There are too many rich text formats. Many are undocumented, proprietary, or otherwise not well-supported. They're hard to revision control, which means harder to sync across computers and among groups of people.
The hardest thing about this project, which I love, is going to be saying "no" to almost every single feature request people propose. I'll be interested to see how well the authors can hold the line.
I'm also very interested in the software development methodology the authors use. If they're seriously about longevity, they'd be smart to take as few dependencies on external libraries/systems as possible and use clean, layered architecture. I'd probably do the lowest levels (the sync and file storage, etc) in close to pure C and put a graphical front-end on top of it, similar to what Netflix does with their high-performance media libraries (streaming and decode) to share code across iOS and Android (via NDK). You want something super-portable that's easy to slap a new front-end on in 10-15 years when things have changed, which IMO pretty much rules out any language except C and maybe Java or .NET.