Google
 

Trailing-Edge - PDP-10 Archives - decuslib10-08 - 43,50506/keywrd.rnh
There are 2 other files named keywrd.rnh in the archive. Click here to see a list.
.page size 1000,60
.noheader
^^The KEYWRD program produces a sequence of tests which can identify
the leading word or the leading phrase formed of a fixed sequence of
words in a line of text without ever having to test a character
which has already been identified. Such a leading word or phrase
will be referred to as a command. A command does not need to include
any characters to the right of the first of the characters which
uniquely identify the command. A word or each of the words in a
phrase can be abbreviated by truncation, leaving at least the left
character in each word of the phrase if additional words or their
abbreviations appear to the right. Spaces are allowed between the
words in a phrase, but are not required. A single sequence of tests
is used to recognize the initial portions of commands which start
with a common series of characters, then the unique portion of each
command is identified by a separate sequence of tests. After the
unique portion of each command has been identified by the separate
sequence of tests, then a single sequence of tests is similarly used
to recognize the final portions of the commands which end with a
common series of characters.
.break
#
.break
The KEYWRD program reads a single input file and produces an output
listing file and an output FORTRAN language file containing DATA
statements which represent the sequence of tests. These DATA
statements must be merged into the FORTRAN program by which these
tests are to be used. The KEYWRD program is written in FORTRAN, and
is machine independent except for the short subroutine which asks
the user for the file names and then opens these files.
.break
#
.break
Each line in the input file which does not start with one of the
reserved characters *, /, ( or ), which are described later,
contains a single word or phrase preceded by a nonzero value by
which the word or phrase is to be identified. The number should not
duplicate a number to be associated with any other command unless
these commands are synonyms or unless some of these commands are
abbreviations which would otherwise be ambiguous. The number can be
preceded by one or more spaces, but leading spaces are not required.
The number cannot contain any characters other than the digits 0
through 9 and a leading minus sign if the value is negative or an
optional leading plus sign if the value is positive. Spaces and/or a
single comma can appear between the leading number and the following
word or phrase, but are not required. Words within a phrase must be
separated by at least 1 space. Extra spaces are ignored. Words and
phrases can be constructed from any characters, but upper and lower
case alphabetic letters are considered to be equivalent. The
sequence of tests produced by the KEYWRD program is independent of
the order in which the commands are defined in the input file. The
input file is terminated by a line containing a number which is not
followed on the same line by any word or phrase.
.break
#
.break
.fill
If the words and phrases are constructed from characters other than
spaces and the alphabetic letters A through Z, then an additional
DATA statement is generated which specifies a third array, LTRXTR,
containing the unexpected characters. KNTXTR, which is specified by
one of the DATA statements which are always generated, is the size
of the LTRXTR array. If the words and phrases are constructed only
from spaces and the alphabetic letters A through Z, then KNTXTR has
the value zero and the DATA statement defining the LTRXTR array is
not generated.
.break
#
.break
The first location in the NOTPNT array describes the first test
which is to be performed. The absolute value of each entry in the
NOTPNT array is the sum of the location within alphabet of the
letter to be matched times (KNTPNT+1), plus the subscript of the
location in the NOTPNT array which describes the next match which is
to be attempted if the current match is a failure. The subscript of
an array location is its serial position within the array, counting
the first value in the array as being at subscript 1, the second
value as being at subscript 2, and so on. If an entry in the NOTPNT
array is negative, then the character starts a word and any spaces
in the input line can be skipped. If the location within the
alphabet is greater than 26, then this minus 26 is the location
within the LTRXTR array of the character to be matched. If the match
succeeds and the value to be associated with the command is positive
or if the value is zero indicating that the match does not uniquely
identify a particular command, then the parallel location in the
MCHPNT array contains the sum of the value of the command times
(KNTPNT+1), plus the subscript of the location in the NOTPNT array
which describes the next test. If the match succeeds and the value
to be associated with the command is negative, then the parallel
location in the MCHPNT array instead contains the value of the
command times (KNTPNT+1), minus the subscript of the location in the
NOTPNT array which describes the next test. If the subscript of the
location in the NOTPNT array which describes the next test is
indicated to be zero, either by the MCHPNT array if the current
match is a success or by the NOTPNT array if the current match is a
failure, then no additional test remains to be performed, and the
command is identified by the last nonzero value encountered in the
MCHPNT array for a match which succeeded.
.break
#
.break
Lines in the input file which start with an asterisk, a slash, a
left parenthesis or a right parenthesis are treated specially. These
initial characters cause the following actions to be performed.
.break
#
.break
.left margin 3
.indent -3
/##An initial slash indicates that the line specifies the names of
the arrays and variables which are to be represented in the DATA
statements which are to be written into the output FORTRAN statement
file.
.indent -3
*##An initial asterisk indicates that the line specifies 5 numbers
which characterize the sequence of tests produced by the KEYWRD
program. Such a line would appear in the input file only when the
result is already known and the operation of the KEYWRD program is
being verified. The numbers can be separated by spaces and/or by
single commas. A sixth group of up to 6 characters defines the label
to be shown to the user.
.indent -3
(##An initial left parenthesis indicates that the rest of the
current line is to be copied into the output FORTRAN statement file
unchanged. This does not interrupt the specification of the glossary
of keywords.
.indent -3
)##An initial right parenthesis indicates that the DATA statements
which represent the sequence of tests are to be written into the
output FORTRAN statement file. This does not indicate that the end
of the file has been reached.
.break
#
.break
.left margin 0
If a line starts with a slash, then the next 5 groups of printing
characters on the line are used as the names of the 3 arrays and the
2 variables which are represented in the DATA statements which the
KEYWRD program writes into the output FORTRAN statement file. These
groups of printing characters can be separated by spaces and/or by
single commas. The names of the 3 arrays must each differ from the
others in their first 3 characters.
.left margin 3
.break
#
.break
.indent -3
1.#The first group of up to 6 characters is used as the name of the
array which specifies the next operation if the match fails. This
name is NOTPNT if a line starting with a slash does not appear in
the input file.
.indent -3
2.#The second group of up to 6 characters is used as the name of the
array which specifies the next operation if the match succeeds. This
name is MCHPNT if a line starting with a slash does not appear in
the input file.
.indent -3
3.#The third group of up to 6 characters is used as the name of the
nondimensioned variable which contains the number of items in each
of the previous 2 arrays. This name is KNTPNT if a line starting
with a slash does not appear in the input file.
.indent -3
4.#The fourth group of up to 6 characters is used as the name of the
array which specifies all characters, other than spaces and the
letters A through Z, appearing in the commands. This name is LTRXTR
if a line starting with a slash does not appear in the input file.
.indent -3
5.#The fifth group of up to 6 characters is used as the name of the
nondimensioned variable which contains the number of characters in
the previous array. This name is KNTXTR if a line starting with a
slash does not appear in the input file.