gcrevell / cs3411prog3 Goto Github PK

                                                                            |

When encoding the encoder.c and encoder files, using gzip and our own encoder, the following results are obtained:

file | No encoding | gzip | ratio ------------+---------------+---------------+--------------- encoder | 13180 | 2512 | 0.1905918058 ------------+---------------+---------------+--------------- encoder.c | 5684 | 1653 | 0.2908163265

file | No encoding | encoder | ratio ------------+---------------+---------------+--------------- encoder | 13180 | 14504 | 1.1004552352 ------------+---------------+---------------+--------------- encoder.c | 5684 | 3376 | 0.593947924

While using encoder on either file didn't match even close with the results of gzip, encoder had a much larger improvment on encoder.c over encoder. I believe the reason for this is with the use of 15 frequent characters. When encoding the english letters of code, many characters repeat and finding these characters in the input is easy. With the encoder object file however, the character values are more or less "random". While not truely random of course, the values are much more spread out, reducing the number of repeated values. Because non repeated characters, which are 8 bits, require an encoding into 9 bits, the file size actually increases.

This leads to the conclusion that while this program is great for encoding text files, its usefullness is less appearent with other types of files.