Grigore,

The primary function of the linker in the context of MS Windows is to create a PE executable image from the OBJ files emitted by a compiler. It will combine all the sections (for example code and data) from the OBJ files (and LIB's) into the relevant final section of the PE, on the way it is calculating the Relative Virtual Address (RVA's) for things like global variables.

Firstly, why would this be of benifit to DOTNET, after all its a managed environment? In fact the only time a linker becomes relevant is if you are using Visual C++.

When you link code statically using the linker utility you pull in all sorts of stuff you don't really need, but it is repeated from one executable to another. How many executable images on an average Win32 machine are bloating our drives with duplicated code?

From a deployment view the runtime (be it VC, VFP or DOTNET) tends to irritate people. Why? Usually size and versioning. One of the key benifits of DOTNET is the ability to create and deploy side-by-side assemblies allowing multiple versions to be installed, identified and ultimately executed.

Why is it of benifit to redistribute the same code every time you deploy an executable? Considering that pretty much everything from the BIOS to the Operating system is just a glorified runtime why do people make such a big deal?

Just my 2p worth.

Neil