I've had the same problem. Not just programming projects but also blog posts, articles, music tracks, and even a novel that I want to launch, but somehow, seem stuck at 90% complete.

I was actually thinking of launching a platform to help people get the last 10% done. Often, people have the entire thing figured out but lack the confidence or the expertise to go live. Perhaps if an expert was to look over the project when it was about 90% done, you would feel more confident about launching.

Either way, the demands of someone looking for help before launching, vs. someone who wants something built from scratch are very different. I thought a dedicated platform for the former would help.

That's actually a really good idea. A lot of times people have big dreams for projects they really enjoy, but little knowledge and expertise in those areas, and so they aren't confident enough to see it through to the end. I myself have a website trading business that is in this state.

That's great, but I think this is even more accomplished than a majority of developers. The majority of developers have a wild or fun sounding idea and pursue it for a few days (guilty of this myself). Most likely, the "fun" part only constitutes a minority of overall functionality, and I find it's generally the non-fun stuff that demotivates people to not finish (e.g., billing, setting up a web site, user management, etc.) The key has to be to either determine if those things are critical, or somehow fit them in before you're left with nothing other than a pile of non-fun stuff.

My editor ozette is like that (http://www.github.com/marssaxman/ozette/). I wouldn't call it "done", or suggest that anyone else use it, because there are a dozen unresolved bugs and maybe half a dozen unfinished or unimplemented features that seem important - but it's done enough that I spend all day every day using it to make a living. I fix bugs when they annoy me enough to justify the time spent, and I add features when I have time and inclination in my off hours. I suppose I assume that it will be "done" when I've fixed all the known bugs and implemented all the features I had intended before I began working on it, but it seems likely that I could spend my entire career using this tool and still manage to retire before "finishing" it, which suggests that "done" is not a very useful idea here.