Note: Descriptions are shown in the official language in which they were submitted.
`8
IMPROVEMENTS RELATING TO TEXT PROCESSING APPARATUS
This invention relates to text processing apparatus in
which text is displayed on a video display.
Modern word processing apparatus use video displays tc
present textual data to an operator as it is entered or for
editing purposes. In the former case, as the operator keys
characters into the apparatus, in much the same way as with a
typewriter, the characters are presented on the video display:
when keying is complete, the operator can check what has been
entered before the text is printed or, alternatively, trans-
mitted to a remote screen for display or a central file for
storage. When text is to be edited, the text is displayea on
the screen and by using the keyboard to enter commands and text
information, the operator can insert, delete or transpose text.
To improve operator productivity, it is convenient if text can
be handled on a word-by-word basis rather than on a character-
by-character basis. It is also helpful if the use of operator-
entered carriage return commands can be minimized.
In the embodiment of the invention which will be des-
cribed below~ characters can be deleted and underscored on
a character-by-character basis. However provision is also
madQ to allow the operator to delete and underscore whole
words. Furthermore, in the apparatus to be described, carriage
return codes are not normally entered by the operator. As
will be seen, the operator only has to enter textual data
and the apparatus will automatically take care of word over-
spill conditions. In other words, if a word at the end of
a line of text extends beyond the right hand margin, the
apparatus will automatically transfer this word to the
beginning of the next line of text.
UK9-77-004 -1-
Various attempts haye heen made in the past to ease
the load on operators by giving assistance at the end of the
line, for example by automatic hyphenation techniques, but
generally these require relatively large and expensive
equipment.
United States of America Patent No. 3,631,957 describes
a system for producing output copy from an input tape. When
word overspill is detected, a choice is given to the operator
as to where to hyphenate. A variable-width control zone gives
varying degrees of "raggedness" and need for hyphenation at
the right hand margin.
United States of America Patent No. 3,688,257 discloses
a formatted text display system in which end-of-line codes
are inserted before words which are detected as causing
Overspill. A register containing the address of the last
space on the line is used to determine where the end-of-line
code is to be inserted.
United States of America Patent No. 3,550,091 uses a
counter to keep track of the last space on a line. Just
before the right-hand-margin, words are entered on the next
line as well as the current line. If a space is subsequently
encountered before the margin, the word on the next line is
deleted: if no space is encountered before the margin, the
word entered on the current line is deleted.
None of this prior art apparatus discloses or susgests
the use of a word boundary counter which contains the address
of the next word with the provision of being able to "freeze"
the contents of the counter to allow text to be entered
and/or edited as a complete word rather than on a character - -
by character basis.
According to the present invention, a text processing
UK9-77-004 -2-
~ ~.
.
apparatus comprises a keyboard for entering text and command
information, a video display for displaying text, a message
buffer for containing text to be~displayed, means for detecting
the presence of a word delimiting character during cyclic
scanning of said buffer to determine the start of a word in
said buffer, a word boundary counter adapted to be loaded in
succession with the address on the display of each word of
text as said buffer is scanned, and means for preventing
loading of said counter with the address of subsequent words
when a processing operation is to be performed on a word and
for allowing loading of said counter with the address of
subsequent words when said processing operation is completed.
In order that the invention may be readily carried
into effect, a preferred embodiment thereof will now be
described, by way of example, with reference to the accomp-
anying drawings, in which:-
Figure 1 is a diagrammatic representation of a text
processing apparatus, -
Figure 2 is a block diagram illustrating a preferred
embodi~,ent of the invention,
Figure 3 shows the format of a command word,
Figure 4 shows the format of text code,
Figure 5 illustrates the use of temporary left margins
with the apparatus of Figure 2, and
Figure 6 illustrates how word overspill is accomplished
in the apparatus of Figure 2.
Referring now to Figure 1, a text processing apparatus
comprises a keyboard 1 by means of which text or commands
can be entered, and a video display 2 on which text can be
displayed during text entry or for editing. The text
processing apparatus is connected to a controller 3 by a
cable 4. Other displays and keyboards, not shown, can be
UK9-77-004
.~
~ ~.
connected to the controller 3~ Attached to the controller
is a printer 5, for example a so-called daisy-wheel printer or
a line printer, by means of which text displayed on the display
2 can be printed. Also attached to the controller 3 is a text
store 6, for example a magnetic disc store, in which text can
be stored electronically for later recall. The stored text
may include standard paragraphs etc. which can be incorporated
into individual letters. Optionally, the controller 3 can be
connected to a host central processing unit 7: such an arrange-
ment would be useful, for example, for electronic mail purposes.
Typically, the video display 2 and keyboard 1 can beconstituted by an IBM (Registered Trade Mark) 3277 Model 2
Video display unit suitably modified as described below.
The controller 3 can be constituted by an IBM (~egistered
Trade Mark) 3791 controller. Since the controller 3, printer
5, store 6 and host processor 7 are not directly relevant to
an understanding of the present invention, they will not be --
further described.
The keyboard 1 has a number of keybuttons 8 by means of
which alphanumeric characters and other symbols can be keyed
in by the operator. Keybuttons 8 can typically be arranged
as in a normal typewriter. Also provided on keyboard 1 are
keybuttons 9 which are used for entering command information
into the apparatus. Typically the keys 9 can be used to
initiate word delete, word underscore, character deletion,
to move a cursor on the display, to set up tabs, temporary
l~ft and right margins, etc.
As indicated above, text is displayed on the video
display 2. Also displayed to assist the operator are a
temporary left margin (TLM) symbol 10, a right hand margin
(RHM) symbol 11, a scale 12 showing the various character
positions across the display and on which tab settings
UK9-77-004 4
B
would normally be indicated, and a cursor 13 showing the
current point of interest. Shown on the display 2 is a
message in which the words `'underscored" and "or" which
were being entered on the third and fourth lines respectively
(thereby causing overspill) have been deleted and displayed
on the fourth and fifth lines respectively: this operation
of word overspill will be described in detail below.
Those familiar with the IBM (Registered Trade Mark)
3277 video display unit will recall that up to 1,920 characters
can be displayed in 24 rows each containing up to 80 character
positions. Each row of characters on the screen is constituted
by g raster scan lines, with each character being formed using,
effectively a 7 x 9 dot matrix. It will be appreciated that
although the preferred e~odiment is described in terms of
such an arrangement, other arrangements are also possible.
All that i6 necessary is that some message buffer containing
the characters to be displayed is cyclically scanned.
Referring now to Figure 2, data to be displayed are
received on line 4 together with command information. The
data are deserialized in interface unit 15 before being passed
along line 17 to decoders 16 whose functions will be described
in more detail later. Characters to be displayed are sent on
line 18 to a message buffer 19 formed from cyclic shift reg- -
isters and which has 1,920 character positions. The message
buffer is cyclically scanned and a line (row) of characters
constituting a line of text is assembled in line buffer 20
containing 80 character positions. Thus line buffer 20 will
contain successive llnes of data to be displayed on the video
display. Before this data can be displayed however, they
must be decoded into video information and to this end line
buffer 20 accesses a character generator 21 in the form of
UK9-77-004 5
%
a read only store. Video information from the Read Only
Store 21 is passed to the video display along line 22. A
buffer address register 23 contains the address on the
screen of the current character of interest. Register 23
is divided into two parts, a first part 23A containing the
character count identifying the position of the character
within a row on the screen and a second part 23~ containing
the row count indicating the row on the screen. In the
example under discussion, part 23A is a 7 stage register and
part 23B is a 5 stage register. A parking register 24 is used,
as will be described in more detail below, during insert
and delete operations. The data stream from the message
buffer 19 on line 25 is rewritten into the buffer 19 through
an OR gate 26, being delayed in parking register 24 under the
control of control logic 27 as necessary when editing
operations are being performed.
Interface unit 15 also allows the keyboard 1 to commun-
icate with the decoders 16 or with the controller along line
4. The apparatus so far described with reference to Figure
2 will be recognized by those familiar with the IBM (Registered
Trade Mark) 3277 video display unit as standard. The dif-
ferences and modifications necessary to practice the
present invention will now be described.
Characters in the data stream can be divided into three
categories, text characters which are to be displayed,
formatting characters which delimit a word, for example
"spaces" and "required carrier returns", and other formatting
characters which are not classed in this e~odiment as wGrd
delimiters" formatting characters may or may not be displayed.
A "space" is a formatting character which normally separates
words. "Required carrier returnH is a formatting character
which is used to force the end of a line of text even though
UK9-77-004 -6-
.%
there may be enough room for more words on that line: it may
be displayed to provide operator assistance during editing
functions. A "null" is a formatting character which is
distinguishable from a l'space" in that it represents a charac-
ter position on the screen which can be over-written with
another character without causing shifting of subsequent
characters to the right. A "temporary left margin" (TLM)
character is used to specify the temporary left margin: it
is displayed during text editing to provide operator assistance.
Other formatting characters, not described in this specific-
ation since they are not required for an understanding thereof,
include "new paragraph space" (NPS), "tab", "new paragraph
tab" (NPT) and attribute characters for protect/non-protect
and display/non-display. A word is normally defined as a
string of text characters having one or more delimiters
preceding and following it. However non-delimiting formatting
characters are treated as a one-character word without the
need for a delimiter to preceed or succeed it.
Within the interface 15 and decoders 16, characters are
coded as 13-bit bytes as shcwn in Figure 3. This is true
of characters to be displayed, formatting characters and
other commands from the controller. Characters within the
message buffer 19, on the other hand, are coded as 8-bit
bytes as shown in Figure 4. The eighth bit of each byte is
used to indicate ~hether the character is underscored. If
this eighth bit is set, the character is caused to address
a different section of the character generator 21 to obtain
the underscored version of the character to be displayed. As
an alternative, circuits, not shown, could be caused to
generate an underscore and to feed this to the screen with
the bit stream forming the character. This is represented by
UK9-77-004 -7-
line 42 in Figure 2~
An essential feature of the invention is a word boundary
counter (WBC~ 31 which keeps track of words in the data stream.
When a delimiter (space or RCR~ is decoded, a latch 32 is
set to prime WBC 31. On the next text character, the row
and character counts in register 23 are loaded in parallel on
lines 33 and 34 into sections 31A and 31B respectively of
WBC 31 so that the contents of WBC 31 always point to the
start of words in the data stream. (It will be appreciated
that it is a matter of choice whether the WBC 31 points to
the first character of a word or the character positisn
immediately preceding the first character). When a non-
delimiting formatting character is decoded (defined above as
a one character word), the WBC 31 is loaded with the values
Of the row and character counters 23A and 23B. During each
scan of the message buffer 19, the word boundary counter 31
will normally point in turn to the start of each word in
the data stream appearing on line 25. As will be explained
in more detail below, when a word operation such as word
overspill, word delete or word underscore is executed, the
normal loading of the WBC 31 is inhibited and its contents
are frozen: section 35A of the WBC 31 may then be stepped
as a normal counter by means of line 32 from control logic
27 until processing of that word is complete.
A hold control 41 is used to inhibit setting of latch
32 by line 43 thereby freezing WBC 31 during a word overflow
operation, word dele~e operation or word underscore operation:
a signal appears on line 43 whenever decoder 28 detects a
delimiting character.
A character decoder 28 examines the data stream on line
25 from message buffer 19. If a temporary left margin (TLM)
UK9-77-004 -8-
character is recognized, the address of the temporary left
margin is set in TLM register 22, Decoded signals from
decoder 28 are also passed on line 30 to control logic 27,
The TLM register 29 is set when a TLM character is
entered by the operator into the position where the TLM is to
start or when a TLM character is detected during the buffer
scan. The TLM is terminated by a RCR character.
Right Hand Margin (RHM) register 36 contains the
address of the right hand margin. In this embodiment, in
the absence of any specific command, it will be the right
hand edge of the display area. The value in the register
36 can be set either by a command from the controller or
under operator control from the keyboard 1. Characters
- entered to the right of the right hand margin must be nulls,
otherwise they will be shifted to the next line.
Compaxators 37, 38 and 39 allow the comparison of the
contents of the word boundary counter 31 with the buffer
address register 23, the column or character count in the
buffer address register 23 with the RHM register 36, and the
column or-character count in the WBC 31 with the RHM register
36 respectively. The buffer address register 23 ccntains
the address of each character of interest in turn within
buffer 19 so that comparator 37 indicates to the control logic
23 whenever scan of the buffer 19 has reached the start of
~5 a word. Comparator 38 indicates to the control logic 23
when the buffer scan is at the right hand margin and com-
parator 39 indicates when the first character of the word is
at the right hand margin. Comparator 40 compares the value
contained within the TLM register 29 with the character count
in register 23: in other words comparator 40 indicates
when the character count is at the left hand margin.
UK9-77-004 9
~,
The operation of the apparatus ~ill now b.e descrihed
in more detail. Assume that the rideo screen is completely
blank and that the operator wishes to "type" a document.
The right margin can be set either from the controller
or from the keyboard. When the right margin command is
received in the decoder 16, the RHM register 36 will be
loaded with an 8 bit byte defining the column position
of the right hand
'
~.
UK9-77-004 -9a-
X
.
-'
1 margin.
The left hand margin will either be set by the con-
troller or by the operator. Assume that a temporary left
margin is to be set by the operator and to this end a left
hand margin key is operated with the cursor in the required
position on the screen. This causes the TLM character to
enter the message buffer 19: when it is scanned and decoded
by decoder 28, its position will be entered into TLM
register 29.
All other character positions will be filled with "nulls".
- The cursor will be located at the top line to the immediate
right of the left margin. Using appropriate keys on the
keyboard 1, the operator can enter desired text and format-
ting information into the apparatus.
Figure 5 illustrates the use of the temporary left
margin. Text characters and spaces 47, shown circled to
indicate that they are not visible on the screen, have been
entered on to the first line. It is desired to indent the
second to fourth lines and to this end the operator moves
the cursor into column "05'~ and then uses a "temporary left
margin" key on keyboard 1. The TLM address will be entered
into register 29 from character counter 23A and a TLM
symbol 46 is displayed on the screen. It will be appreci-
ated that during key-stroking, the message buffer 19 is
being continually scanned. Character positions before the
TLM 46 are filled with nulls 48 shown circled because they
are not displayed on the screen. Text is entered on the
second line of the displayed text. When text is entered on
the third line, the value in TLM register 29 ensures that
the first text character is placed in column "05", that
is the temporary left margin. A similar process happens
on the fourth line of displayed text except that the
UK977004 10
1 operator decides to terminate the temporary left margin. To
this end a required carrier return (RCR) key is operated
when the cursor is positioned in column "12". As a result the
4 RCR character
UK9-77-004 lOa
B
\
44 is displayed and the contents of TLM register 29 are
erased. Thus on the fifth line of text, the first text
character will be displayed in column "01".
The temporary left margin can also be used for indent-
ing paragraphs. Thus the first eight characters entered in
the sixth line of displayed text can represent the heading
of a paragraph. The TLM key is operated with the cursor in
column "09" and then text is entered. As will be seen, text
entered on the seventh and eighth lines of displayed text
10 will be indented with columns "01" to "08" being filled with
nulls 48. To terminate the TLM, the RCR key is operated
with the cursor in column "15". Text entered on the ninth
line begins at column "01". In Figure 5, the cursor 45 is
shown in column "04" indicating the position of the next
character (whether text or formatting) to be entered.
When the text shown in Figure 5 is printed out, the
characters on the second and sixth lines to the right of
the TLM character 46 will all be shifted one position to -~
the left. This compensates for the TLM character which is
; 20 displayed but not printed.
Each time that a TLM character is encountered during
a message buffer scan, its address will be entered into
the TLM register 29 via character decoder 28. Comparator
40 compares the TLM register 29 with the character count
in buffer address register 23 to control the time that an
overflow character from the hold register 24 should be
inserted into the new line.
The operation of word overspill will now be described
with reference to Figure 6. In Figure 6, the words "WORD
30 SPILL" are being entered at the end of a line. The
- operator enters the characters W,O,R,D, space (symbol 47
circled to show it is invisible) S,P,I,L and L in that order
- UK977004 11
.~
1 but makes no attempt to enter a "carriage return". As shown
2 in Figure
; .
;~
UK9-77-004 lla
. . .
6A, the right margin is set at column "8Q"~ When the "space"
character 47 in column "73'` is decoded by character decoder
28 during the scan of the message buffer 19 next following
its entry, the word boundary counter 31 will be primed by
5 means of latch 32 and when the character "W" is decoded, the
contents of register 23 will be entered into counter 31 ie
column 74. When the next space character 47 in column "78"
is decoded, the counter 31 is again primed and the start of
the next word, character S in column "79", is entered. Normally
10 the counter is set whenever a delimiting character is scanned
from the message buffer. However when comparator 38 indicates -~
to the control logic 27 that the right hand margin, column 80,
is occupied by the character "P", the contents of the word
~oundary counter 31 are frozen by a signal from the control
15 logic 27 to the hold control 41 and will not be set to the
address of the next word until the word overspill operation
is complete. During the next scan of the message buffer 19,
a "null" character 48, Figure 63, will be entered into column
79 by the control logic 27. The contents of counter 31 are
20 incremented by 1 on line 35 to give the new start of the word.
Thus the character "S" which previously occupied column "79"
will now occupy column "80". This condition is shown in
Figure 6B which also shows the start of the next line. In
this next line, "null" characters 48 (shown circled to indicate
25 that they are invisible) are shown in columns "01" and "02".
(The TLM character 46 is shown in column "02" of the previous
line). During the next scan of message buffer 19, a null
will be inserted before the start of the word, the counter -
31 will be incremented by 1, and the remainder of the line
30 shifted by one position to the right. This process will be
repeated until comparator 39 detects that the right hand
UK9-77-004 -12-
X
.
:,. - - .. , - . .
margin, column "80" is occupied by the start of the word:
this allows the process of overspill to be completed at the
next scan when normal operation of the word boundary counter
31 will continue. As characters are shifted from the end
of the line, they will appear in the first vacant position
on the next line. This is ensured by placing the shifted
character in parking register 24 and placing it in the
data stream after the temporary left margin character has
been decoded. The completed operation is shown in Figure
6C. What has been described is the operation of word overspill
when a document is being created. An essentially similar
operation will occur when displayed text is being edited.
If a word or character is inserted, this can cause word
overspill on the same line of text and possibly subsequent
lines. In this event, the word overspill condition, ie
the presence of a character other than a null at the right
hand margin (set to column 80 in this case), would be
detected by the comparator 38 during scanning of the message
buffer l9, the current contents of the word boundary counter
would be frozen by hold control 41, and nulls would be
inserted into the data stream.
It will be appreciated that scanning of the buffer 19
occurs at a much higher rate than the operator keying rate.
Therefore there is normally no need to buffer the keyboard~
however if certain of the keys are typamatic in operation,
ie continued depression causes multiple entry of the
corresponding characters, a buffer may be provided in the
interface 15 to avoid any possibility of lost characters due
to the time taken to execute word overflows. (Typamatic
operations are typically approximately 10 characters per
second and the normal scan speed of the message buffer is
UK9-77-004 -13-
approximately 2Q milliseconds]. When a word operation is in
progress with the contents of the word boundary counter 31
frozen, fast scanning of the buffer 19 can be initiated with
each scan taking approximately 2 milliseconds. During fast
scanning, it may be desirable to blank the screen by means
of a signal on line 49 to prevent degradation of the picture
or changes from distracting the operator. When the word
operation is complete, normal scanning of the buffer 19 can `-
; be resumed.
Another word operation which can be performed on the
apparatus of Figure 2 is word delete. The operator can key
in a request for word delete by using the appropriate key
on keyboard 1. The cursor can be at any position within the
word. When the word delete request is decoded, the word
lS boundary counter 31 is frozen at the start of the word when
the cursor is decoded during the next scan of the buffer 19.
Each character in the word so specified will be deleted from
the message buffer and screen, one character being deleted
- ~ for each scan. The word counter 31 is incremented by 1 for
each character deleted. As each character is deleted, the
row of characters shifts to the left, the cursor remaining
stationary. The word delete mode is reset and the counter
31 released to operate normally when the start of the next
~ word is detected. Continued (typamatic) operation of the
- 25 word delete key will cause deletion of the next word or
words to the e~d of the line. As an alternative, deletion
can be commenced from the cursor position to the end of the
word: those skilled in the art will appreciate that only a
minor modification to the apparatus descxibed will be
required.
~; As mentioned above, underscore is achieved by adding
a bit to each character code in the message buffer 19. The
UK9-77-Q04 -14-
.
~3 3~
operator can call for word underscore by using the appropriate
key on the keyboard 1, The word to be underscored is defined
by the position of the cursor on the screen which may be at
any position in the word or at a delimiter of the word or a
null following the word. When the cursor is detected in the
data stream from message buffer 19 by decoder 28, the
contents of the word boundary counter 31 are frozen. If the
cursor is at the start of the word, a condition recognized
by comparator 37, then underscore begins immediately with
the control logic 27 causing the underscore bit to be added
to each data character in the word. Underscore will be
completed in one scan of the message buffer 19. If the cursor
is not at the start of the word, the contents of the counter
31 are frozen when the cursor is detected by decoder 28 and
is held for the next cycle of the data stream. One under-
score bit is added to each character as it is scanned from the
message buffer 19: the contents of the counter 31 are
incremented by 1 for each character and underscoring is com-
plete when the first delimiter or non-data character is
detected by the character decoder 28. When the value in
counter 31 is equal to the character and row counts in
register 23, the start of the word has been reached and
insertion of the underscore bit commences and continues as ~ -
described previously. In this manner, the operator can key
in a word of text and then underscore it without having to
reposition the cursor to the start of the word and then move
it on to key in the next word. If word underscore is
called and the cursor is at a formatting character, only
that one formatting character (defined above as a one
character word) will be underscored.
Those skilled in the art will appreciate that many
UK9-77-004 -15-
. .
.
l'L l~l~B
modifications to the apparatus described can be made. Thus
although mention has been made of the use of particular
keys for use during the word delete and underscore operations~
the decoders 16 are easily able to be designed to recognize
particular combinations of keys operated at the keyboard
to signify a re~uest for those operations. The apparatus
has also been described in relation to the entry and
display of text read from left to right. It will be apparent
that the invention also has application to those scripts,
for example Hebrew and Arabic, which are read from right
to left. In these cases word overflow would occur at the
left hand margin. Similarly the invention can be used in
text processing apparatus in which text is read from top
- to bottom.
UK9-77-004 -16-
