Compression Wars

by Ken Hancock — [email protected] — khancock on AOL

With all the recent hubbub regarding the various file compressors, I thought it was well past time for a review and comparison of the leaders: StuffIt, Compactor (now Compact Pro), and DiskDoubler.

Before I begin, let’s lay down the foundation used for the timing tests:

All tests were done on a Macintosh SE/30 with 8 MB memory, running 6.0.7 under Finder on a Wren IV drive. The only INIT running was Desktop Manager (no, I wasn’t going to wait for my desktop to be rebuilt). Three different tests were performed on each application and each compression mode, each trying to reflect different real-world situations:

• "Binaries" – a 1,857,129 byte folder containing four applications: Microsoft Word, HyperCard, Compactor, and StuffIt Deluxe 1.0.
• "Graphics" – a 1,868,118 byte folder containing eight PICT files used for backdrops. Mostly scanned images.
• "Text" – a 227,521 byte folder containing 60 TEXT files of varying lengths.

I tested six different applications: StuffIt 1.5.1, StuffIt Classic 1.6, StuffIt Deluxe 1.0, StuffIt Deluxe 2.0, Compactor 1.21, and DiskDoubler 3.1. In each case, the times were based on the time to compress an entire folder from the respective application in a one-step process. For DiskDoubler, this meant using its "Combine" feature since DiskDoubler is most commonly used to compress single files. Since DiskDoubler combines and then compresses, only one compression mode is used throughout all the files. For the DiskDoubler "Smallest" mode test, smaller files might be possible by first compressing the folder and then combining the folder. All the other applications compress on a file-by-file basis.

Please note that these tests do not take into account any interface issues or shortcuts that one might use. In each case, I only recorded the actual time spent by the Mac in compressing or decompressing the file. You should also be aware that there are more variables that were beyond the scope of these tests, such as still more different file types and how the programs worked with smaller files. Working on a single small database file, for instance, might produce somewhat different results. There is simply no way to test all the possibilities, sorry.

So, who won? Good question. As with everything, it depends on the game being played. I judged the applications on two criteria:

1. Absolute compression: which program compressed the files the most.
2. Best Efficiency: which program yielded the best compression/minute.
   

In the case of a tie, the application that compressed the fastest garnered the top spot. All compression percentages are carried out to two decimal places since different applications calculated/rounded the values differently. "Efficiency" was calculated by taking the "% saved" and dividing by the number of minutes it took to compress the folder. For those who wish to see the gory details, take a look at the enclosed Excel spreadsheet.

Since I tested four separate versions of StuffIt and a total of 12 different modes/versions altogether, I’ve summarized the results below, with the following condition:

All StuffIt Deluxe 1.0 trials have been removed as well as all timings with StuffIt Deluxe/Classic "Better" mode. In almost all of the trial modes, "Best Guess" provided identical compression to "Better" mode, but at a faster speed. (One questions why the "Better" mode is even included.)

The results (a note for the less mathematically inclined – the larger numbers are better):