There have been quite a few times when I’ve worked on building a portion of a tool, gotten it near utility, but haven’t seen it all the way through to being reusable. I think this probably follows a Pareto effort for developing a polished product. It’s often a much smaller part of the effort to build the core functionality than it is to make it actually usable. And this comes in waves.
The first 20% is something that just proves it works—like printing a log. Then there’s the 80% of making it something I can personally use, such as wrapping it in an app.
That stage is the 20% of it becoming a real project. Then there’s the 80% of setting it up in VCS (gotta come up with a fun name) and taking enough notes so I can come back to it in a year.
Once it’s a dogfoodable project, I might be 20% of the way to an early demo. A demo probably needs some sort of navigation, things labeled in ways that make sense, and a reasonable GUI.
A demo is 20% of a beta. A beta may break, but it still needs to be fairly complete. At this point, you have to add a lot of the big-boy stuff like logging, crash reporting, analytics, and billing.
A beta product is only 20% of a launch. Then there’s all the planning, website, content.
By breaking it down this way, I obscure the total hassle of creating something people might actually like and just keep doing the next 20% of a larger step.