About the only thing you have to worry about is a new version of Windows not being able to run/install the Visual Studio 6 IDE and/or not being able to run the VB6 runtime libraries in the future.

There will never be a time where maintenance and band-aids won' be cheaper than a complete rewrite... look at the Y2K scare: Companies were paying programmers to come in and repair 30-year old software! Plus, with faster hardware, you'll be able to EMULATE older hardware fast enough to still run ancient software.

But it never made sense to me either to not convert. My recommendation is to start all new projects in .NET. I've never met a VB6 program that wasn't 10x simpler to do in .NET. If VB6 is like a toolbox, the .NET framework is like a fully stocked mechanic's garage.

Heck, I've known programmers and IT departments to "loose" their VB6 install disks just so they can force the corporate wonks to agree to the conversion of legacy applications.

After you're proficient in .NET where you'd never even consider touching VB6 again (should take about 6 months), then bring up the topic of converting the program as a side-project. Put in like an hour a day rewriting it. Aim for 1:1 functionality. Don't do anything fancy or add any new features unless you're sure they won't take any extra time. After a few months, you'll be done. You'll have a perfect conversion of your program in .NET 3.5 or 4.0 or whatever.