Just a couple of quick notes from me here.
First, it's always good to hear from ya Samson, whether you're doing stuff or not.
Second, I'll agree that there's no need for any apologies Kayle, as others have said, one should enjoy their hobbies.
Now, with that out of the way... SmaugFUSS++ would be a good idea. I think the vast majority of people are sick of dealing with C by now. Seriously, younger people learn C++ out of the gate, and most of us old farts that cling to C still complain about it. I would not put much stock in the whining of people who complain about having to learn...
Point in case, the RaM project. We argued for *MONTHS* about switching from C to C++, heck we argued about just making g++ the default compiler in the makefile! In the end, we forked the Ice branch specifically to remain pure C, and you know how much interest remained? None. All the folks who argued so hard to keep things in C, ran away despite having a whole branch where they could have what they supposedly wanted.
So, you'll excuse me if the arguments about C++ being too different, too hard, or "not right" fall on deaf ears.
I would love to hear about a revision control system that really does make merging branches simple. I've used cvs and svn, both suck, but svn sucks less. Svn is workable, but only if you constantly keep your branch up-to-date with changes made in the trunk, so when it comes time to merge, it's only your changes going back into the current version.
As far as rewriting the codebases from scratch.... not a bad idea, but first one needs to document how they currently work. That's kindof what I was getting at with my "what makes Smaug, Smaug" question. If you wanted to re-implement Smaug (or SWR, etc) from scratch, you need to know what it is you're going to be re-inventing. AFAIK, none of these codebases fully documents how their components work.