So, let's talk feature creep. And I'll pick on ST:R specifically because I take full responsibility for its numerous and extensive delays.
ST:R officially began work in August 2002, and we could have released a perfectly playable campaign in summer 2004. The initial FREDding was complete, the initial beta testing was complete, and the campaign would have been easily recognizable to anyone who played the actual release. (In fact, one of these days I want to release the campaign as it was back then, since I have backups of all the missions at various points in time. I'd be interested to hear people's comparisons of it.)
ST:R is probably unusual among feature-creeped campaigns because none of the delay was due to HTL modelling. Instead, I kept pushing for more and more polishing and enhancement. We did several rounds of beta testing, and I rewrote several missions extensively (a few from scratch). This is all in addition to the standard occasional periods of inactivity, not to mention the four month detour to put together version 3.0 of the Port. And finally, when the campaign was essentially DONE in summer 2007, we spent a year and a half on voice acting.
If we had stuck to our original plan and released in summer of 2004, we would have released without the following things:
1) The Exodus, Blood of the Innocents, and Abandon Hope missions
2) The current versions of More than Meets the Eye, Ghosts, Secrets Reborn, and Last Stand
3) Voice acting (except for the sound effects in He Who Rides the Tiger)
4) The two mystery videos
5) The two trailers
6) Phoenix
7) A lot of mission polishing and bugfixing
And I suspect people would have been perfectly happy with it. But I wouldn't have been happy with it, because I would always be thinking of all the cool stuff I wanted to add.
The other question is whether we could have released in stages: the first draft, plus an enhanced draft, plus a final "director's cut" complete with voice acting. Instead, we waited until the director's cut was finished and released that. The reason I didn't do it that way was that I suspected many (perhaps not most, but a significant number) wouldn't bother playing the updated version if they played the original. Plus, the final release made much more of a "wow" impression than the 2004 release would have.
And then there's the danger of so many delays that the developers lose interest and the campaign essentially dies. Sometimes raw perseverance wins over feature creep (in the case of ST:R); sometimes feature creep wins (in the case of TVWP). TVWP was essentially abandoned, due to Lightspeed's disappearance and my preoccupation with ST:R, until Admiral Nelson came along and whipped it into shape. Even so, without a new project leader, TVWP is going to collapse once it staggers across the Chapter 1 finish line.
It's often hard to tell from the outside whether a campaign will conquer feature creep or not. And sometimes, even from the inside, the campaign leaders refuse to admit it to themselves.
