It's somehow interesting to me how often this comes up on HN. It's also a problem I tend to have. But if you had asked me 10 years ago based on what I'd casually heard outside of tech, it was much more common for me to encounter and worry about the exact opposite failure case in creative projects: not the person who has 30 half-finished novels, but the person who's spent 30 years writing their Masterpiece Novel that is probably not going to be a masterpiece, but nobody can convince them that they really should be trying out other things instead of sticking with this one project to the bitter end. Basically, not enough experimentation and bouncing around between projects, and too much unwavering fixation on one Big Project that they've tied their ego to.
Curious what the distribution of those two opposite problems is like! Among creative types, it seems that the "stuck with it too long" problem is more prevalent in certain areas, like writing or filmmaking, but not as prevalent among hackers. I suppose one compensation for it being so common for hackers to fall into the "100 unfinished projects" trap is that it's less common to find hackers who fall into the trap of, "I stuck with the first idea I ever got for 20 years because I thought it was my One Big Idea".
It's partially because the technology industry moves, while the film or book industry doesn't.
Film and novels derive their quality from their plots. Plots transcend time. Although the tools involved in creating a movie or book may change (improved camera systems), the tools don't impact the plot much. Therefore, a book written thirty years ago can have the same quality as a book written today.
The quality of an application is unable to transcend time. Sure, the idea of creating a website to connect people might, but the quality of an application isn't usually based on the idea, it's based on the implementation. Implementations cannot transcend time; that's why technological masterpieces don't exist. A programmer who locks himself in his room and starts working on "the next Facebook" now won't be very successful in thirty years: when he launches, the whole technological landscape will have changed.
Programmers rarely work on something for too long, because they understand that their products are dependent on technology, which changes over time. What is amazing today is not amazing ten years from now.
I think the person with 30 unfinished books and the person spending 30 years on their novel are really just flavors of the same thing - inability to finish a job. They're both re-starting and revising over and over again. I would guess that if the 30-year novel ever gets completed, it probably bears little resemblance to the original draft. There's probably various reasons that could apply to either case - perfectionism, lack of attention span, unable to commit on decisions, self doubt, etc.
Curious what the distribution of those two opposite problems is like! Among creative types, it seems that the "stuck with it too long" problem is more prevalent in certain areas, like writing or filmmaking, but not as prevalent among hackers. I suppose one compensation for it being so common for hackers to fall into the "100 unfinished projects" trap is that it's less common to find hackers who fall into the trap of, "I stuck with the first idea I ever got for 20 years because I thought it was my One Big Idea".