Abstract: A client can operate with a parser or generator for both text (e.g. XML) and binary (e.g. WBXML) mark up languages; the client uses a unique integer value that can be interpreted in an index of elements, attributes and attribute values needed to describe a particular type of mark-up document, the index maps that unique integer value not only (a) to a token associated with predefined element, attribute or attribute value to enable a token based mark up language to be handled but also (b) to a string associated with a predefined element, attribute or attribute value to enable to enable a string based mark up language to be handled. This greatly simplifies the implementation of clients. Further, clients that require parsing of different mark-up languages will require less memory, as they will communicate with only one common API to parse both binary and text mark-up languages. Also, clients will not need to know what the source document is (i.e. text or binary) as the APIs are the same.