PS: I've bought Elite but not Star Citizen ;-)
PPS: I've also bought Elite in 1984 and was blown away.
I've done some ambitious things in my engineering career... nothing like Star Citizen in absolute scope but relative to the resources I had they were ambitious. But I always made sure there was some core I could show off before the project was complete.
If nothing else, Agile (by which I mean more the orginal manifesto rather than any specific thing like "scrum") recommends this sort of approach for a reason. The more parts you develop in isolation in your "ambitious" project, the higher the odds are that when it comes time to put them all together, they won't go together. Games have the extra problem that they can both technically fail to go together as expected, but they can also artistically fail to go together as well, making it even riskier to do this in gaming. Many is the game where the obvious fault lines between the various pieces are plainly visible in the final product. (Heck, The Ubisoft Game™ that they keep remaking and just slapping new skins on is all but a game structure designed to make this approach possible, which is basically why they keep remaking it over and over; it's not that they're artistically committed to it, it's that they've probably structured their entire game making organization around it and the rest is just Conway's Law.)
I still see little reason to expect Star Citizen to ever be anything like the hopes and dreams. I still expect that whatever finally comes out of this process will be profoundly disappointing. The process just can't work.
PS: I've introduced Agile in several companies as CTO, no need to convince me ;-)
Excatly, and it will exactly produce the games we currently have.
Even if they keep going to 10 years, they're doomed. To be relevant 10 years from now, SC is probably going to have to switch engines. That will probably mean starting from scratch on many game systems and assets.
There's a reason you don't see many (or any?) successful games that have a 10+ year development time.
Game engines went through a period where they evolved very quickly, and they are still evolving. But the rate of change is slower, like word processors. It starts to make more sense to evolve them rather than replace them.