I know what you mean, but you have to understand the trend.
Since the open sourcing of Roslyn and the availability of a free (non-paid) Visual Studio edition that has support for extensions, the demand for an alternative, free .NET IDE on Windows has kind of vanished. That being sad, there's also no real point in continuing the development of a competing analysis/refactoring library (speaking of NRefactory here) because it's a pretty complex piece of software that requires a constant effort in maintainance and enhancement since both languages, C# and VB, are further evolving.
SharpDevelop would require a fundamental change - again - to keep up with the competition: The replacement of NRefactory with Roslyn. I think this would be a necessity because right now, NRefactory hasn't support for C# 6 features, and no VB support at all. Using Roslyn would make those available almost for free, plus support for code fixes/diagnostics. And there would be more time to improve other aspects of the IDE.
Another point I can think of is that the main developers of the past years seem to have less time to (and maybe less interest in) contributing so much of their time to this project. Therefore, unless someone is willing to seriously spend some time on it right now, I don't see this project going anywhere.
But, again, the thing is: There is nothing in SharpDevelop that I can't do with Visual Studio, and if there is something missing, you could just write a VSIX as well (for example, I just build a prototype of a semantic formatter that kind of resembles the behavior of SharpDevelop, with italicized field identifiers, etc. because that's something I actually miss in VS) Sure, SharpDevelop is more lightweight, but memory consumption of Visual Studio is not so much the problem, especially if you're not using any heavy-weight VSIXs.