.PDF

Portable Document Format

Introduction

PDF was created by Adobe in 1993 to show a document exactly as intended on any computer or printer. It became an open standard as ISO 32000: PDF 1.7 in 2008 and PDF 2.0 in 2017. A PDF holds fixed pages of text, vector graphics and images, with the fonts embedded, and can also carry links, forms, annotations, attachments and digital signatures.

Software

Structure

A PDF file starts with a header line such as %PDF-1.7, usually followed by a comment of a few bytes above 127 so that transfer programs treat the file as binary. The body is a series of numbered objects, each written as 12 0 obj … endobj: dictionaries, arrays, numbers, strings, names and streams of compressed data. Pages, fonts and images are all objects that refer to each other by number.

The cross-reference table, starting with xref, gives the byte offset of every object so a reader can jump straight to it. The trailer names the root object, the document catalog, and the file ends with startxref, the offset of the table, and %%EOF. An edited PDF can append a new body, table and trailer rather than rewrite the file. Since PDF 1.5 the table and many objects may themselves be held in compressed streams.

  1. 0Header %PDF-1.7 and a binary commenta line or two
  2. …Body numbered objects: catalog, pages, fonts, images, content streamsvaries
  3. …Cross-reference table the byte offset of every object20 bytes per object
  4. …Trailer root object, size, startxref, %%EOFa few lines
The four parts of a PDF file as first written.