Nifty Assignments

1
Nifty Assignments

• Shamelessly Borrowed from Nick Parlante at
  Stanford University

2
A Little History

• Began as a small panel discussion at SIGCSE in
  1999
• Friends of Nick Parlante would present cool
  assignments
• Now the 800 pound gorilla of SIGCSE
• original web site
• nifty.stanford.edu/
• AP Central has a similar page

3
Start With the Problem

• Greg Lavender - Start with the Problem
• Motivate what needs to be learned with an
  interesting problem
• Biology teacher I knew started his year with the
  question What is food?
• I think CS actually has it easy because we can
  actually solve interesting problems in our labs.
• students get immediate feedback and can create
  non trivial artifacts

4
What Makes an Assignment Nifty?

• Not the same answer for everyone
• I think determining if a 400 digit number is
  prime is an a very interesting problem
• algorithmically interesting with many approaches
• look at efficiency of solutions
• data representation is interesting
• relevant -gt Encryption techniques such as RSA
• A lot of my students DONT find this as
  interesting as I do

5
What Makes an Assignment Nifty?

• they are fun, playful, interesting
• they are often, but not always visual.
• they are scalable. Top students can run with
  them, others can complete the basics
• they fulfill Astrachans Law
• Owen Astrachan Do not given an assignment
  that computes something that is more easily
  figured out without a computer such as the old
  Fahrenheit / Celsius conversion problem.

6
Example 1 Name Surfer

• From Nick P.
• The Name Surfer
• Social Security Administration popular baby
  names web site.
• www.ssa.gov/OACT/babynames/
• Data on names of children born in US
• Assignment uses list of 1000 most popular names
  by decade stored in a text file

7
Name Surfer

• Students must read in and store the names, search
  the names, and complete a GUI to show the names

Where did all the Ethels go?
8
Name Surfer

• Handout has fairly detailed step by step
  instructions on how to approach the problem
• Components
• modular design
• multiple classes
• using ArrayLists and other classes
• calculations for lines on display
• Scalable
• do simple text based, then add window, then add
  GUI
• dont provide data file, have students create it
  from multiple files

9
Name Surfer

• Use the tool you have created to investigate
  naming trends
• plot grand parents names
• Rock, Trinity, Dwight
• Jose, Mohammed
• Mike and Michael, Dave and David, Matt and
  Matthew
• J, D, M

10
Example 2 Word Ladders

• Proposed by Owen Astrachan
• Begin with two 5 letter words and a list of valid
  5 letter words
• brain, smart
• Change one letter of start word for next word
• New word be in the list of valid words

11
Word Ladder

• smartscart (a type of audio / video
  connector)scantslantplantplaitplainblain
  (an inflammatory swelling or sore) brain
• My word list was derived from a Scrabble word
  list.

12
Word Ladder

• Actually exploring a graph
• the nodes are the words
• connections (or edges or links) exist between 2
  words if they differ by a single letter

13
Word Ladder Assignment

• I present an algorithm to students that does a
  breadth first search of the graph using a queue
  of stacks
• Students must implement the stack and queue
  classes and then implement the algorithm
• Components
• implementing data structures
• implementing algorithms
• comparisons of efficiency

14
Word Ladder Assignment

• How do you find words one letter different?
• do a linear search of all words in the word list
• O(N)
• but the word list is sorted.
• Given a word generate all possible 5 letter
  combinations that are one letter different
• smart -gt amart, bmart, cmart, smarx, smary,
  smarz (125 in all)
• take these 125 words and search the word list for
  each one using a binary search
• This cant be faster can it?

15
Word Ladders

• smart to brain
• Linear search method 0.991 seconds
• Binary search method 0.260 seconds
• How can that be?

16
Word Ladders

• Another extension
• Do a depth first search
• Ladders are much longer
• smart brain, 521 words in 0.140 seconds
• The word list has about 8500 words
• Other possible extensions
• start from both ends and work towards the middle
• find all the connections up front
• dont provide the words in sorted order
• map out all the independent sections, which is
  the biggest?
• which word is closest to the center of the
  largest graph?
• what is the largest ladder that exists?
• other rules from Wikipedia