Compressing Compound Documents

Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 348 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 349

Electronic documents are commonplace. From PDF files [27] to fax transmissions, including internal raster representations, a number of electronic formats are used to convey text and pictorial information. Documents are present in a wide spectrum of printing systems and are basically represented in vectorial or raster forms. A document in a vector form is composed of a series of primitives, or instructions such as “render this text at that position.” It often comes as a pair of object and instruction, e.g., the text is the object and the instruction was to render it at a particular position. The same happens to graphics. The document is imaged by “drawing” all “vectors” and applying instructions to image a number of objects. Examples are PostScript [29] and PDF files [27], printer languages such as PCL [13], graphics primitives in operating systems, and even ancient presentation protocols such as NAPLPS [2]. In contrast, raster documents are composed of images of the information to be rendered, e.g., a compressed image or a scan of a magazine page. The advantage of raster images is that they are ready to image, as opposed to vector formats that need to be rasterized first.