Could this be done?

[whitebread]

So I was thinking that it might be nice to be able to write an application in VB .NET (since it's easy to use and learn) and compile it to native code for faster execution (and so I don't have to worry about the .NET runtime being installed, and to prevent decompilation to MSIL -- I realize native code can be decompiled and such, but it isn't as easy as MSIL since it is quite easy to decompile MSIL to recognizable source code). A tool like this exists (http://www.remotesoft.com/linker/ ) but it costs $1249 for a license that allows 1 developer to use the tool, and $2219 for a license that allows 2 developers to use the tool.

I know that there is a free Delphi IDE called Lazarus... So I was thinking it wouldn't be too hard to write a program that converts simple programs to Object Pascal which can then be compiled to native code. Obviously it would be very difficult and time-consuming to re-write the whole .NET Framework, but basic stuff wouldn't be too hard.

Assuming I was able to accomplish this, would there be any market for this? I wouldn't really be looking to make much money; I'd probably sell the program for $0.99 just to help cover hosting costs.

[Phouchg]

Conversion between languages with brute force, statement by statement, would probably not be an overly complicated thing.

But why not use Lazarus to begin with? If basic functionality doesn't give you the things you need, there are many quality components and units (on Torry) that, afaik, can be converted with Lazarus itself and work well enough.

That said, there's always market for every crud if the marketer is good enough.

[+LogicalApex MVC]

Is there a market? Not really.

Your intended goal is to make it harder to reverse engineer MSIL compiled programs (C#, VB.NET, F#, etc.) and that capability already ships with Visual Studio in a light form (and has since .NET 1.0) via Dotfuscator Community Edition. There is a paid version of Dotfuscator too that supports more features as well.

So, there is a market for tools that can help a programmer to make reverse engineering his software harder, but there isn't a huge market for taking it through major hoops to get there. Obfuscation is pretty standard fare on .NET and Java (both are IL based platforms) for a reason (by using the same underlying framework and just doing alternations that don't affect the way the program actually executes you reduce your chance of introducing new bugs that are hard to find).

As a result, I would recommend you look at tools like Dotfuscator and see what you could realistically do better. But most devs won't want a language conversion then a recompile to be added to that process. As there are just too much room for new bugs that they aren't responsible for. Also, the "pain" of running the .NET Runtime have really subsided as Windows Vista and 7 ship with .NET out of the box. Users have no clue when an app uses the runtime and really don't care either way.

[whitebread]

^ Excellent answer. Thank you. :)