This talk discusses ways to reduce the size of Unicode text in files and protocols by choosing a compact encoding, optionally combined with general-purpose compression.

Unicode is often perceived to be "too big", and to cause an increase in text size compared to traditional codepages. Concerns are raised especially for systems with limited connection bandwidth, e.g., dial-up or long-range wireless networks, and for computers with small memory sizes, like PDAs and cell phones.

There are several Unicode encodings available with different encoding size characteristics. After an overview of UTF-8/16, SCSU, and BOCU-1, their use in different environments is discussed and compared with general-purpose compression and traditional codepages. Comparison numbers are presented, based on the ICU implementation. Software support for the encoding and compression schemes is discussed.