Hi Rick,
I have seen a few different ways users have done this. The first is to write their reporting code in .NET and expose through COM interop; which I think Ludek has suggested and is the most common. I have seen one or two cases where the users doing this has had issues that could not be duplicated in pure .NET. But others have had no problems at all... no known rhyme or reason.
If I had to hazard a guess I would say that our .NET objects are actually calling a COM object in the background through a COM interop. This means your application goes through an interop from COM to .NET, .NET goes interop to our back ground platform then all the way back and perhaps something goes wrong in this process?
The other method I have seen people do is create a .NET application that you either shell out to, or that is running as a service. Actually the service solution I have a seen a few times. Developers write a reporting service that ships with their application. This service then can be called from their front end application which was COM/Win32 based, then was available to them and working when they upgraded to .NET code. And the great thing was when they upgraded to their .NET code it still worked whether or not it was a Win or Web forms application.
Kind regards,
Trevor Dubinsky