Brazil. Here we have a lot of big sites using J2EE. Brazil´s biggest bank (Bank of Brazil) uses J2EE in most of its systems, including the main web site (www.bb.com.br), which has a real BIG load everyday. Besides, it´s connected to a lot of legacy systems (which would be impossible, or at least very hard to do using other technology, like .NET), giving real time data about all investments, account movimentations, no matter where you are.

    Another big site is www.ig.com.br. It´s a great portal, interconnected with a lot of news feeding systems.
  78. Rolf,

    Let me just fix a misconception here. I am the Lead Developer at the Legacy->J2EE Transition Team at iG (www.ig.com.br), and I also worked a lot of time in the development of several of the content publishing systems and most of the news feed connection, and it is *not* J2EE at all.

    Until recent times, the connection with news agencies was not even Java-based - and it would never be if our first solution - a custom Tcl-based Vignette solution - didn’t become so expensive (largely due to Vignette's "Solaris or W2K" policy).

    Most content publishing is made either with Vignette or with a PHP-based tool created in-house. And they have handled the job so far - a job that demands HUGE load for both static and dynamic content, and lots of daily conceptual change in content management systems. I don't see any benefit that J2EE (or even Java) would bring to this "wild" side of the company.

    Just to clarify (and make clear we are not making clueless decisions), our tech people is not exactly an MS-fond crowd. In fact, the only non-UNIX systems we have are legacy systems from acquired sites. And the investment on Solaris and Linux infrastructure makes it easier (i.e., cheaper) for us to choose the J2EE road. It is quite ironic that it is a choice made largely due to the "monopoly" of J2EE as an enterprise application platform in the UNIX world... ;-)

    However, our use for J2EE stops at our internal corporate systems. Our experience in developing J2EE web application showed us that the gains in the long-term are not so appealing for content publishing in our web sites - and there are much less variables to deal (comparing to lighter solutions) when we talk about performance tuning for the #2 portal audience in Brazil (#1 sometimes in some measurements). This position may change sometime in the future - maybe when more "KISS" technolgy arises and matures in the J2EE world (like Java Data Objects, for instance), but not right now.

    From Brazilian to Brazilian: there seems to be a lot of big sites around WISHING to use J2EE. Reality check: most of them are in transiction state, and some are getting frightened by the greater-than-expected costs and pitfalls - including the IDE/productivity original discussion (good, I am not off-topic anymore :-) ). I would personally like it to see how well this technology adapts to our reality - whether we like it or not, MS technologies deal with some of our specific needs in a visible, widespread way.
  79. (in fact, it was not for Rolf, but for Andre Aragao. Whoops)
  80. Carlos,

    I didn´t know iG does not use J2EE in its main web site, only inside the corp.

    I work for TCO Celular, and here we are in the transiction state you mentioned, but we already have some applications using J2EE (minus EJB), with real performance (yes!!) and mainteinability gains.

    It´s true, there are a lot of people getting frightened. But I think it´s more because of paradigm changes ( an almost procedural aproach to systems, using ASP + VB + ORACLE/Stored Procedures to an oriented object analysis/implementation ) than because of J2EE itself.

    Microsoft development technologies hides a lot of aspects involved to get a system running. It´s good when you´re dealing with small to medium systems, but not when you´re working with real time big systems. In this case there are a lot of technologies better, and J2EE is just one of them.

    Thanks for clarifying,

    André
  81. So where are the big sites?[ Go to top ]

    <quote>
    So where are the big sites? Can somebody enlighten us? Even with the 70% failure J2EE server failure rate according to Gartner, the must be someone left? Maybe there are J2EE that are good for "small, noncritical sites" or as a hobby for impractical theorists?
    </quote>

    Am I the only person tired of these condescending comments? No one makes you use J2EE. Why do you come here to talk trash?
  82. Some people need growing up[ Go to top ]

    Am I the only person tired of these condescending

    >comments? No one makes you use J2EE. Why do you come here >to talk trash?

    No Lotsa,

    you not the only person who is tired of these childish arguments! These comments of "Ive spent X yrs using Y technology" nature, which 'implies' that Y technology is better than Z technology and/or that I am more correct than somebody else who may have spent only (X-1) yrs using that technology are very stupid! Guys, please grow up!

    And for all those Java and J2EE bashers (or better yet losers), please give others the right to choose! (I would consider myself a loser is I spent that much time convincing others that the technology that they like/use/prefer will fail/(is not good!), or at least if I was spending that much time on the site dedicated to technology that 'sucks' when there many sites out there about the technology that is great and I enjoy, and time is more than money!)

    Regards All,
  83. .NET Dissed at Comdex, SD East 2002[ Go to top ]

    I have been doing Java development since 1996. I switched to Java from MS platform (VB and VC++) because I was initially sold on the concept of separating intentions from the implementation. Microsoft was, at that time, too focused on the implementation, too close to the metal for my taste. Java gave me the promise of a nice, detached, hands-off approach to software development.

    But nowadays I'm severely disappointed in Java (read: J2EE). Apparently, gone are the days of keeping the concepts cleanly separated from the underlying implementation. Nowadays, J2EE is ridiculously obsessed with the infrastructure. Look at any J2EE-oriented product, all of them talk about infrastructure, and only about infrastructure. It's all about MVC architecture, portals, messaging, connector architecture, and such. I must say I am very tired of that. Same as I'm very tired of talking about databases, XML, email servers, file servers, etc. It's all very old, very tiring, and should by now recede into the background.

    Whatever happened with "you do the thinking, we'll do the implementation" philosophy? I was expecting J2EE to do the legwork for me, not to keep throwing all those ugly infrastructure issues at my face. In that respect, it failed miserably.

    So, I couldn't take it anymore, so I finally decided to join the dark side, and tried to switch to .Net. I was expecting to find much more reasonable philosophy in the more user friendly environment that Microsoft is seemingly peddling. But, oh man, what a rude awakening. Somehow, anything that these guys do is even more offending than J2EE! It just sucks (I don't want to go into the very depressing discussion here on what's so appaling about .Net).

    My bottom line question is: isn't it high time we stop obsessing over the computing infrastructure, move on, and let the underlying devices (such as J2EE or .Net) take care of that whole charade? I honestly have better things to do with my development time than worry about stupid quirks that J2EE and Turbine and Jetspeed and other stupid frameworks are throwing at me. Right now, I find them next to useless.

    Alex
  84. .NET Dissed at Comdex, SD East 2002[ Go to top ]

    "Nowadays, J2EE is ridiculously obsessed with the infrastructure."

    I believe the aspect-oriented frameworks that currently popup in the J2EE world will help a lot to reduce the amount of code written that is dedicated to "infrastructure".

    Some interesting projects in this field include jboss, nanning, jAdvise and of course Rickards much waited for AOP-framework
  85. .NET Dissed at Comdex, SD East 2002[ Go to top ]

    I honestly have better things to do with my development

    >> time than worry about stupid quirks that J2EE and Turbine
    >> and Jetspeed and other stupid frameworks are throwing at
    >> me. Right now, I find them next to useless.

    So... you develop it all yourself?
    I cant see that being a worthwhile investment of resources...

    It would be lovely not to have to think about the complexity of building enterprise systems - but it is a sad reality.

    -Nick
  86. .NET Dissed at Comdex, SD East 2002[ Go to top ]

    I honestly have better things to do with my development

    >> time than worry about stupid quirks that J2EE and Turbine
    >> and Jetspeed and other stupid frameworks are throwing at
    >> me. Right now, I find them next to useless.

    So... you develop it all yourself?
    I cant see that being a worthwhile investment of resources...

    It would be lovely not to have to think about the complexity of building enterprise systems - but it is a sad reality.

    -Nick
  87. .NET Dissed at Comdex, SD East 2002[ Go to top ]

    I honestly have better things to do with my development

    >> time than worry about stupid quirks that J2EE and Turbine
    >> and Jetspeed and other stupid frameworks are throwing at
    >> me. Right now, I find them next to useless.
    >
    >So... you develop it all yourself?

    Pretty much. Is there an alternative?

    >I cant see that being a worthwhile investment of >resources...

    And wasting months on fiddling with and learning those half-baked frameworks, only to ditch them in the end, is a worthwhile investment of resources?

    >It would be lovely not to have to think about the >complexity of building enterprise systems - but it is a >sad reality.

    I'll tell you what is a sad reality -- the fact that vendors are giving us crappy frameworks, and expect us to be enormously happy for that. They can definitely do better, we all know that, they only need some kick in the butt. But if we're keeping silent, why should they be busting their hump?

    Alex
  88. .NET Dissed at Comdex, SD East 2002[ Go to top ]

    Pretty much. Is there an alternative?


    Er.. Yes! is the short answer.

    >> And wasting months on fiddling with and learning those half-
    >> baked frameworks, only to ditch them in the end, is a
    >> worthwhile investment of resources?

    We happily use Struts and J2EE (CMP2, MDB, EJB, JMS, J2EECA, JAAS, Servlet/JSP, ...).
    We are now looking at Webwork and Ofbiz entity engine.
    We havent *spent* (let alone wasting) months looking at them.
    We spent less than an hour looking at webwork - and while we are far from experts, we saw enough to see the value.

    If you want to write everything from scratch, then good luck to you. Common sense should tell you however, that its probably been done before - more likely by smarter people - and they have been through a few iterations to get their framework where it is.

    However, if wheel re-invention is your game...

    -Nick
  89. .NET Dissed at Comdex, SD East 2002[ Go to top ]

    We spent less than an hour looking at webwork - and while >we are far from experts, we saw enough to see the value.


    Same is with us -- we spent an hour looking at Jetspeed, for example, and that was enough to see its value. But then when we tried to use it in real life, we found that it's heavily based on Turbine, and that we'd need lots more time to invest into learning the Turbine... and so on. After a month or so, we've realized that it was time to throw the towel in. So, one month's time down the drain. Not something we can write off that easily...

    >If you want to write everything from scratch, then good >luck to you. Common sense should tell you however, that >its probably been done before - more likely by smarter >people - and they have been through a few iterations to >get their framework where it is.

    We definitely don't want to write everything from scratch. Otherwise, we wouldn't even be looking at those frameworks. However, the problem is that even though there are snippets of code written by smarter people available around, it doesn't mean that such code in organized in a manner that's going to save us time and effort. Some of those open source frameworks openly brag about the elite status of their code, and tell potential customers to buzz off if they can't figure out their exalted engineering practices. You see, if I want to reuse someone's piece of code, I want to do it in the simplest possible way, with the smallest possible fuss. I definitely don't want to make a career out of it (which is what frameworks such as Struts is forcing you to do). I don't want to go through a complicated song and dance in order to get a piece of code to work. Which is what it feels like if you want to utilize frameworks such as Turbine, for instance.

    Turbine fans tell me that, if I can't stand the heat, I should stay out of the kitchen. That's why I ask -- is there an alternative to reinventing the wheel, or to be subjected to a torture by those snotty engineers?

    >However, if wheel re-invention is your game...

    Where did you get the idea that reinventing the wheel was my game?

    Alex
  90. The thing that infuriates me is that why are M$ trolls always hanging around in the server side. Do we have the freedom to choose any platform to develop in ? If MS see that’s as a threat then I am sorry. I am not sure creating phony results with TCM and showing Java is slower will not change our mind rather completely avoid .Not as a whole.

    Don’t they have any thing better to do than just come here to comment on .Not is so good or is it that no one is using .Not and their .Not site is empty.

    This has been history of MS to dog other products that they copied from and then show off their stuff runs better.

    We know the history of proprietary behavior of MS and in the end the product will just run in windoze.

    I know ms trolls will say but we submitted to open source then why is the .Not server proprietary and why does M$ dish Linux in every breathe.

    We have tried .Not here and it is as complex as java ..they are also trying use design patterns and any VB programmer will be perplexed by .Not.

    Technically .Not is still short coming I know you ms trolls will argue till the cows come home but please go back to ms website and post there and I know it is empty there but as the wise men says talk to the hollow wind and you should gain wisdom.
  91. .NET Dissed at Comdex, SD East 2002[ Go to top ]

    .NET vs java.

    I think the point here is not which is better Java or .NET. The point is what business wants and how business wants it. These days companies are very serious about how much do they want to spend on IT. And, Java is great if you know what you are doing. The ones who know what they are doing know how much to ask for what they know. So, java developers are paid big Dollars. Those can develope in Java fast and efficient and can come up with great solutions, enterprise or small company. VB developers (probably 80%) started as non-programmers and converted to IT from other professions because VB was easy. But, now VB.NET is not really as easy as it was cause it is OOP like.
    So my point is business is always going to look for ways to cut costs to make money. Business is about money not Programming languages. If C# with VS.NET is going to safe them big dollars they will go for it even though it may be a Java monkey. If Java will safe them big dollars than Java it is.
    However, these days they will take any out-of-US resource to do the job on any language. No matter if it is Java, Python, Ada, or C++. The goal is not to invent more typing and more developing. The goal is to invent and automated process to completely or at least partially eliminate human envolvement in building software. This is when business is going to make a TOP dollar. Better IDE, less coding, more automation, robots in Cafeteria means no big salaries to pay, no employee benefits to worry about, more money in the pocket, big bonuses, steady tax payments, GOOD for Economy as a whole.
    And, as for us (developers), we are going to service the ROBOTS.

    Cheers!
  92. .Not is dead.... Java and Linux rules[ Go to top ]

    Seems to me .Not and windoze is dead

    http://www.vnunet.com/News/1137229
  93. Good architects do not limit themselves to petty bigotry over different computing strategies. They choose the right toolset to solve the problem at hand. You get bad J2EE solutions as a results of bad J2EE architecture. Ditto for .NET. Both of the competing frameworks have their respective bright spots and both have their 'black holes'. What the software community should be more focused on on is solving the resounding unheeded complexity one encounters in a LOT of the projects out there. 5 year experienced 'architects'. 2 year experienced 'senior developers' - that's where the problems with J2EE and .NET *arise*. Buiding Escher like woven balls of crap just to suit some deep seeded intellectual need of the complex (mental masturbation) is what is going to kill the cat - Not .NET vs J2EE.