480 Python Projected Miscellany
Lect 13, 11-15-05

Beginning of in-lecture exercise:

1. What would you NOW type at the >>> prompt below to see 
   lect13_1's spam's value of 40?
   (after the 

   >>> from lect13_1 import spam

   )

2. What would you NOW type at the >>> prompt below to see
   lect13_1's spam's value of 40?
   
   (after typing, in a new python session,

   >>> import lect13_1

   )

*   recall: built-in function named dir

    returns a LIST of the attributes of the argument object

    >>> import lect13_1
    >>> dir(lect13_1)

*   a name that begins and ends with two underscores
    (e.g. __name__, __doc__)

    ...is a system-defined name in Python.

    (ref: Section 2.3.2, Python Reference Manual,
    docs/python/org/ref/ref.html, follow the link to 2.3.2)

    *   these are names defined by the interpreter and its
        implementation (including the standard library);

    *   "applications should not expect ot define additional
        names using this convention"

*   __doc__ is set to a string (if you've put one) that is the
    FIRST thing in a module, a def, a class, etc. --- that is then
    stored in the __doc__ attribute of the corresponding object.

    (such a string is called a docstring...)

*   the PyDoc tool makes for "nicer" reports using docstrings
    and other info...

    ONE way to use it: is with the built-in help function

    help(object) --- invokes PyDoc to generate a UNIX-man-page-like
                     textual report on that object

*   LOOKS like Python MAY be lacking in the kind of true
    information-hiding capabilities you may be familiar with
    in C++, Java, Ada, etc.;

    BUT there ARE ways to discourage users from seeing/mucking with
    SOME things;

    *   1st way: IN a module, if you begin a name with a single
        underscore, it isn't copied out when a user uses from *

        (it IS imported using import, however)

        *   hmm... help(that_module) doesn't include such names,
            either...

        *   intended to "minimize namespace pollution"... [Learning
            Python, p. 279, Ch. 18]

    *   2nd way: define a list named __all__ in the module;

        ...then ONLY those names in the __all__ list will be
        copied over using from *.

        (STILL get 'em all with import...)

    *   NOTE: Python looks for an __all__ list 1st --- if
        it finds it (when from* has been done), only the
        names in that list are copied out.

        OTHERWISE, from* copies out all but the _blah names

    *   (yes, it behaves oddly, IMHO, with help and __all__...)

*   __name__ and __main__

    ...lets a module be conveniently used for command-line or import use

    *   point in fact: when you import a module, __name__ for that module
        is set to be the module's name as known by the client

    *   BUT --- if a file is run as a "top-level" program file --- __name__
        is set to the string '__main__'

*   then: intro to OBJECTS (but didn't add to this projected file for
    that; see lect13_7.py, lect13_8.py for examples used to spur
    discussion.)