Unicode for prettier code?

It's pretty easy to directly translate a source program into a much prettier unicode representation, especially in a functional setting where math-y features are quite common.

I am not advocating the use of unicode everywhere. Rather, I am pointing out that a text editor could present prettier code to the user while leaving the underlying source code unchanged. This way existing languages and compilers do not need to adapt at all. We can potentially reap the benefits of unicode without the hassle.

Would this make programming more enjoyable?

Demo

Input in the left box will be actively prettified in the right box. Try it out!

Ideally, these two views would be merged into one window so that the programmer only sees the translated text. This would almost certainly be problematic for languages such as Haskell which have meaningful whitespace, but I still think it is a cool idea.

Questions:

Notes:

The unicode characters look best in Chrome with the Consolas font.

This is a very naive implementation, so many undesirable things are possible (e.g. floating subscripts, pi only works when surrounded by spaces, etc.). This would not be the case in a serious implementation!