diff --git a/src/p2/writeup/WriteUp.md b/src/p2/writeup/WriteUp.md deleted file mode 100644 index 904e3dc5462bc599761af76785355a098310d9f5..0000000000000000000000000000000000000000 --- a/src/p2/writeup/WriteUp.md +++ /dev/null @@ -1,102 +0,0 @@ -# Project 2 (uMessage) Write-Up # --------- - -## Project Enjoyment ## -- What was your favorite part of the project? Why? - <pre>TODO</pre> - -- What was your least favorite part of the project? Why? - <pre>TODO</pre> - -- How could the project be improved? Why? - <pre>TODO</pre> - -- Did you enjoy the project? Why or why not? - <pre>TODO</pre> - ------ - -## Experiments ## -Throughout p1 and p2, you have written (or used) several distinct implementations of the Dictionary interface: - - HashTrieMap - - MoveToFrontList - - BinarySearchTree - - AVLTree - - ChainingHashTable - - In this Write-Up, you will compare various aspects of these data structures. This will take a significant amount of - time, and you should not leave it to the last minute. For each experiment, we expect you to: - - Explain how you constructed the inputs to make your conclusions - - Explain why your data supports your conclusions - - Explain your methodology (If we wanted to re-run your experiment, we would be able to.) - - Tell us what inputs you used. If generated in the code, tell us where and how. If from a file, please describe the files and include them in the experiments folder. - - Include your data either directly in the write-up or in the experiments folder - - You should include graphs of the outputs for at least a few of the questions. - You can add a link to an image following the instructions found here: -https://courses.cs.washington.edu/courses/cse332/18wi/handouts/markdown.pdf -An example can also be found at the end of this file. - - We recommend that you keep your "N" (as in "N-gram") constant throughout these experiments. (N = 2 and N = 3 are reasonable.) - - You should probably run multiple trials for each data point to help remove outliers. - - You should not need to wait for hours and hours for your program to run in - order to answer a question. Do use large values for N, but not so large that - you are waiting overnight for your program to run (N=1,000,000 is likely larger - than you need.). - - -### BST vs. AVLTree ### -Construct inputs for BST and AVLTree to demonstrate that an AVL Tree is asymptotically better -than a Binary Search Tree. Comparing the worst case for each structure is fine here. -To do this, we expect you to show trends. You might consider fitting a curve to -your results. Explain your intuition on why your results are what they are. -<pre>TODO</pre> - -### ChainingHashTable ### -Your ChainingHashTable should take as an argument to its constructor the type of "chains" it uses. Determine -which type of chain is (on average, not worst case) best: an MTFList, a BST, or an AVL Tree. Explain your intuition on why -the answer you got makes sense (or doesn't!). -<pre>TODO</pre> - -### Hash Functions ### -Write a new hash function (it doesn't have to be any good, but remember to include the code in your repository). -Compare the runtime of your ChainingHashTable when the hash function is varied. How big of a difference can the -hash function make (on average, not worst case)? (You should keep all other inputs (e.g., the chain type) constant.) Explain your intuition on -why your results are what they are. -<pre>TODO</pre> - -### General Purpose Dictionary ### -Compare BST, AVLTree, ChainingHashTable, and HashTrieMap on alice.txt. Is -there a clear winner? Why or why not? Is the winner surprising to you? -<pre>TODO</pre> - - -### uMessage ### -Use uMessage to test out your implementations. Using N=3, uMessage should take less than a minute to load using -your best algorithms and data structures on a reasonable machine. - - - How are the suggestions uMessage gives with the default corpus? (here we mean spoken.corpus or irc.corpus, not eggs.txt) - <pre>TODO</pre> - - - Now, switch uMessage to use a corpus of YOUR OWN text. To do this, you will need a corpus. - You can use anything you like (Facebook, google talk, e-mails, etc.) We provide - instructions and a script to format Facebook data correctly as we expect it will be the most common - choice. If you are having problems getting data, please come to office hours and ask for help. - Alternatively, you can concatenate a bunch of English papers you've written together to get a corpus - of your writing. PLEASE DO NOT INCLUDE "me.txt" IN YOUR REPOSITORY. WE DO NOT WANT YOUR PRIVATE CONVERSATIONS. - * Follow these instructions to get your Facebook data: https://www.facebook.com/help/212802592074644 - * Run the ParseFBMessages program in the p2.wordsuggestor package. - * Use the output file "me.txt" as the corpus for uMessage. - - - How are the suggestions uMessage gives wth the new corpus? - <pre>TODO</pre> - ------ - -A sample image: - - - -To show you how it is done. - -## Above and Beyond ## -- Did you do any Above and Beyond? Describe exactly what you implemented. - <pre>TODO</pre>