The worst kind of documentation is no documentation. If you want more docs in the world, you need to make it more likely that they will be written. Removing a barrier to doing that (like learning/remembering a markup language) cannot possibly be a bad thing. Many, many OSS projects are one person's small gift to the community; it is ungrateful of us to tell them to suck it up and learn a new tool for our benefit.

Markdown can be just as extensible and rich as reStructuredText given the right syntax for extensions. Don't dismiss it until you see a direct comparison. And even if it's a little worse, see last paragraph.

And finally, it's okay if this project is not for you. Other people will do what they like.

Sphinx is a fundamental piece of the Python community. This is great for the community, but it also means that radically changing Sphinx could be both difficult and harmful.

I think Git and Subversion are an analogous pair of projects to Computer Words and Sphinx. Subversion exists and still works great, but Git introduced some new ideas and a lot of people made the switch because they saw a benefit.

Whether Computer Words has a similar benefit over Sphinx in the long term remains to be seen. My point is that it is still productive to start a new project even when there is already an established one.