The current specification of java.nio.Charset refers to RFC 2278 as its character encoding naming standard. Unfortunately, this RFC doesn't provide enough guidance on how to name character encodings that are related but not identical. This is a pretty common problem among character standards and comes in various forms:
- coded character sets may be revised over time to add characters or change code points (e.g., the various revisions of JIS X 0208)
- different companies may define different extensions of the same basic coded character set (e.g., the various extensions of JIS X 0208 and its derivative Shift-JIS)
- the same coded character set may be mapped differently to different versions of the Unicode character encoding (e.g., HKSCS)
- a single character encoding may represent an abstraction of several character sets, and needs to be mapped to various concrete Unicode subranges depending on context (e.g., ISCII).

In many cases, the creators of these variants don't bother defining or registering name variants for them. However, applications sometimes need to be able to specify a precise variant of a character encoding, and different variants of a character encoding may need to be supported within the same JRE. A better defined character naming scheme is therefore necessary.