Compression of SMIL documents

Summary form only given. The World Wide Web Consortium (W3C) standard synchronized multimedia integration language (SMIL) is an HTML-like mark-up language to describe temporal behavior and presentation layout for multimedia objects. SMIL is widely used in today´s Internet. Many video clips send SMIL documents to clients before transmitting video and audio streams. Data compression is a process to reduce the number of bits in a representation of data. It can save storage space and reduce bandwidth requirement. Data compression technologies may be classified into two categories: lossless and lossy. The compression technology that can recover data perfectly from the compressed bits is called lossless compression. The one that cannot is called lossy compression. There are many lossless data compression algorithms, such as Ziv-Lempel methods, that have been developed in recent years. However, SMIL document compression performance with such methods is not very impressive because they do not exploit the special format of the SMIL document. As a result, the rate to compress a SMIL document is almost the same as that to compress English text, i.e., 2-4 bits/char. For better compression, a method based on a new parsing technology is proposed. The proposed compression algorithm contains two major procedures: parsing and coding. Parsing is a procedure to segment a SMIL document into non-overlapping phrases (a string of consecutive characters). Coding is a procedure to assign a codeword to the phrase. Both the phrase and codeword are of variable length