aboutsummaryrefslogtreecommitdiffstats
path: root/src/documentation/xdocs/hdf/docoverview.xml
blob: fe864bc2d8fa2936a6c2d18e4c4fb83daa1187d3 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE document PUBLIC "-//APACHE//DTD Documentation V1.0//EN" "../dtd/document-v10.dtd">

<document>
 <header>
  <title>HDF</title>
  <subtitle>Word file format</subtitle>
  <authors>
   <person name="S. Ryan Ackley" email="sackley@cfl.rr.com"/>
  </authors>
 </header>

 <body>
  <s1 title="The Word 97 File Format in semi-plain English">

   <p>The purpose of this document is to give a brief high level overview of the
      HDF document format. This document does not go into in-depth technical
      detail and is only meant as a supplement to the Microsoft Word 97 Binary
      File Format freely available at <link href="http://wotsit.org">Wotsit.org</link>.</p>
   <p>The OLE file format is not discussed in this document. It is assumed that
      the reader has a working knowledge of the POIFS API. </p>

   <s2 title="Word file structure">
    <p>A Word file is made up of the document text and data structures
       containing formatting information about the text. Of course, this is a
       very simplified illustration. There are fields and macros and other
       things that have not been considered. At this stage, HDF is mainly
       concerned with formatted text.</p>
   </s2>
   <s2 title="Reading Word files">
    <p>The entry point for HDF's reading of a Word file is the File Information
       Block (FIB). This structure is the entry point for the locations and size
       of a document's text and data structures. The FIB is located at the
       beginning of the main stream.</p>
    <s3 title="Text">
     <p>The document's text is also located in the main stream. Its starting
        location is given as FIB.fcMin and its length is given in bytes by
        FIB.ccpText. These two values are not very useful in getting the text
        because of unicode. There may be unicode text intermingled with ASCII
        text. That brings us to the piece table.</p>
     <p>The piece table is used to divide the text into non-unicode and unicode
        pieces. The size and offset are given in FIB.fcClx and FIB.lcbClx
        respectively. The piece table may contain Property Modifiers (prm).
        These are for complex(fast-saved) files and are skipped. Each text piece
        contains offsets in the main stream that contain text for that piece.
        If the piece uses unicode, the file offset is masked with a certain bit.
        Then you have to unmask the bit and divide by 2 to get the real file
        offset. </p>
    </s3>
    <s3 title="Text Formatting">
     <s4 title="Stylesheet">
      <p>All text formatting is based on styles contained in the StyleSheet.
         The StyleSheet is a data structure containing among other things, style
         descriptions. Each style description can contain a paragraph style and
         a character style or simply a character style. Each style description
         is stored in a compressed version on file. Basically these are deltas
         from another style.</p>
      <p>Eventually, you have to chain back to the nil style which is an
         imaginary style with certain implied values.</p>
     </s4>
     <s4 title="Paragraph and Character styles">
      <p>Paragraph and Character formatting properties for a document's text are
         stored on file as deltas from some base style in the Stylesheet. The
         deltas are used to create a complete uncompressed style in memory.</p>
      <p>Uncompressed paragraph styles are represented by the Pargraph
         Properties(PAP) data structure. Uncompressed character styles are
         represented by the Character Properties(CHP) data structure. The styles
         for the document text are stored in compressed format in the
         corresponding Formatted Disk Pages (FKP). A compressed PAP is referred
         to as a PAPX and a compressed CHP is a CHPX. The FKP locations are
         stored in the bin table. There are seperate bin tables for CHPXs and
         PAPXs. The bin tables' locations and sizes are stored in the FIB.</p>
      <p>A FKP is a 512 byte OLE page. It contains the offsets of the beginning
         and end of each paragraph/character run in the main stream and the
         compressed properties for that interval. The compessed PAPX is based on
         its base style in the StyleSheet. The compressed CHPX is based on the
         enclosing paragraph's base style in the Stylesheet.</p>
     </s4>
     <s4 title="Uncompressing styles and other data structures">
      <p>All compressed properties(CHPX, PAPX, SEPX) contain a grpprl. A grpprl
         is an array of sprms. A sprm defines a delta from some base property.
         There is a table of possible sprms in the Word 97 spec. Each sprm is a
         two byte operand followed by a parameter. The parameter size depends on
         the sprm. Each sprm describes an operation that should be performed on
         the base style. After every sprm in the grpprl is performed on the base
         style you will have the style for the paragraph, character run,
         section, etc.</p>
     </s4>
    </s3>
   </s2>
  </s1>
 </body>
</document>