Screenshots

	The normal startup, using some of its shell-like builtins (ls) and running a program with an error in it via the magic run command. The magic command system can be extended by the user without needing to modify any of IPython's code, examples are provided. The error is shown with a verbose, colored traceback which provides better context and easier to read information than Python's default.
	Here are some of its time-saving features: TAB completion, automatic parentheses and quotes for function calls and information about defined variables, including details on any available object. You can also view the function definition line (for knowing how to call a function), constructor information, the full source code for an object or directly view the file where it is defined.
	You can  change the style of exception reporting to a shorter one, and when developing code, you can load the python debugger pdb at any uncaught exception. The pdb debugger can walk up and down the stack, showing you details on the status of your program when it crashed (all data is still available). This can make for a very efficient debugging cycle without the need for code modification.
	IPython supports an easy system for changing configurations based on profiles. The included 'tutorial' profile allows pasting of code which has >>> and ... at the beginning (such as is found in the standard Python tutorial). This is done via a mechanism for extending the processing of the user's input, which transforms on the fly what would otherwise be invalid python code. A similar profile is provided for doing calculations with physical quantities with units. These can be used as examples to adapt IPython with your own syntax extensions for particular problems.
	It can easily interface with other systems which support pipes-based communications. The provided 'numeric' profile loads some utility functions for conveniently using the Gnuplot plotting package, and can be used as an example for similar tasks.
	IPython can show syntax-highlighted source code for objects whose source is available.  This screenshot was generated by typing at the IPython prompt: import code code??