Comments on assignment due week 4, Nov 1: most_frequent

Some students had most_frequent return the sorted list, in addition to
printing it.  The problem statement did not ask for this, but did not
forbid it either.  In fact, it's usually good for functions to return
something useful.

Some students noticed the problem statement said "letters", not
"characters", so they removed non-letter characters and converted all
letters to the same case (upper or lower).  I didn't expect this; it
shows scrupulous attention to the exact wording of the problem
statement.

Some students chose particularly informative strings to confirm their solution
was correct, before using it to check samples of natural language.  In this string,
a has frequency 1, b has frequency 2 etc.

 test = 'abbcccddddeeeee' # etc. ...

Some students created additional alternative solutions, besides the ones in
my most_frequent.py:

- Instead of reversing the tuples derived from the histogram
dictionary and sorting those, build a new dictionary where they keys
are the frequencies and the values are lists of all the letters that
have the same frequency, for example:

 collect = dict() # create empty dictionary
 for letter in freqs: # freqs is dict returned by histogram
   freq = freqs[letter]
     if freq not in collect:
       collect[freq] = [letter] # start list with one element
     else:
       collect[freq].append(letter) # add letter to list

-  Instead of deriving a list of tuples from the histogram dictionary by d.items()
 and then reversing them, build the list of reversed tuples directly from the dictionary:

   inverted = []
   for letter in freqs:
     inverted.append((freqs[letter], letter))

- Instead of using ts.sort(...) to mutate the same list of tuples, use
  sorted(ts, ...) to return a new list

- Instead of using Downey's histogram function to build a dictionary, use
  the string count method as in s.count(c) to count each character c in string s, 
  and build a list of tuples (freq,c).  Then sort that list and remove the duplicates.

- Instead of using a dictionary and then sorting tuples, use the
 Counter class from the collections module (new in Python 2.7), which
 does all the work for you: its constructor creates the frequency
 histogram, and its most_frequent method does the sor ting.

- etc. ...

I am impressed by the number of useful variations, even in this simple exercise!