intertext.me

早期freebsd实现

.bp
.sh 1 "A Textual Representation for \*(IP"
.sh 2 Introduction
.lp
This section describes a language used to represent an Interpress master in
human readable form.
In order to make an Interpress master compact,
an encoding scheme is used to represent the operators and the values.  
Although this encoding uses less disk space and results in shorter
transmission times, it is practically impossible to examine or edit an
Interpress master.  For this reason, the utilities \fIiptotext\fP and
\fItexttoip\fP represent the master as a text file.
For someone to completely use the Interpress utilities and understand the
output, an understanding of the text representation is necessary.
It is recommended
that the reader of this section have a working knowledge of Interpress.
.sh 2 Tokens
.lp
The keyword \*(lqHeader\*(rq indicates that the following string is the
header that appears at the front of the Interpress file.  This keyword can
only be used once at the top of the file.  It is not checked for validity.
.sh 2 "Integers"
.lp
A series of digits optionally preceded by a plus or minus sign is an integer.
.sh 2 "Short numbers"
.lp
Any integer that starts in the first column is translated into an Interpress 
\fIshort number\fP.  This number must be in the range \(mi4000..28767.
.sh 2 "Operators"
.lp
Any word that starts in the first column
is translated into an operator, with the exceptions noted in \fIkeywords\fP.
The \*(lqpseudo\*(rq operators are handled as follows.
The word \*(lqBEGIN\*(rq marks
the beginning of a block.  The word \*(lqEND\*(rq indicates the
end of a block.  The character \*(lq{\*(rq indicates the beginning
of a body and the character \*(lq}\*(rq indicates the end of a
body.  All other words refer to an Interpress operator.  The names found in
Appendix B of the \fIInterpress Electronic printing Standard\fP
are recognized.  The case of the letters is insignificant.
.sh 2 "Comments"
.lp
Anything that appears in parentheses is assumed to be a
comment and is ignored.  These comments should not be confused with a
comment sequence.
.sh 2 "Sequences"
.lp
If the first
character on the line is \*(lq>\*(rq then the remainder of the line
describes a \fIsequence\fP.
If the second character is also \*(lq>\*(rq then the
sequence described is \fIlong\fP, otherwise it is \fIshort\fP.
The keyword that starts
in the third column describes the type of sequence.  The value of the
sequence appears after the colon.  The following names are valid sequence
names:  Comment, Identifier, Insert file, Integer, Rational, String.
.sh 2 "Strings"
.lp
Strings are represented in the same way they are in C.
A double quote is used to mark each end of the string
and a backslash is used to indicate the start of a special
character.  If a digit follows a backslash, it starts an octal number that
does not exceed three digits.  This number is the octal value of the
character represented.  For example, the string "a\e244b" consists of an
ascii \*(lqa\*(rq followed by the octal value 244 and ending with an ascii
\*(lqb\*(rq.  Two successive backslashes represents a single backslash and a
backslash followed by a double quote represents a double quote.  It will not
be uncommon in program listings to use strings like:
"#include\ \e"file.h\e"".
.sh 2 "Rationals"
.lp
A rational number is represented as a fraction with a slash separating the
numerator and the denominator.
For example:  \*(lq1/2\*(rq represents the rational 1 over 2.
.sh 2 Format
.sp
Here is a Backus-Naur style grammar that describes the text format
(tokens are in italics, literals are contained in double quotes):
.sp
.(l
textFile ::= prelim body | body

prelim ::= file header | file | header

file ::= "File: " \fIstring\fP \fInl\fP

header ::= "Header: " \fIstring\fP \fInl\fP

body ::= line \fInl\fP\^ body | line \fInl\fP

line ::= \fIinteger\fP | \fIoperator\fP | short-long sequence

.ta \w'sequence ::= 'u
sequence ::= "Comment: " \fIstring\fP |
	"Identifier: " \fIidentifier\fP |
	"Insert file: " \fIstring\fP |
	"Integer: " \fIinteger\fP |
	"Rational: " \fIinteger\fP\^ "/" \fIinteger\fP |
	"String: " \fIstring\fP

short-long ::= "> " | ">>"
.)l
.sp
The token \fInl\fP\^ is a newline or line terminator.
The other italicized tokens are explained in the previous section.
.sh 2 Omissions
.lp
There is currently no way to express large vectors or any
of the pixel vectors.  There is also no way to specify a continued
sequence.