Not that octal floating point is commonly (or ever?) used, but I was experimenting with stream manipulators and found that std::oct is ignored when outputting a floating point value to display(just displays in dec). Curious - is there a std method to show fp in other bases?
In programming (or at least in C++) numerical bases other than decimal (as hex or oct) are used only for integral data types as they somehow reflect the binary representation of that number. Floating points numbers work differently so standard C++ doesn't provide ways to handle non decimal I/O with them. Writing your own function to do that shouldn't anyway be much hard if you really need this.
You can represent floating point numbers in other radices, but it is uncommonly rare. It is natural, but it is a bit of a pain, and so oddball that it isn't worth the time and code to do it in the STL.