Another perfectly acceptable reason is education.

By developing my own fairly complete (albeit simple) framework, I can better understand how things work at a lower level. This will, among other things, help me to make a good decision about what framework to use for more complex applications.

That’s the nice thing about wsgi, you can take bits and pieces from all over the place and glue them together with relatively little effort.

Unless you’re creating an application which fits perfectly into the model of a given framework, it will probably take you less time to make your own framwork which matches your needs exactly. Not from scratch, but from existing standalone pieces. Otherwise, you’ll end up spending a large chunk of time working around quirks in the framwork you’ve chosen, or even worse, you’ll find yourself locked into a given framework even when it proves to not fit as well as another would.

This is especially bad when you run into issues of scalability. If you make your application out of loosely coupled components, you can easily swap out the chunk that doesn’t scale well. With some framworks, this (swapping out bits and pieces) can be difficult to do.

In my next post on this subject I’ll be going into more detail of using preexisting components to build you own application, so you’ll see I’m not advocating building a whole application at such a low level. This starting post was just to give an idea of what will lie below everything.