CSCI 587 - Solutions to Assignment 2 Text Compression

1. The dictionary used by the Unix system command spell is in the file /usr/dict/words. Write a small C program that will calculate the frequencies of each character.

   #include 
   
   
   main() {
      int c;
      int charCount[256];
      int i;
   
      for(i=0; i < 256; ++i) charCount[i] = 0;
   
      while((c = getchar())!= EOF)
          ++charCount[c];
   
      for(i=0; i < 256; ++i)
          if(charCount[i] != 0) printf("%15d\t%c\n", charCount[i], (char)i);
   }
   

   • What major differences do you see from the frequencies presented in class?

     No spaces
     

   • Why is this text not a good sample text?

     It is just a list of words, not a good representative example of
     an English  text.
     

2. Using the Huffman coding encode the phrase "USC wins".

   00110 0110 10101 111 001011 1001 0100 0110
   

3. Compare the performance of compress, compact, and gzip calculating the compression ratio for each on /usr/dict/words. Also record the time required using the command "time."
   • To do this create a symbolic link to the dictionary from your home directory with the command
     ln -s /usr/dict/words words
   • Then run each of the commands on the file using time using the elapsed time.
     time compact words
     Use the manual ( man 1 time ) to interpret the output.
     u = user time, s = system time, then total or elapsed time.
   • Check the size using the command "wc"
   • You will need to uncompact, uncompress, and guzip to get back to the original file

   What I got was:
   compress    102727 bytes    1.91 seconds
   compact     111646 bytes    5.43 seconds
   gzip         79269 bytes    5.76 seconds
   Note these are actual sizes. 
   Compression ratios are calculated by 
       CR = (OriginalSize - NewSize)/OriginalSize
   

4. Find a spelling/grammar checker on a PC or Mac.
   • Which wordprocessor are you using?
   • What is the grammar checker's evaluation of
     An hoarse is one thee gulf curse.
     Out the window, the bird flew.
   • How does the spelling checker respond to
     fastly, greenly, et cetera (and I mean the phrase)

   I didn't do this one.
   

5. Extra Credit 3 points A digram is a two character sequence.
   • Using /usr/dict/words calculate a static model for digram compression.
   • For the top ten digrams compute a Huffman code