PDF Specifications
Portable Document Format, PDF, is a file format similar to Postscript. It enables users to
exchange and view electronic documents independent of the platform on which they were created.
This allows any user with Adobe Acrobat Reader to view a PDF document.
The PDF implementation of an Employee Directory was created using basic structures of the
PDF format. The following is an overview of the structures used to create the document.
Basic Document Structure
The PDF document is defined in segments called objects. Each object has a number assigned
to it so that it can be referred to by other objects. The objects do not have to appear in chronological
order in the document, but is much easier to follow when they are.
The first part of a PDF document is the Catalog. The Catalog is the root object and contains
references to the pages and outline contained in the document. The reference is a pointer to another
object that defines the pages and outlines. Each page can be represented by with it’s own object. The
Catalog references an object that contains the number of pages and the object number of each page.
The reference to the outline contains the number of outlines and their object numbers. If no outline
is used, the number of outlines is set to zero.
The Page object defines how the page will appear. The fonts that will be used, which are
defined in their own object, are referenced, as well as the parent object, the ProcSet, MediaBox, and
the object that contains the actual information. These are explained in more detail later. The
information that the page contains is defined in a stream. The length of this stream in characters must
be know, and is referenced in another object. If the length of the stream is known before the PDF
document is written, the length of the stream can be written into the document without the need of
another object.
After all of the objects have been written, the xref is made. This is a listing of the objects and
position that the object starts in the file in bytes. It is possible for PDF readers to jump to a particular
page in a document. The xref tells the reader where in the file to go to for the page without having
to read through or download the whole document. After the xref is the trailer. This tells the reader
the number of objects (including one free object) are in the document and the object number of the
root object. Finally, the total length of the file up to the xref is written, followed by an end of file
marker.
Syntax
The first line of the PDF document is the version of the PDF format being used. For simple
documents only containing text, version 1.0 can be used.
The first line of an object starts with the object number, followed by the generation number,
and finally the keyword obj. The generation number is generally set to zero, and obj lets the reader
know that the following information is an object. The content of the object is contained within
opening << and closing >>. The object is closed by the endobj keyword.
The first object is the Catalog. The object is defined as a Catalog using the Type keyword.
The Pages keyword is used to reference the object that contains information about the page objects.
If an outline is used, the Outlines keyword references the object that contains the outline objects. The
following is part of a PDF file showing the version and Catalog object.
%PDF-1.0
1 0 obj
<<
/Type /Catalog
/Pages 3 0 R
/Outlines 2 0 R
>>
endobj
2 0 obj
<<
/Type /Outlines
/Count 0
>>
endobj
3 0 obj
<<
/Type /Pages
/Count 1
/Kids [4 0 R]
>>
endobj
The page information is contained in object three and the outline information is contained in
object two. Object two is of type outline and /Count 0, shows that it has no information defined for
the outline. Object three is of type pages, so it defines the page objects for the document. The
number of pages is defined using the count keyword. /Kids references the objects that contain the
page format information. If there were two pages, /Kids [4 0 R 25 0 R] would be used where object
25 is an object containing the page format information.
4 0 obj
<<
/Type /Page
/Parent 3 0 R
/Resources << /Font << /F1 7 0 R >> /ProcSet 6 0 R >>
/MediaBox [0 0 612 792]
/Contents 5 0 R
>>
endobj
Object four is of type page and defines the page format. It references it’s parent object, which
is the object that calls it, and the object that contains the page information. The /Resources tells the
main format of the page to be used. The objects that define the fonts and ProcSets are defined inside
the opening << and closing >>. The MediaBox defines the size of the page in points in a rectangular
coordinate system. The lower left corner is (0,0) and in this example the upper right corner is
(612,792). One inch is equal to 72 points. Positioning of characters on the page is done using this
coordinate system. /Contents references the object, 5, that contains the page information.
5 0 obj
<< /Length 44 >>
stream
BT
/F1 24 Tf
100 100 Td (Hello World) Tj
ET
endstream
endobj
Object five is the main object that defines what appears on the page. Since the information
is contained in the stream, the length of the stream has to be known. In this example, the length if
the stream is known to be 44 bytes. This can be verified by counting the characters, including an end
of line character, starting with the BT and ending with the ET. In general, if the length of the stream
is not know, an object that appears after the page object must be referenced that contains the length
of the stream. This is accomplished by changing << /Length 44>> to /Length X 0 R, where X is the
number of the object that contains the length of the stream. This object is very simple:
X 0 obj
44
endobj
After the length declaration, the information stream begins. The keywords steam and BT start the
stream. Likewise, ET and endstream end it. Before text can be written, it has to have a font
referenced. The name of the font and the point size for the text is declared, followed by the keyword
Tf. Tf sets the font name and size.
This way of defining a page makes it necessary to have three objects per page. These objects
are of types Pages, Page, and the length object.
The next line contains the text and it’s placement on the page. If this is the first text to be
written to the page, the current x and y position of the page are both zero. After text has been
written, the current x and y positions take on the value that x and y were increased or decreased by.
The Td keyword is used to move x and y positions. The difference in the x position, the difference
in the y position, preceeds Td to move the text position. So to move to (50,50), before any other text
has been placed, “50 50 Td” would be used. Now to move to (100, 25), “50 -25 Td” is used. The
length of the text does not change the x and y positions.
The text to be written to the page follows the Td and is contained within parenthesis. The
Tj keyword tells the reader to show the text. If the text has hanging parenthesis, this will cause the
page containing the hanging paren to not load.
6 0 obj
[/PDF /Text]
endobj
7 0 obj
<<
/Type /Font
/Subtype /Type1
/Name /F1
/BaseFont /Helvetica
>>
endobj
Object six defines the ProcSet to be used. For a simple document that only contains text,
/PDF and /Text are the only ProcSet’s needed. There are other ProcSet’s that need to be used of
images are to be included.
Object seven is of type font which defines the font to be used. The Sybtype keyword defines
the type of font to be used. Type1 fonts are special-purpose PostScript language programs used for
defining fonts. The name of the font, as it is to be used in the PDF document, is defined using the
name keyword. The name of the font has the format /FX, where X is an integer. Multiple fonts can
be defined in this way. There are 14 Type1 fonts that are guaranteed to be present in the reader.
These fonts do not need any attributes defined for them. The BaseFont can be one of these 14 fonts
or a special font can be defined.
xref
08
0000000000 65535 f
0000000009 00000 n
0000000074 00000 n
0000000120 00000 n
0000000179 00000 n
0000000322 00000 n
0000000415 00000 n
0000000445 00000 n
After all of the objects have been written, the xref is made. Each object in a PDF file is
referenced in the xref table at the end of the file. This table begins with the xref keyword. The next
line begins with a number representing the subsection of the xref table. For simple PDF files, this
number is usually 0. The following number is the number of objects used in the file, plus one for the
next free object to be used. This free object is defined on the following line. The remaining lines are
objects that are in use. The first set of numbers is the number of bytes into the file where the object
starts. The second set of numbers is the generation number of the object. This number is originally
set to 00000, (unless it is a free object, where it is set to 65535). Each time that the object is updated
or changed, the generation number is increased by 1, until the maximum number, 65535, is reached.
The generation number is followed by either an f, for free object, or an n, for a used object. There are
seven objects in use, plus the one free object for a total of 8, as seen on the second line. The first
object in use, begins at 9 bytes into the file. It has not been updated or changed because the
generation number is still 00000. The last object begins 445 bytes into the file, and also has not been
modified.
trailer
<<
/Size 8
/Root 1 0 R
>>
startxref
553
%%EOF
After the xref is the trailer. This reports the total number of objects, including the free object.
It also tells the root document, which is the Catalog, object number one. The trailer is contained in
opening << and closing >>. When the reader opens the the PDF document, it starts at the end of the
file and reads the trailer and the xref first. This is why startxref comes after the xref has been defined.
Reading the end of the file first allows the reader to know how far into the document each object
starts so that individual pages can be accessed without having to read or download the entire
document. After startxref, the number of bytes into the file where the keyword xref begins, is
reported. Finally, %%EOF marks the end of the document.
The 14 Type1 BaseFonts Guaranteed to be in Adobe Acrobat Reader
Courier
Courier-Bold
Courier-Oblique
Courier-BoldOblique
Symbol
Helvetica
Helvetica-Bold
Helvetica-Oblique
Helvetica-BoldOblique
Times-Roman
Times-Bold
Times-Italic
Times-BoldItalic
ZapfDingbats
PDF 1.2 Reference Manual November 27, 1996 :
APPENDIX B
Summary of Page Marking
Operators
Following is a list of all page marking operators used in PDF files, arranged
alphabetically. For each operator, a brief description is given, along with a
reference to the page in this document where the operator is discussed in detail.
Words shown in boldface in the summary column are PostScript language
operators.
Table B.1 PDF page marking operators
Operator Summary Page
b closepath, fill, and stroke path 227
B fill and stroke path 227
b* closepath, eofill, and stroke path 227
B* eofill and stroke path 227
BDC begin marked content, with a dictionary 240 PDF 1.2
BI begin image 238
BMC begin marked content 240 PDF 1.2
BT begin text object 233
BX begin section allowing undefined operators 239 PDF 1.1
c curveto 223
cm concat. Concatenates the matrix to the current transformation 213
matrix.
cs setcolorspace for fill 220 PDF 1.1
CS setcolorspace for stroke 221 PDF 1.1
d setdash 215
d0 setcharwidth for Type 3 font 239
d1 setcachedevice for Type 3 font 239
Do execute the named XObject 236
333
: November 27, 1996
Table B.1 PDF page marking operators
Operator Summary Page
DP mark a place in the content stream, with a dictionary 241 PDF 1.2
EI end image 238
EMC end marked content 240 PDF 1.2
ET end text object 233
EX end section that allows undefined operators 239 PDF 1.1
f fill path 226
F fill path 226
f* eofill path 226
g setgray (fill) 219
G setgray (stroke) 220
gs set parameters in the extended graphics state 217 PDF 1.2
h closepath 225
i setflat 213
ID begin image data 238
j setlinejoin 216
J setlinecap 214
k setcmykcolor (fill) 220
K setcmykcolor (stroke) 220
l lineto 223
m moveto 223
M setmiterlimit 217
MP mark a place in the content stream 241 PDF 1.2
n end path without fill or stroke 226
q save graphics state 213
Q restore graphics state 213
re rectangle 224
rg setrgbcolor (fill) 220
RG setrgbcolor (stroke) 220
334 Adobe Systems Inc.
November 27, 1996
Table B.1 PDF page marking operators
Operator Summary Page
s closepath and stroke path 226
S stroke path 226
sc setcolor (fill) 221 PDF 1.1
SC setcolor (stroke) 221 PDF 1.1
scn setcolor (fill, in pattern and separation color spaces) 221 PDF 1.2
SCN setcolor (stroke, in pattern and separation color spaces) 221 PDF 1.2
Tc set character spacing 229
Td move text current point 233
TD move text current point and set leading 233
Tf set font name and size 230
Tj show text 234
TJ show text, allowing individual character positioning 235
TL set leading 230
Tm set text matrix 233
Tr set text rendering mode 232
Ts set super/subscripting text rise 232
Tw set word spacing 229
Tz set horizontal scaling 230
T* move to start of next line 233
v curveto 223
w setlinewidth 216
W clip 227
W* eoclip 227
y curveto 224
' move to next line and show text 234
" move to next line and show text 235
: 335