File:Targzip.svg

Schema of tar-creation and compressing via gzip. In this diagram, the entropy of data corresponds with how much of the data can fit in a square of the same width and height: the more that can fit, the more it has been compressed and the higher the entropy. The files are represented as circles and the gzip stream is represented as a square to show the change in entropy.

The tar container data occupies space equal to the area of the green rectangle (shown in the middle), and the file data occupies space equal to the area to the blue circles. Mathematically, they are in a ratio approximately 25:39, respectively. 64 blocks are shown in the gzip stream: the 25 green blocks are compressed tar container data and the 39 blue blocks are compressed file data. In real-world situations, certain data would be difficult to compress, such as a JPG image, and in the gzip stream, it would consume several times more data than the tar container data as tar data is relatively simple to compress. Other types of data, such as those with unusually long run-lengths, might be easier to compress than tar data.