Discussions

News: Intel and HP Back Open Source Version of .NET

  1. Computing heavyweights Intel and Hewlett-Packard said Monday they will lend support to an effort to create an open-source version of Microsoft's .NET initiative, called Mono, which aims to aims to develop a version of the .NET programming environment that runs on Linux and UNIX. If this project is successful, J2EE will lose the cross-platform advantage it always had over .NET.

    Read more at
    Open-source competition for Microsoft's .NET

    What is HP doing? HP is support J2EE by releasing free J2EE Application Server to J2EE 1.3 supporting server to developers, at the same time, the company also plans to support .NET initiative?

    How will "Mono" really affect J2EE vs. .NET marketshare?

    Threaded Messages (32)

  2. It makes perfect sense for HP to back this initiative. Mono would allow .NET projects to choose unix-based HP servers instead of only sticking to WinTel. Their decision makes good business sense - I don't think it in any way reflects a dropping commitement to J2EE on HP's part.

    They have invested a lot of time and effort into completely re-writing their new HP-AS J2EE 1.3 server from scratch.

    Floyd
  3. There's also the question of how portable .NET applications really will be.

    The basic .NET infrastructure of the CLR is not all of the picture. There are also the supporting class libraries and services, many of which Microsoft delivers as part of Windows, not .NET. Unless these classes and services are also ported, then very few .NET applications will really be truly run-anywhere.

    I'm very much in the "wait-and-see" camp on this issue.
  4. Robert - quite.

    The class libraries do form quite a substantial part of the offering. Which, if any, they choose to implement independently will determine portability. For web apps, luckily alot of the classes are irrelevent, but nonetheless you will need some ADO, Response & Request stuff at the very least.

    The other hurdle is the db - if you want the elegent MS dev environment through and through, the natural data source is SQL Server. To then deploy on UNIX, you'd probably want to retain SQL Server as the DB. So the partnership of UNIX middle tier and SQL Server backend partnership would need firming up.
  5. I think it's great when there is an OpenSource .NET. One more plattform to choose for your projects. As .NEt is more or less a Java clone (with some nice enhancements), there is no steep learning curve for us.

    But, there are two problems:
    1.) Will there be WORA like in J2EE ? I believe this, when I see certification tests. I have doubts about the effort to reimplement the APIs, it's always difficult to code behind Microsoft.

    2.) There might be patent issues hanging around. MS could probably stop OS .NET any minute, if it's starting to become a threat and myServices doesn't pay the rent. They stopped OpenMail from HP, because it was a threat to Exchange.

    3.) Ximian products are not always very stable.

    bye
    -stephan
  6. Learning C# is quite straightforward, if you are comfortable with java. .NET is more like learning J2EE. There is little common ground here apart from what they are trying to achieve.

    Quentin
  7. strong the force that lies in this competition is. good the competition for both sides will be.
  8. Right on target, Robert! .NET is tightly integrated to the Microsoft OS it runs on. Just getting a CLR port does not do it, they will have to code a lot of "overhead" to emulate/simulate those Microsoft services.

    Nobody has come up with a commercially succesful COM/DCOM/COM+ framework on Unix just because of that. Does Intel and HP know better? We'll see.
  9. ".NET is tightly integrated to the Microsoft OS it runs on. Just getting a CLR port does not do it..."

    Why do you think .NET is integrated to Windows? By .NET in this case, are you talking about ASP.NET?

    "Nobody has come up with a commercially succesful COM/DCOM/COM+ framework on Unix just because of that."

    Native .NET applications do not use COM to communicate with each other.. that's the point. The CLR provides a much simpler substrate for interfaces and common data types. The CLR will also allow communication with "legacy" COM objects.
  10. MSoft has no reason to ever want support from another OS. And they have shown again and again they will sow FUD, and do not care. Windows XP is acually amazingly solid. Its really simple to install and doesn't crash every 5 minutes. 3GHz PC's will be out this year. MSoft has no reason to want to run under Unix. Most apps will run on wnidows platforms now.

    The Unix vendors, on the other hand, are all wetting themselves. And Linux has to fight hard (even harder now XP is out) to come out and be used on the corporate desktop.
    The only Linux advantage is cost, but if msoft only costs £100 per year, per employee then its pretty cheap (eg v's tax, rent etc).

    From msofts point of view .NET should be marketed as 'open' and 'cross platform', but there is zero reason to make it so.

    I imagine that anything using the .Net brand on any other OS will have msoft lawyers stomping on them pretty hard.

    Jonathan
  11. Nobody has come up with a commercially succesful COM/DCOM/COM+ framework on Unix just because COM/DCOM/COM+ was highly integrated into Windows, depending on a lot of Windows services in general and OLE in specific.

    I didn't mean to say that .NET depends on COM.

    .NET is not using COM for local communications (except when talking to unmanaged code). Fine.

    Where is the security component in the .NET picture? In the OS (Active Directory). Policies? In the OS. Web services? In IIS. Asychronous support? In MSMQ. Instrumentation? In WMI. The list goes on... I had the .NET beta-2 on 4 CDs. 3 included development tools, goodies, whatever. 1 included OS updates.

    Anybody trying to implement .NET on a non-Microsoft platform will have to implement a lot of Microsoft OS services as well.

    On a conspiracy theory dimension: could it be Intel and HP are riding shotgun for MS? MS itself cannot come out in the open to go for open-source now, though, it will no doubt benefit from open-source .NET innovations.
  12. Should they back Halcyon's iNET in stead?
  13. Nothing like a naming an idea after a kissing disease.


  14. Don't know why they are wasting resources on this BUT they better keep it far off of Microsoft's OS if they expect it to work. Real far.


  15. The more I think about .Net and in particular the CLR, the more suspicious I become that MS have a good chance of pulling this one off.

    While the competition has been led into a 'Web Service War' MS has been slowly advancing the real threat to Java: .Net and the CLR. These two components form the core of MS's response to J2EE and Java, and in fact are arguably a little more compelling because the CLR is multi-lingual, the tools are easier to use, the choices are more limited and therefore less daunting. The CLR offers the possibilitty (and it is just a possibility at this stage) that organisations will have the flexibility to develop applications in the language most suitable for the task at hand and have them run in a common environment and framework.

    The standardisation of both C# and CLR is also a factor, and Sun has long been criticised for not letting Java go a little (if you love something set it free...).

    The thing that scares me the most about the whole .Net thing is how impressed and uncritical a lot of the normally anti-MS zealots have been once they have seen it in action and delved into some of its technicalities.

  16. HP-AS J2EE application server seems to be very solid and its free!

    I think it makes sense that HP is getting behind .Net.

    Its good for HP to hedge their bets and give MS diehards an alternative server environment.

    Competition will make Java and J2EE stronger!

    I heard someone in an earlier post praise .Net multilingual support. Java is multilingual as well. Just look at Rhino (JavaScript), NetRexx (Rexx), Bistro (Smalltalk-like), Jython (Python) and more.

    A while back, I wrote a series of articles on programming languages that work in the JVM.

    Series on Scripting Languages for the JVM for JDJ

    The series highlights the different languages and how to integrate with other non-Java languages with BSF.

    I am a big Python fan, and I can use it in the JVM already.
  17. The links are broken. Can you update them.
  18. HP-AS J2EE application server seems to be very solid and its free!

    I think it makes sense that HP is getting behind .Net.

    Its good for HP to hedge their bets and give MS diehards an alternative server environment.

    Competition will make Java and J2EE stronger!

    I heard someone in an earlier post praise .Net multilingual support. Java is multilingual as well. Just look at Rhino (JavaScript), NetRexx (Rexx), Bistro (Smalltalk-like), Jython (Python) and more.

    A while back, I wrote a series of articles on programming languages that work in the JVM.

    Series on Scripting Languages for the JVM for JDJ

    The series highlights the different languages and how to integrate with other non-Java languages with BSF.

    I am a big Python fan, and I can use it in the JVM already.
  19. Hmmm... I'd say "Let's wait and see". I remember all the hype behind the DCOM being ported on Unix, etc. It turned out that things were tightly coupled with the OS, a little too tightly ;-)

    But I also beleive that the general competition is good for the quality of the competitors.

        Yagiz (http://www.erkans.com)
  20. Wouldn't hold my breath on Mono. It smacks of the kind of VC and management thinking that's completely divorced from the realities of enterprise computing.

    1)Very questionable portability...

    2)It's taken about five years for Sun, IBM and a whole slew of other partners to get J2EE in the shape it is today. So Ximian and Intel and HP are going to re-write virtual machine, class libraries, enterprise APIs from scratch. And we all know how much support they're going to get from Microsoft.

    I like Open Source when it makes sense i.e. JBoss. However, these guys are talking about re-inventing the wheel in a combination of equally blind Microsoft admiration and paranoia.

  21. Chip,

    You do realize that Mono currently has a VM and a compiler working, right? I don't know how stable it is, but I know the compiler can compile itself. Class libraries and APIs are a lot of work, but I'm impressed they've made it this far already.
  22. and the guys are the guys behind gnome that sun considers interesting enough to be adopting for their desktop in solaris... These guys are no idiots.

    One moment I think that this mono project is just trying to achieve something that isn't possible, or there’ll be so many differences between the platforms that portability is too hard, the next moment I think they might just pull it off. Only time will tell.

    Quentin
  23. Though I resent M$ for a variety of reasons, I have to admit they have amazingly good business sense (mainly Gates, I think.)

    First they tried to tie to Java to Windows. When they lost that battle, they figured they needed to attack Sun where it was weakest. Ironically, Sun's biggest mistake thus far has been to keep Java proprietary. An even bigger mistake is closing off J2EE to anyone without $500,000 for certification. M$ has successfully exploited these weaknesses to create the perception that .NET is a more open version than J2EE. In addition, it would not surprise me to find out that .NET is better than J2EE just by virtue of the fact that it is written from scratch.

    It's not too late for Sun, though. They are still in time to save Java and J2EE by making them non-proprietary. I just hope they take action before it's too late.

    Guglielmo





  24. HALCYONSOFT.COM(Inet) IS SLEEPING WITH THE ENEMY.


    *HUSH*

  25. Hi guys, please read and read Ximian reasons.

    Ximian boys want leverage Ximian-GNOME to compete with Microsoft desktop, but current used technologies makes it very difficult (using C, CORBA ...).

    I think that multilanguage integration could be got by developing interpreters in Java, because relevant languages other than compiled C,C++,Java,C# (ok, today not but will) are interpreted (JavaScript, Python, TCL, perl ...).

    These interpreters running in the same JVM can interoperate in the same memory address space: interpreted applications can share data.

    If Java standard class libraries are reflected in the scripting language by the interpreter (thanks to Java Reflection), we have a powerful multi-language platform ¡already! without reinventing the wheel.

    Moreover, ¡simple scripting applications might run in managed environment as J2EE-EJBs, and EJBs could call scripting code!

    JDK 1.4 will enhance Java reflection to support interoperability with scripting languages in any way, by example JavaScript: Netscape 6.2.x/Mozilla 0.9.x currently have LiveConnect (Java access to JavaScript code) disabled because are waiting to JDK 1.4 to resolve some issues (security?).

    Many languages are already running inside a JVM, the
    list is impressive.
    An article about it : IS JAVA LANGUAGE NEUTRAL.

    Ximian approach is intelligent, but very hard to do because some things must be developed from scratch (compiler, classes not addresed by GNOME libraries).

    Interoperability of Microsoft .Net programs in Mono-Linux is a utopia, because Microsoft will use native Win32 calls to avoid it (a Mono developer told me, that it isn´t a important issue to justify Mono strategy).

    Leverage GNOME with more powerfull development tools as .Net, can be achieved alternatively using perhaps JNI from Java calling GNOME libraries. In this way is not necessary rewrite all GNOME to use Java framework. New components can be developed in Java but using existing infrastructure.

    Note: other approach is use Mozilla-XPCOM because marshaling overhead to communicate components in the same address space is lower than CORBA (I´m not a expert).

    Java is not truly opened, ok, but is more opened than .Net, Java programs running already in many platforms, there are gateways to interoperate with native code, to be reflected in other languages ...
    .Net is a only a promise today (but perhaps ...).


  26. Portability of .NET[ Go to top ]

    The MONO effort is very interesting, particularly with Intel's and HP's backing. I wonder what those two are contributing that caused the switch from GNU GPL to X11 licensing.

    But here is my take on .NET portability at this time. I just attended a .NET versus J2EE "debate" (not really one) in Reston, VA. There were two Microsofties there, and I quizzed them on Microsoft's stance versus open-source, Corel, etc.

    First, the C# compiler and CLR are not at all equivalent to J2EE. You have to include ASP.NET and ADO.NET, and neither of those technologies will go the way of standardization. Now Microsoft can't prevent reverse-engineering, but the evangelist was skeptical that it could be done (and he might be right there). There is also a question of intellectual property and whether any efforts to create "open" ASP.NET and ADO.NET might run into legal problems (my thoughts, not his).

    Second, it appears clear that Microsoft is positioning non-Windows .NET implementations for academic uses. They want people to tinker with it, realize the coolness of C# (or any other .NET language) and CLR, and then move to the Windows .NET platform for industrial-strength programming. If Corel puts out a .NET implementation on FreeBSD, it would most likely be targeting educational markets. Microsoft wants mindshare among the future programmers and that starts in the colleges.

    So, .NET will indeed be portable, but it depends on what you are building. If you are trying to build an entire e-commerce application, I think you'll have to stick with Windows-based .NET. If you are learning languages and writing simple applications, you will probably have Mono and Corel's platform sitting on non-Windows OSes.

    The Halcyon transcoding approach (iNET) sounds like a much more direct competitor to Windows-based .NET, since it includes analogues for ASP.NET and ADO.NET. Time will tell if their solution is robust.

    If Microsoft ever allows "open" ASP.NET and ADO.NET, they will effectively remove the greatest advantage of J2EE - and that's future-proofing through vendor freedom. I would urge Microsoft to go this route since they will always lead other .NET implementations due to better tools and performance metrics.
  27. Portability of .NET[ Go to top ]

    I think that multilanguage integration in .NET is nothink but marketing becouse I can not create application to run in Common Language Runtime using regular compiler and development IDE. There is only possibility to create new compiler and IDE to produce code to run in CLR. Such possibility also exesit for java JVM. CLR is nothing more than runtime for C# in the same way as every C++ compiler has his own runtime. MS sugest that this runtime can be use by over languages.

  28. So what?[ Go to top ]

    Java is already running on many plataforms, I use it over Win32 and Linux with *zero* problems. This is only a marketing strategy, the biggest voporware startegy I have seen in many years. Like some people have said, porting the CLR enviroment it's only part of the problem, where do .NET API calls go? Win32 of course... so you will have to port a great portion of W32 to have real .NET plataform. So.. we're going to use WINE to emulate .NET? Wake up people... the only real cross-plataform solution is Java, let's leave this ".NET over <your favorite OS>" for science fiction writers.
  29. So what?[ Go to top ]

    Tyler

    You may well be right on this. Strangely though in ’95 I was an NT/Windows and Solaris/Motif C/C++ developer. I seem to remember being equally dismissive about my colleagues’ enthusiasm for some new cross platform language / technology that Sun and Netscape were making noises about. I don’t mind saying that I got that all wrong. There is some similarity between the vision that those guys had back then and what these guys are looking to do now.

    Quentin
  30. So what?[ Go to top ]

    Big difference, M$ is involved with this.
  31. So what?[ Go to top ]

    Big difference, M$ is involved with this.


    So what?
  32. So what?[ Go to top ]

    Where have you been living for the past several years ?

    Look at ALL the other dev tools that attempted to play in the M$ sandbox. HP, Intel, iNet, etc are just ramp-up tools to herd the masses to M$'s OS. They are all tight M$ partners. Shit will start to fall apart and stuff will magically work in m$'s OS.



     


  33. And from what I gather, for the money they are paying, Intel and HP will make proprietary changes to Mono and they will not put in open source. Nice touch!