 Ongoing Discussion on the Value of DotNet
started 3/13/2002; 8:38:51 AM - last post 3/18/2002; 3:12:25 PM
|
|
Ehud Lamm - Ongoing Discussion on the Value of DotNet 
3/13/2002; 8:38:51 AM (reads: 2255, responses: 10)
|
|
| Ongoing Discussion on the Value of DotNet |
|
Peter and Patrick are discussing the merits of .Net
Most of the issues they raise were mentioned here in the past (e.g., compiling various languages into Java bytcodes/CLR). Still the discussion is worth following.
When I first saw Java I thought that the languages isn't that exciting, but I did like the language-as-platform idea. Seems to me like the next step is going to return to the distinction between language and implementation (think Java on CLR), and we'll just be stuck with the (not very exciting) languages Java and C# (and VB.Net). This is going to be boring.
The next step after that, let me venture a guess, is going to be allowing interoperability between the two platforms, using emulation and other tricks.
I think there's a good chance that these phases will happen at roughly the same time, so there's still hope.
Posted to cross-language-runtimes by Ehud Lamm on 3/13/02; 8:47:07 AM
|
|
|
|
Peter Drayton - Re: Ongoing Discussion on the Value of DotNet
3/13/2002; 6:48:14 PM (reads: 1086, responses: 1)
|
|
|
|
Peter Drayton - Re: Ongoing Discussion on the Value of DotNet
3/13/2002; 6:50:32 PM (reads: 1074, responses: 1)
|
|
|
What form of interop are you referring to?
|
|
Ehud Lamm - Re: Ongoing Discussion on the Value of DotNet
3/14/2002; 1:10:10 AM (reads: 1100, responses: 0)
|
|
|
Cut&Paste is dangerous thing
|
|
Ehud Lamm - Re: Ongoing Discussion on the Value of DotNet
3/14/2002; 3:53:19 PM (reads: 1080, responses: 0)
|
|
|
Well, we had languages that didn't really cater for interoperability: you could link object files, but to make sure things work you had to sweat out all the details. Next we celebrated the idea of language-as-platform which allows for(supposedly) easy interop between languages running in the same env (JVM). Now have two such envs CLR and the JVM. So we are likely to want to interop between components available as Java bytecodes and code compiled for the CLR (and the other way around).
|
|
Chris Peterson - Re: Ongoing Discussion on the Value of DotNet
3/14/2002; 7:09:27 PM (reads: 1040, responses: 0)
|
|
|
Having just read the other CLR papers posted on this weblog, couldn't Microsoft simply (pay another company like Corel to?) write a JVM-to-CLR layer for the CLR runtime?
|
|
Dan Shappir - Re: Ongoing Discussion on the Value of DotNet
3/17/2002; 1:27:59 AM (reads: 991, responses: 3)
|
|
couldn't Microsoft simply (pay another company like Corel to?) write a JVM-to-CLR layer for the CLR runtime?
Microsoft certainly could, but I doubt it would want to. Microsoft has made a publicized move of yanking the JVM off of the Windows platform. This was done intentionally in order to undermine Java's claim to run everywhere (obviously you can install a JVM manually, but if the browser wars have taught us anything is that being preinstalled makes a dramatic difference). I would therefore be highly surprised at such a move (not so say that somebody might do it on their own).
|
|
Ehud Lamm - Re: Ongoing Discussion on the Value of DotNet
3/17/2002; 1:52:29 AM (reads: 1051, responses: 2)
|
|
|
Seems a reasonable explanation.
Notice that the war is fought on more than one level. As we all know most programmers (and even professionals and academics) are so preoccupied with syntax, and language semantics that they forget about the implementation details. Thus, it seems to me that M$ want to eliminate Java the language just as much as they wish to eliminate the JVM. The logical way to do this is to encourage (as in 'make them an offer they can't refuse') programmers to use C#/VB.Net instead of Java.
I am quite interested to know how M$ will do this in the academic market. So many CS1 courses are moving to Java that it is going to be a up hill battle to switch them again, this time to C#.
|
|
Dan Shappir - Re: Ongoing Discussion on the Value of DotNet
3/17/2002; 7:17:13 AM (reads: 1086, responses: 1)
|
|
|
MS's proposed settlement of its lawsuit with the DOJ (donate $900M of MS software to schools) is one indication of how. Basically, they have lots of money, and software, to give away. This can be done in the form of grants, contributions, kickbacks or whatever.
If MS would approach the huji CS department and offer to donate all the appropriate hardware and software, as well as set up grants for numerous students and so forth in return for switching from Java to C# I believe they will get a receptive ear. The downside from the University's POV is minimal: C# and .NET are arguably as good as Java and the JVM, and as likely to succeed in the market.
|
|
Ehud Lamm - Re: Ongoing Discussion on the Value of DotNet
3/17/2002; 11:08:13 AM (reads: 1156, responses: 0)
|
|
|
We'll just have to wait and see.
I think it is more about getting many textbooks published, and convincing some high profile schools.
|
|
Peter Drayton - Re: Ongoing Discussion on the Value of DotNet
3/18/2002; 3:12:25 PM (reads: 976, responses: 0)
|
|
|
I'm expecting that the release of Rotor (aka the Shared Source CLI implementation) is going to be very attractive to the academic community: you've got Microsoft pushing C# and the *CLR* in the commercial space, and then a very, very similar shared source implementation of both the VM and the compiler in the academic space, for use as a teaching tool and research platform. Heck, there's even a special version of Visual Studio.NET specifically tuned for use on class projects.
Admitted the JVM source is also available, but I don't see Sun going after (or recently, defending) this market nearly as agressively. Basically, no-one markets to developers like Microsoft, and the SSCLI gives them some pretty interesting ammo.
A while back I met Stuart Reges from the UoA - he's been doing some great work getting C# into the Comp. Sci curriculum there, as an alternative to Java for people learning OOP.
|
|
|
|