.DOC File Format – Stories, fields, pictures and objects

Word 97–2003 document

The whole document is one sequence of character positions (CPs). The FIB’s character counts divide it into stories, which come one after the other.

Stories

  1. ccpTextthe main documentstory 1
  2. ccpFtnfootnote textstory 2
  3. ccpHddheaders and footersstory 3
  4. ccpAtncommentsstory 4
  5. ccpEdnendnote textstory 5
  6. ccpTxbxtext boxes in the main documentstory 6
  7. ccpHdrTxbxtext boxes in headers and footersstory 7
The stories in CP order.

A story of 0 characters is not present. A document with several stories ends with an extra paragraph mark, after the last story.

Footnotes and endnotes

PlcffndRef holds the CPs of the footnote reference marks in the main document. PlcffndTxt holds the CPs where each footnote’s text starts in the footnote story, relative to the start of that story. PlcfendRef and PlcfendTxt are the same for endnotes.

Headers and footers

PlcfHdd divides the header story into pieces with no data elements: the first six are separators for footnotes and endnotes, then six for each section, in the order even header, odd header, even footer, odd footer, first header, first footer. A piece of length zero means that header or footer is absent.

Comments

PlcfandRef gives the CP of each comment’s reference and an author record for it. PlcfandTxt locates each comment’s text in the comment story.

Fields

A field is written in the text as the characters 0x13, the field code, 0x14, the result, 0x15. A PlcfFld for each story gives the CP and a two-byte Fld of each of those characters: its character, and at the begin character the field’s type (88 HYPERLINK, for example).

Bookmarks

SttbfBkmk names the bookmarks in a string table (see Styles and fonts for the shape of a string table). PlcfBkf and PlcfBkl give the start and end CP of each.

Inline pictures

An inline picture is the character 0x01 with sprmCFSpec and sprmCPicLocation. The operand is an offset in the Data stream where a PICF structure starts.

  1. lcbthe size of the whole picture entry4
  2. cbHeaderthe size of this header, 682
  3. mfpthe picture kind: 0x64 or 0x66 Office Art, 8 metafile, 0x62 a linked TIFF, 0x63 a linked BMP8
  4. Header restcrop, scaling and borders54
  5. Picture dataan Office Art inline shape, whose image is a BLIPvaries
The start of a PICF.

Drawn objects

Shapes that float over the text are the character 0x08 with sprmCFSpec. PlcSpaMom gives each one’s position and shape identifier; the Office Art data is at DggInfo in the table stream.

Embedded objects

An embedded OLE object is a storage inside ObjectPool, named from the sprmCPicLocation value with an underscore in front. Its CompObj stream gives its type name. The object’s own streams are kept as they are, ready for the program that made it.