Re: [DotGNU]Bending the twig of .NET (large -- sorry)

[Martin Coxall]

> Where can we find information on this that has been officially
> published by Microsoft (or by others in a Microsoft-approved
> manner) that allows to verify these claims?
>
> Greetings, Norbert.

Briefly, the VES offers the following feautures that JVM does not:

Type-safe Enumerations
Extensible metadata
PInvoke (JNI is clunky)
Value types
extensible loader (java.lang.ClassLoader is woefully coarse-grained)
An intrinsic IL emitter for generating types on the fly
An extensible context model
An intrinsic XML-based serializer

IL itself offers at least the following advantages over java bytecodes:

It provides greater type neutrality (helps implementing generics)
It provides greater language neutrality through the open-standard CLS.
It is engineered to always be compiled to assembly language before executing, 
and never interpreted. This should offer significant performance advantages 
over Java bytecodes, which were designed to be interpreted. I don't 
understand the full details of this, but it has something to do with making 
guarantees about the type of data that will be found on the stack at certain 
times, in a way that Java cannot, allowing far more optimised JIT compiler 
output.
It allows additional declarative information to be added to classes, methods, 
for use with System.reflection.emit() for on-the-fly IL code generation, and 
self-modifying code.

---
Martin
---

"Where laughing and smiling are not allowed"