US3102254A - Apparatus for minimizing coding errors - Google Patents

Description

M. M. LEVY Aug. 27, 1963 APPARATUS FOR MINIMIZING CODING ERRORS 6 Sheets-Sheet 1 Filed March 17, 1.960

w MW 7 7 7 W 7 w ll \MN NE n3 3 u] MUN Him 024 I uni Yd 014 I .ruMJ

Aug. 27, 1963 M. M. LEVY APPARATUS FOR MINIMI'ZING comma ERRORS 6 Sheets-Sheet 2 Filed March 17, 1960 mzadwa m b NH Aug. 27, 1963 M. M. LEVY v APPARATUS'FOR- MINIMI-ZING CODING ERRORS 6 Sheets-Sheet 3 Filed March 17, 1960 .Tnven W; H V V, n H

6 SheetsSheet 4 Av a:

M. M. LEVY APFARAT S FOR MINIMIZING CODING ERRORS Aug. 27, 1963 Filed March 17} 1960 M. M. LEVY Aug. 27, 1963 APPARATUS FOR MINIMIZING CODING ERROR S 6 Sheets-Sheet 5 Filed March 17, 1960 Aug. 27, 1963 M. M. LEVY 3,

APPARATUS FOR MINIMIZING CODING ERRORS I71 em a rice/f, Lev

War 7:4 7

Amway:

primary destination designation of such articles.

United States Patent 3,102,254 APPARATUS non MINIMIZING comm; ERRORS Maurice M. Levy, Ottawa, Ontario, Canada, assignor to Her Majesty the Queen in the Right of Canada as represented by the'Postmaster General, Ottawa, Ontario, Canada Filed Mar. 17, 1960, Ser. No. 15,721 9 Claims. (Cl. 340-1461) This invention relates to apparatus vfor minimizing the harmful effects of operators errors in the application a code designation can be imposed and which are destined subsequently to be sorted in accordance with such coding.

The invention is concerned with a coding system operating in accordance with the binary system. 'In a typical such system, the operator has five keys zfior each hand, one key corresponding to each finger, and each character tobe coded corresponds to a selected combination of finger operations. Thus, unlike a typewriter where each key corresponds uniquely to a given character, in the binary system each combination of. keys cornesponds to a given character (or, in some instances, characters, since the same combination may be 'used'for both a letter and a numeral). 7

The broad object of the present invention is the reduction of operators errors in such coding apparatus. It

has been found that the majority of errors which occur in practice result from either one desired key notbeing quite adequately depressed, or one key that should not be depressed being accidentally touched and operated. That isyto say, the majority of errors made by an experienced operator are one-finger errors, the combination of keysactually depressed differing only in respect of one finger (or one key) from the combination intended. It is not considered a one-finger error when one finger is displaced from its correct position so as to depress the wrong key. This is a twoger or two-key error, since, not only is one key depressed which should not be, but also one key is not depressed which should be. As a practical matter, the probability of such a displaced finger type of error is low.

.One important object of the present invention is the provision of apparatus for detecting one-finger errors in the operation of a keyboard. Once such errors have been detected, the detector may be employed simply to warn the operator of the erroneous coding attempted, or it may be employed automatically to block the application of the erroneous code to the article.

As will later be explained, the basic principles of the present'invention may be further extended also to detect some two-finger, three-finger and more complex errors.

i In its specific application to a mail coding and sorting system, the invention will normally be employed to detect all one-fingerand some multiple-finger errors in the code combination applied to articles of mail in respect of the In this way the number of total code combinations available is not universally restricted; it is only restricted in respect of the first code designation applied to the article. This is in keeping with practical requirements. If an occasional error causes a letter to be directed to a town within 39%2254 Patented Aug. 27, 1963 the sarne geographical area as the town intended, the matter can be readily corrected with only a small delivery delay. On the other hand, if a letter destined for Vancouver from Ottawa is erroneously sorted and sent to Great Britain, the resulting delay before the error has been rectified would be very serious. The avoidance of errors in the primary destination designation (country or large geographical or political area) is thus of special importance.

It is not intended to give the impression that any form of sort-ation error is considered trivial. With conventional' manual sorting methods errors have never been entirely avoided, and indeed with the modern improve ments. in mechanisation of postal sorting methods, it is believed that errors can be substantially reduced. The

principal purpose of the present invention is to add still further to this improvement by providing apparatus that n will automatically check the operators first chosen code combination (which will signify the main geographical or political area of the letters destination) and will compare this chosen code combination with a set of acceptable code combinations, and indicate when an error has been made. The present apparatus cannot check against all possible cperators errors, such as the substitution of an erroneous acceptable code combina tion for the correct acceptable code combination, but it can prevent all other errors in the initial coding step. Moreover, it can ensure that all one-finger errors (the most likely to occur) are amongst those detected during this initial coding step.

It is believed that the manner of operation of the present invention will become more readily apparent after reference has been made to the accompanying drawings 'whichdemonstrate by way of example one method of carrying the invention to practice. In the drawings:

FIGURE 1 shows a plan view of the keyboard panel of a coding desk;

FIGURE 2 shows a side view of FIGURE 1 cut away to show part of associated electrical circuits;

FIGURE 3a shows a portion of a matrix and an associated switching circuit controlled by the left hand key-;

board;

FIGURE 3b shows the switching circuit operated by the right hand keyboard and controlling the same matrix;

FIGURE 4 is a fragmentary circuit diagram;

FIGURE 5 is another similar fragmentary circuit diagram;

FIGURE 6 is a general circuit diagram illustrating the manner of operation of a printing and feeding mechanism in accordance with the arrangement of the matrix;

FIGURE 7 is a more detailed view of the printer of FIGURE 6; and j I FIGURE 8 is -a diagram showing some only of the parts of FIGURE 3a arranged in a manner to demonstrate the particular choice of coding combinations adopted in accordance with the present invention.

FIGURE 1 shows the keyboard panel of a coding desk 10, such panel comprising a first binary keyboard L having five keys, L1, L2, L3, L4 and L5 crepe-ration by the left hand and a second binary keyboard R having five'keys R1, R2, R3, R4- and R5 for operation by the right hand. In both instances the thumb is considered as'the first finger, so that the thumbs operate keys L1 and R1, the other fingers being numbered consecutively from the thumb towards the smallest finger. FIGURE 2 shows the manner in which each of the keys L1 to L5 .and R1 to R5 is mechanically arranged to close a pair of corresponding contacts L0]. to LCS and RC1 to RC5. Each such pair of contacts is arranged in series between a positive voltage supply 11 and ground with a corresponding relay coil LR]. to LRS and RR1 to R intereach numeral.

which reduces the number available to '31.

which follows:

posed. As will be apparent, depression of each one'of the finger-operated keys, will cause energization of the correspondingly designated relay coil.

Attention .is now directed to FIGURE 3a in which a binary switching system is illustrated for connecting the positive supply 11 to each one individually of thirty-two horizontal lines in a matrix 12. As far as this switching system is concerned, relay coil LR} operates only one movable contact arm LRlll, while relay coil LRZ has two such arms LR'Zll and LRZZ, relay coil LR3 has four arms LR31 to LR34, relay coil LR l has eight arms LR ll to "LR48 and relay coil LRS has sixteen arms LRSl to LR5-16; In this way any one of the thirty-two horizontal lines can be connected to the positive voltage supply 1-1 'byemploying a different combination of relay energiza- .tion. Each such line thus corresponds uniquely toa par- Q F ticular combination of operated keys. The switching arms are assumed to be in their. upper positions, as shown in FIGURE 3a, when the corresponding relay is un-e'ner- "gized. jEnergizati-on of a relay will cause its entire bank" of contact arms to move to their lower positions.

A particular combination of key operations is then chosen to' correspond to each letter of the alphabet and This is essentially an arbitrary choice. The letters required to be used most often may conveniently be made one-finger operations; the next most fre- 1 quently used letters, two-finger combinations and so on.

In this connection, it should be appreciated that it is the letters that will appear-most commonly in the cod-e designations, and not the letters which-appear most frequently in ordinary language, that are so chosen. With a five bit binary system, that is to say using'five keys, the number of combinations available is 2. that is 32. This number of combinations includes the one in which no keys are operated. It is inconvenient to use this combination, However,

since there are only twenty-six letters in the alphaibet,an-

, other five combinations can be unused. As a practical coding matter, there is no disadvantage inherent in using a particular code combination to signify both a letter and a numeral, so that some of the combinations are used for this purpose. In choosing the table of code designations consideration is also given to the manual ease with "which certain combinations of keys can readily be degpressed by the human hand.

A typical-list of code designations is shown in Table 1 In the left hand column, the key combination is shown,

the numerals inthis column referring to the numbering of the keys shown in FIGURE 1. The samecode system is conveniently used for both the left and the right hand, although there is nothing inherent in the system that demands this.

nation. This table will require to be memorized by the operator and will represent the principal mental human etfort involved. I i

The right hand column of Table 1 shows the code character that corresponds to each key combiquent purpose.

a I Taking the binary code set out of Table I and applying it to the matrix lines shown in FIGURE 3a, it immediately becomes possible to label each ofthe horizontal matrix lines with a corresponding code character. F or example the letter M is represented by the combination of keys 1, 2 and 4. Now, if the corresponding relay coils LRI, LRZ and LIM- of the left hand are assumed to be energized so that their banks of contact arms are movedto their lower positions (While the contact arms of relays LR3 and LRS remain in their 'upper positions) the,posi-- tive voltage from source 11 is conducted down through the first "contact arm LRll it encounters, down through the secondcontact arm LRZZ it encounters, up through the next contact arm LR34, down through the fourth contact arm LRl7 and finally up through the last contact arm LRSM- to impose such voltage on the sixth horizontal line from the bottom ofthe matrix 12. It will be observed that this line has been labeled at the extreme right hand end with the letter M, to signify this fact. All the other linessh-own in the matrix 12 have been similarly labeled to agree with the key combinations to which they correspond. Forsyrnmetry of illustration,

all thirty-two horizontal lines in the matrix have been shown, including those that are unuseds- In practice, such lines can be simply omitted, because they serve no subse- Indeed some others of the lines may also be omitted, depending on the coding procedure I. adopted, as will be more fully explained-below.

As shown in FIGURE 3b, a similar switching arrange ment isladopted for control by the relay coils RRI to RRE that the keys of the right hand keyboard control. The switch arms RRll to RRSM shown in FIGURE 3]) serve similarly to connect the source of positive voltage 11 to one or other of thirty-two matrix lines. These matrix lines extend vertically so that each one crosses every one of the horizontal matrix lines. The vertical lines are shown brokenin FIGURE 3b, but appear in i 7 full in FIGURE 3a. Since the; same code system has been chosen for the righthand, each of the vertical'matrix lines will oorrespondin the same sequence to the letter or numeral that is to form the code designation. These characters have been applied to each of the vertical matrix lines in the same way as to the horizontal ones.

Now, before proceeding with the description of the apparatus provided, it will be desirable to explain that, in the sorting of mail by post ofiice authorities, it is conventional to designate a certain number of large centres as DIRECTS. Other major cities in the same country 'as the sorting is being carried out in, areusually desi-gnated DIRECTS, but occasionally very large-cities in other countries and foreign mail generally may also be so designated. The choice of DIRECTS will be arbitrary.

and will depend on the volume of mail for various centres being handled by the sorting post offiee, and by the proportion of such mail which is destined for the major centres. Taking the Ottawa post office as an example,

there arecurrently nineteen designated DIRECTS. Mail which is destined for a DIRECT is coded} accordingly.

Each DIRECT has its own individual bin in the primary sortation stage and after sorting is dispatched directly to Y the centre concerned without furthersorting. Thus, as far 'as'the originating station is concerned, it is unnecessary to impose any code designation on such mail beyond that of the DIRECT itself. I g

All ma'il to centres which are not designated as DIRECTS will require subsequent breakdown. For example, again assuming the Ottawa post oifice to be the originating point, the mail to a small town in Ontario will require firstly to be sorted under Ontario and then sub jected to a secondary sortation in which all the Ontario mail is further sub-divided into individual towns ordistricts. In some instances, a tertiary and even a quaternary I sortati'on maybe necessary to elfect the necessary fineness of breakdown. Since the post ofii'ce system'of Canada is being taken as an example, and since the principal areas into which non-DIRECT mail is divided are defined by provincial boundaries, non-DlRECTS are referred to as PROVINCES. In other countries, this type of mail may well be referred to as STATES or COUNTRIES or DIS TRICTS. The name applied is of no consequence; what distinguishes this mail [from the DIRECTS is the fact that at least one lurther breakdown of such mail is required to be made at the ofiice of origin. I

Each piece of mail is thus lCOdfiCl initially either with a DIRECT or as a PROVINCE code designation. Taking the Ottawa post ofiice system as a further example, there is shown below in Table 2 a list of the DIRECTS proposed llor use in the Ottawa post office, together with the code designations chosen tor each.

TABLE 2 Directs TT Toronto. VV Vancouver. WW Winnipeg. MT Montreal. HA Halifax. HM Hamilton. FF Q British & Foreign. US U.S.A. (other than NYC). OA Ottawa (local mail). GD Government Departments. LD London, Ontario. PB Petenborough, Ontario. QB QuebecCity. IN I St. John, N.B. EO Edmonton. KG Kingston. I-IL Hull, P.Q.- NY New York City. WS Windsor, Ontario.

It will be noted that some of the DIRECTS represent not a major city but a major geognaphical area, such as British and Foreign and United States of America. Such mail will obviously require substantial subsequent to the names of the centres to which they apply. The tables given herein are furnished merely to facilitate an understanding of the description which follows.

Whenever an operator codes a letter, the first thing that he does is to apply the DIRECT or PROVINCE code, as appropriate. This must happen to every letter handled. When the code applied is that of a DIRECT there is no subsequent coding at this point, and that is an end to the matter. When the :code applied is that of 21 PROVINCE, the operator goes on to code the letter further in accordance with additional code designations that have not been listed herein. It follows that the combination of code letters first applied bythe operator must be chosen from one or other of Tables 2 and 3. {Li the operator presses any other combination of letters during the initial coding step, he \has made an error. Many other combinations loan be used later when in effecting a subsequent breakdown of a letter initially coded by PROVINCE it is necessary to code the town, suburb and street name and number. But initially only the few selected code combinations shown in Tables 2 land 3 are breakdown, but for the purposes of the first sortation stage, these areas are treated as DIRECTS.

Table 3 shows a list of PROVlNCES'employed at the Ottawa post ofiice, together with the code designation chosen for each.

TABLE 3 Provinces OX Ontario. QX Quebec. MI Manitoba.

AJ Alberta. 7 PX New Brunswick and Prince Edward Is. CI British Columbia & Yukon. SI Saskatchewan. FX Newfoundland. NJ Nova Scotia.

It cannot be too highly stressed that the choice of the particular destinations that make up the list of DIRECTS and PROVINCES as well as the choice of the individual code designations, is entirely arbitrary. It will vary from one post office to another and may vary from time to time in the same post ofll'ce, asthe volume of mail varies. The code designations are chosen with a view to being companatively easy to remember. They usually include either the initial letter of the city repeated, such as TI for Toronto, or two prominent letters from the word, such as QB for Quebec City. All the PROVINCE code designations are chosen by adopting the first letter of the PROVINCE and following this either with an X or a J. Such a system makes the task of memorizing the code somewhat easier for the operators, but the apparatus to which the present invention relates would be equally applicable if the code designations were entirely unrelated acceptable. All other code designations are unacceptalble. In the examples chosen there are nineteen DIRECTS and nine PROVINCES, which provides a total of 28 acceptable code combinations. Batch of these acceptable code combinations has been shown in FIGURE 3a by a circle or square placed around the point of intersection of the matrix lines corresponding to such combination. The circles represent DIRECT acceptable combinations DC, while the squares show PROVINCE acceptable combinations PC. It will be remembered that the horizontal lines always represent the first letter'of the combination While the vertical lines represent the second letter.

It is necessary to distinguish in the apparatus the matrix cross-over points DC and PC, from all other matrix 'cnoss-over points, that is distinguish acceptables from unacccptables. In the specific example illustnated,'this is done 'by connecting the DC and PC cross-overpoints together in the manner illustrated respectively in FIG- URES 4 and 5. FIGURE 4 shows on a larger scale the cross-over point of matrix lines H and A, representing the DIRECT, Halifax. The vertical matrix line is connected through resistor R10 to a common point 13 with a second resistor R11 which connects such point to the horizontal matrix line. The common point is also connected to ground through resistor R12 and to one electrode of a rectifier R13 and hence to a bus bar DB. When a voltage from source 11 is applied to the horizontal matrix line H only, the voltage at the common point 13 will rise in relation to ground by an amount determined common point 13 will be raised to a higher voltage than when either of the lines is individually energized. "Ibis voltage is transmitted through rectifier R13 to the DI- 0 RECT bus bar DB. As will'be seen later, the lower voltage resulting from energization of only one of th matrix l1nes is arranged to be too low to actuate the swbsequent circuit, whereas the higher voltage acquired through 'energization of both matrix lines is sufiicient t0 etlect subsequent actuation. All the cross-over points DC are similarly interconnected, each having a rectifier R13 which feeds to the common DIRECT bus bar DB. 'I he rectifiers prevent a signal fed from one cross-over point being dissipated by return through the resistors R12 of one of the other cross-over points DC. In an identical manner, the matrix lines at each of the P RO VINCE crosslover points PC, of which the example for Ontario, OX, is shown in FIGURE 5, are joined together and transmit their combined signal to a PROVINCE bus bar PB'separate from bus bar DB.

. nected between the positive source 11 and ground.

thisway a small positive bias is placed on the grid of the The'eilect of a signal on one or other of these bus bars will now be described in connection with FIGURE 6.

This figure shows two triodes DT and PTconnected for the controlgrids of both triodes are subjected to la negativebias supplied (firom a source of negative potential 14' .through contact arms DRCl and PRCl. grids are also connected to their respective bus bars, the

These control control grid of triode PT being directly connected to bus bar PB, while the grid of triode DT is connected to bus bar DB through the normally closed contact arm PRC4.

It will be lhelpful to explain the operation of the system while proceeding with the description of its structure, for which purpose let it be assumed that a positive signal is supplied to bus bar PB from one or other of the PROV- INCE code crossover points PC on the matrix'lZ. If

thisis a low voltage, resulting from energiz'ation of only one of the matrix-lines, it will be insufiicient to overcome the negative bias on the grid and the triode will remain non conducting. If, however, the positive potential on bus bar PE is that produced by simultaneous energization.

to the position opposite from that shown in FIGURE 6.

Contact arm PRCl acts as a holding contact, the grid of and connected through a pair of normally closed contacts Relay PR will actriode PT being disconnected [from the negative source 14 v FBI to the centre tap of a potential divider PD contriodePT, which thus remains in its conducting condition until contacts FBI are opened, regardless of subsequent changes in potential of bus bar PB. Contact arm PRCZ Serves at the same time to connect the positive source 11 through normally \closed contacts 18 .tothe operating coil PM of a printer shown generally at 15. Such printing mechanism 15 will include additional contacts LR'lP to LRSP and RRIP to RRSP, operated rrespectivelyiby the corresponding key-controlled relay coils LRI toLRS and RRl to RRS, to setthe printing bars inreadiness to print the chosen code on the letter which the operator is coding and which is thus currently. situated in the printing mechanism. Such printing bars having been set by the key-controlled relays, the printing operation will r be efiected as soon as the operating coil PM is energized. The particular manner in which this printing operation is carried out is shown only diagrammatically in FIGURE 6, but is further described below in connection with FIG- URE 7. The only other part of the printing mechanism 7 shown in FIGURE 6 is a pair of contacts 19 which are closed by the movement of the frame carrying the type bars as the printing operation is etIected.

The third contact arm PRCSwhich is moved upwardly when, the triode PT conducts, connects the positive sup-- ply 11 to a lfeeding advance coil FA mounted in a .feeding mechanism forming part of the printer 15 land which serves to position the letter properly for such printing. More specific description of one dorm that the printer 15 may take will be described below with reference to FIG- URE 7. Before the feed advance coil FA can be ener gized, the printing operation must be completed and con: tacts 19 closed, the latter serving to complete a circuit to [a feed operating coil FO, normally open contacts F01 of which are connected in the circuit to coil FA; Coil FA when so energized advances the letter one line in readiness for theimposition thereon of a further set of code symbols. It will be remembered that when imposing a PROVINCE coding on a letter; the'letter must remain in the printer fior additional subsequent coding, although mally open contacts 22 which when closed energize a I lamp 23 on the coding desk (see also FIGURE 1).

Before explaining in further detail the manner of operation of these latter parts, the triode ET and. its operation will be described. Its holding contacts DRCl operate in a similar manner to theholding contacts PRCl, connecting the grid of the triode initially to the negative bias, but, once the triode has started to conduct and energized is relay coil DR, connecting such grid to the positive bias so that the triode will'be maintained in conducting condition so long as contacts FEl remain closed Contact arm DRCZ serves a similar purpose to contact arm PRCZ, connecting the printer operating coil PM to the positive supply when the relay coil DR is energized. Contacts DRC3, when closed, connect the positive supply 11 to a feeding-ejecting coil FE of the printer 15. Whereas coil FA advances the letter one position to be in readi ness for another line of coding, the coil FEejects' the printed. letter completely from the printer. This is required because no further coding is applied after a D1- RECT has been coded. Coil FE, like coil FA, only becomes operative when coil F0 is energized, since itssun ply passes through normally open contacts F02.

The practical operation of the system is as follows:

. if one of the selected DIRECT codes is correctly applied to the keyboards by the operator the bus bar DB will be energized to its full voltage. This voltage will pass through contacts PRCdto the grid of triode DT and will be sufiicient to overcome the negative bias on that tube. As a result this tube will conduct and relay coil DR will be energized to reverse the position of contacts DRCl, DRCZ and DRCS. Contacts DRCl will hold the triode ET in its operating condition, not withstanding the fact that the operator may immediately release the keys. Depression of the appropriate keys will also. have actuated the printer contacts LR1P to LRSP and RRlP to RRSP to prepare the correct marking bars in position for printing on the letter. The

printer l5 willbe then operated by its coil' PM which receives voltage through contacts ERCZ. The printing operation will close contacts 319 thus energizing coil F0 and, closing contacts F02. These latter contacts, to.-

gether with contact arm DRC3 will now establish a circuit energizing the feeding-ejecting .coil FE, so that the letter, after having had the DIRECT code markings imprinted on it,will be ejected from the mechanism to make room for the next letter. Energization of coil FE will also open contacts FEl Whichwill then break the supply of positive bias to the grid of triode DT. The latter will then cease to conduct and will revert to its quiescent condition, as will the'remainder of the circuit,

awaiting subsequent operation. Theoperator will know that a correct combination of keys has been pressed,'bythe fact that a printing operation has'taken place and" ejected." A new letter =will'then' invention.

9 and contact arm PRC3 will energize the feeding advance coil FA of the printer as soon as contacts P01 in that circuit are closed by coil PC which follows closing of contacts 19 on application of the code markings to the letter.. Coil FA will advance the letter ready for a second coding operation. The operator will then proceed to code the next aspect of the address, which will normally be the town. The code combination for this town will be different from any of the PROVINCE or DIRECT code combinations and accordingly it will not energize either of the bus-bars PB or DB. For example if the PROVINCE coding were OX (Ontario) the next code might be GA (for the town of Galt, Ontario). GA is an unacceptable combination, but this isof no consequence. The matrix is no longer effective and acceptability is no longer a requirement for operation of the printer, once the first coding step has been completed. This effect is achieved byreason of the triode PT remaining energized, so that the operating coil PM of the printing mechanism also remains energized. Thus, as soon as the operator presses the keys for the next code combination it will be immediatelyprinted regardless of the nature of such combination. The contacts 19 will be closed again, so that coil FOwill be energized and hence coil EA will be energized, The letter will be again advanced in readiness for yet athird coding. As soon as the operator has impressed all the desired stages of code markings on the letter, he will-depress an eject button 24 (FIGURES 1, 2 and '6) which will close contacts arranged to provide direct energization of the feeding ejecting coil PE. The letter will thus be ejected from the printer, contacts FEll opened,

and the circuit returned to its quiescent condition.

Should, however, the operator depress a DIRECT combination after having coded a PROVINCE combination (a series of actions which must necessarily be erroneous), the voltage which will appear on the directbus bar DB will pass through the raised contact arm PRC4 to energize coil 21. This will open contacts 18 supplying power to the printer operating coil PM, thus preventing the erroneous codebeing printed, and at the same time it will close contacts 22 energizing a warning lamp 23 which appears in front of the operator and indicates that an erroneous coding has been attempted. As soon as the operator releases the wrongly pressed keys, coil 21 will be de-energized and the circuit will be ready to receive a proper coding combination, the erroneously selected code combination having been ineffectual. a

If, initially, on the first coding operation for each letter, the operator chooses any code combination which is not one of the selected PROVINCE or DIRECT acceptable code combinations, neither of the triodes PT and DT will be energized and the operating coil PM of the printer will not be energized. Accordingly the keys will be ineffective to impress any code upon the letter. This fact will be apparent to the operator by virtue of the fact that clearly no operation has taken place and the letter has not been fed forward. The operator will then realize that an error has been made and will select the correct code. If desired, a Warning light arrangement could'also be associated with this form of erroneous operation, but it will not normally .be

necessary, as the absence of theinovement associated with a printing operation will be readily apparent to the operator.

In further sup-port of the general method of operation described in FIGURE 6, a simplified, schematic illustration of one possible form of printer 15 is shown in FIGURE 7, although it will be understood. that any other convenient form of printing mechanism may be substituted for the one shown,'since the particular nature of the printing mechanism forms no part of the present FIGURE 7 shows a letter L in printing position'between a glass sheet 40 and a perforated metal plate 41.

The printing mechanism itself comprises a frame 42 carrying ten printing bars L131 to LBS and RBI to RBS each operable by an associated solenoid controlled by respective contacts IRlP to LRSP and RRIP to RREP, which, in turn, are controlled by the key-operated solenoids LR]. to IRS and RRl to RRS (see FIGURE 2). FIGURE 7 shows bars LB LE2, LE4 and RBI extended to print the binary marks corresponding to the code designation MT. With the printing bars thus set by the keys L1 to L5 and R1 to R5, "the printing operation is effected by movement of frame 42 towards the portion of the rear surface of letter L exposed through plate 41. This movement is effected by coil PM cooperating with an armature 43 connected to frame 42. As such printing operation is completed the contacts 19 which are mounted on plate 41 are closed, with the effect already described in connection with FIGURE 6.

Forward movement of the letter L is controlled by an endless belt 44 hearing projections 45 to engage the rear edge of the letter. Belt 44 is driven by roller 46 on shaft 47 which carries one part of a clutch 48, the other part of which is connected to a shaft 49 driven by a motor 50. Gearing 51 drives a further shaft 52 from shaft '49, shaft 52 carrying one part of a second clutch 53, the other part of which is mounted on shaft 5'4 that drives through gearing 55 to shaft 47. Clutch 48 is operated by feed advance coil FA and clutch 53 is operated by feeding-ejecting coil FE, the energization of Which has already been described. Each of the clutches is of the known type that, upon energization, completes one revolution transmitting rotation and then automatically disengages itself. Thus each of clutches 4'8 and 53 will connect the constantly turning motor 5d to roller 46 driving belt 44 for one revolution of the clutch in question, regardless of the shortness of the duration of the impulse energizing coil PA or FE. I

When clutch 48 is closed by coil PA, the single revolu-, tion of shaft 47 that results, advances the letter L the required short distance for the new printing operation to take place beside the first such operation on the surface of the letter. When clutch 53 is closed by coil FE, the single revolution of shaft 5 that results, causes a number of revolutions of shaft 47, sufiicient to propel letter L entirely out of the mechanism.

I In a practical printer, other mechanisms, such as inking rollers and devices for feeding a fresh letter into correct initial position eachtime a previous letter has been ejected from the printer, will be required, but there appears to be no purpose in further describing these mech anisms, as they are entirely conventional and form no part of the present invention. Indeed neither do any of the mechanisms shown in FIGURE 7, but these have been includedin order to render a complete understanding of the operation of the circuit of FIGURE 6.

As has already been explained, the twenty-eight selected code designations for DIRECTS and PROVINCES are arbitrarily chosen. However, once they have been-chosen, certain other requirements must be met in accordance with the preferred form of the present invention, to avoid any possibility of one-finger. errors in respect of such chosen code combinations. This feature is explained with the aid of FIGURE 8 which shows the switches ofthe left hand switching relays LR1 to LRS in position for energizing the matrix line 30 corresponding to the letter M. An example is'being taken of the DIRECT code combination MT for Montreal. The right hand switching relays RRl to RRS are not shown in FIGURE 8 but are assumed to be in position to energize the vertical matrix line 31 for T. These matrix lines are shown somewhat heavier than the other matrix lines in FIGURE 8, since they are being considered for the purposes of the present example as the Principal matrix lines. Now let it'be assumed that the operator makes a one finger error with the left hand. This means that, keeping four of the relay coils in the proper condition corresponding to the letter 1 ll M (that is with relay coils LRL LRZ and'LRd energized) the fifth relay coil is assumed to be in the incorrect posi tion. 7 That is it is assumed that one key is either erroneously operated or erroneously not operated. This is done five times, once -for each key, since five such one-key errors are possible and the switching arrangements produced are such as'to energize the other horizontal matrix lines. designated 32 in FIGURE 8, i.e. those correspond- 'ing to the letters Y, P, X, F and one unused line near the bottomfofthe matrix. This group of lines 32 may be referred to as First Order Associates of the Principal line 30. This is because they represent lines that may be energized with a first order or one-finger error. If the same procedure is carried out for the vertical Principal line 31, First Order Associate lines 33 are produced. It will be apparent that, for each selected Principal line, there will be a different group of five First Order Asso- T ciate lines. There will be some measure of reciprocity in this regard. For example, since the line for P is a First Order Associate for the line for T, when the latter is a Principal line, then the line for T will be a First Order Associate when the ,line P is chosen as a Principal line. But when the line P is chosen as a Principal line the other First Order Associates will not be the same as those which accompany the Principal line T.

" In its broad purpose, the preferred form of the invention sets out to avoid First Order errors in the primary coding step. Ten possible First Order errors can arise,

sinceten fingers are used in the coding. These are represented on the matrix by the five points at which the horizontal Principal line crosses the First Order Associatesof the vertical Principal line and the five points at which the vertical Principal line crosses the horizontal First Order Associates. that all these ten points are unacceptable points, that is to say they do not correspond to ,a DIRECT or PROVINCE code combination. of acceptable code combinations that no one of them By so selecting the set is a combination which could result from a First Order error in one of the other acceptable combinations, all possibility of a one-finger error in the first sortation is eliminated, because any one-finger error must necessarily produce an unacceptable code combination which will not operate the printer.

All the code combinations are chosen in this way. That is to sayfor each combination in which the matrix 1 lines are associated with each other in the so-called .acceptablernanner, there is an associated group of combinations of matrix lines in which said lines are associated with each other in the so-called unacceptable manner, this associated group including every comb-ination of matrix lines that corresponds to a combination i of key operations differing only in respect of a single key from the key combination corresponding to the Principal becausea one-finger error in setting up a combination involving an odd number of keys must necessarily result .in the depression of an even number of keys, and all the even number combinations will have been made un-' acceptable. Similarly, if all the even number keycombinations are made acceptable, there can be 512 of these, the unacceptable combinations then being constituted by all the operations involving depression of an odd num ber of keys. i v

In practice, seldom would as many as 512 acceptable It will be seen from FIGURE 8 oretical maximum in post oflice systems permits a hybrid. choice, that is a mixture of some odd number combina-,

tions and some even number combinations. This mixing of the odd and even types of combination reduces the number of acceptable combinations possible, but permits more latitude of choice, and hence enables combinations to be made acceptable that are comparatively easyfor the operator to remember, by virtue of a similarity to the city concerned. Such a hybrid choice has been adopted in the examples of code combinations given in thev tables above. bination MT (Montreal) is an even number key combination-3 keys (L1, L2 and L4) for'M and 1 key (R1) for T, making a total of 4 keys while the acceptable DIRECT code combination NY'(New York) is an odd number key combinationkeys (L3, L4 and L5) for 'N and 2 keys-( R2 and R4) for Y, making a total of 5 keys. Similarly PROVINCE acceptable combination OX (Ontario) is an odd combination .(3 keys), while PROVINCE acceptable combination an even combination (4 keys). 1 i

All the possible code combinations must fall into one or other of the two classifications or sets that have, tor

QX (Quebec) is convenience, been referred to as acceptable and sun-- acceptable. In the matrix an acceptable combination is distinguished by'the connections which extend from the corresponding matrixelements (FIGURE 4 or FIG- URE 5) to the operating circuit of FIGURE 6, the unacceptable combinations "being simply constituted by a lack of connection between the matrix elements. As will be apparent, this arrangement could be reversed with the unacceptable combinations connected and the acceptable combinations unconnected. Moreover, any suitable method of connection maybe used which will provide the necessary indication. The essential feature is that there should be two different manners of connecting the matrix elements together (one of the manners of connecting usually, being a failure to connect), and that some means should be provided for distinguishing between the simultaneous energization of a combination of matrix elements that are connected together in the first manner and the simultaneous energization of a combination of matrix elernents that are connected together in the other manner.

Moreover, in the specific example given, those code i combinations that are connected in the first or acceptable. manner are further divided into two categories DI- REC .,and.P ROVINCE, between which the indicating circuit can discriminate.

Furthermore, it is to be understood that energization. is here used in the broad sense of a change of condition. Although the most convenient method of operation is to change the selected combination of rnatnix elements, as in the foregoing example, it is apparent that it would be possible to maintain all the matrix elements normally charged and then to discharge the. selected ones. These would I to come under the DIRECT classification, but never- ,theless large enough to be adapted readily for comparatively direct sorting. In coding a letter destined for a For example acceptable DIRECT code com- 1 such a modification.

' to the piece of mail is the two letter code of the SUB- .DI-RECT. To indicate that a SUB-DIRECT code is beingapplied the operator presses a special operating key or bar (not shown). In this modification, for each pro- Mince there will be provided a secondary matrix .(considering the matrix 12 as the primary matrix), and the first two letters of the secondcode applied to the keyboard will be compared in the secondary matrix of the province concerned. This matrix will be set to distinguishSUB-DIRECT destinations (which require no 'further coding) from all other breakdowns (which do require further coding). If the specially operating subdirect key has not been depressed then the secondary matrix will have no effect. But if-this key has been depressed (indicating that the operator intended to apply a SUB-DIRECT code), the matrix will be employed to indicate the comparison that it has made of such code, and specifically to indicate it such code is not in fact ,one of the preselected SUB-DIRECT codes. Inthis way the apparatus furnishes a check againstv a coding error at the SUB-DIRECT stage, analogously with the principa check made-by matrix 12.

, The discriminating mechanism that distinguishes, between energization of the two different manners of matrix element combination may be, as in. the example illustrated, connected to a printer so as mechanically to prevent an erroneous code marking being imposed. It should be appreciated, however, that this aspect of the apparatus .could be simplified if desired, while still remainingwithin theambit of the present invention. For example, the depression of a combination of keys corresponding to an unacceptable code combination could simply illuminate a warning light or sound a warning buzzer to advise the operator that the wrong keys have been depressed, sub sequent reliance being placed on the operator. to make sure that such code marking is not applied to a letter, or

it has been applied, that such letter is deflected to a reject bin for manual'sorting. In such a simplified arrangement, it might be preferable to isolate the matrix elements making up each acceptable combination, while connecting together those of the remaining unacceptable combinations to enable a circuit to the Warning device to be established whenever an unacceptable combina- {tion of keys is depressed by the operator. .Although it is believed that by far the majority of errors result from the operator making a one-finger error, it would be possible, if desired, to so select the codes as also to eliminate all Second Order errors. For example, thesixteen points of crossing of the horizontal First Order Associate lines with the vertical First Order Associate lines shown. in FIGURE 8 each represents a thesam e hand, then it will be necessary to determine the Second Order Associates of each matrix line required to be. a Principal line, and to ensure that the points of cross- .ingiof such Second Order Associates with the Principal line extending in the other direction are unacceptable.

I .This would have to be done for every acceptable crossover point. The Second Order Associate lines would be ,constructed by changing over the position of two switches simultaneously and taking every possible combination of There will be found to be ten Second Order Associate lines for each Principal line. They have not been illustrated.

As will be seen fromFIGURE 8, no attempt has been made in the present system to eliminate all second Order .longer properly be so called.

errors, since the intersection points of First Order Associate lines P and F with First Order Associate line X are acceptable, both as PROVINCE code designations. Nevertheless in any particular system it could readily be done, although it will reduce somewhat the number of possible acceptable code combinations. Since experience has shown that by far the majority of operators errors are First Order errors, the added complexity of also eliminating all Second Order errors would not normally be justified. It will be undestood that, in the apparatus illustrated many Second Order errors are in fact eliminated, although, unlike the First Order errors, not all the Second Order errors are necessarily eliminated. Since only 28 combinations are connected to be acceptable, all other conrbinations, which will include many Second, Third and higher Order errors, will be detected and rejected.

It was mentioned earlier in this description that the UNUSED matrix lines could be omitted, as they serve no purpose. Similarly, all those matrix lines may be omit- .ted on which no acceptable crossover point DC or PC right'hand and one for the left hand. Although this will be the convenient way to operate, the same final result could be achieved with a single, five-key keyboard. The first time this keyboard was operated, a selected combina tion of the left hand relay coils LRl to LRS would be energized and held in their energized positions, and then the second time (the keyboard was used, it would energize the right hand relay coils RRl to RRii. In other words, a

single keyboard could, in this way perform at different times the respective functions of the two keyboards of FIGURE 1. A code combination representing two letters or other symbols is thus produced and fed to a matrix as in the foregoing example.

Moreover, a possible simplified coding system could, in

certain instances, be used as an alternative, with only one code designation employed. This would be equivalent .to the relay coils LRl to LRS controlled by a single, five- .key keyboard L and operating switch arms feeding to the horizontal lines of the matrix. There would be no second keyboard and no vertical lines, so the matrix would no It'wou-ld not have ac ceptable and unacceptable cross-over points, but instead acceptable and unacceptable lines. Supposing all the lines corresponding to an even number of finger operations (that is those in the foregoing example representing the letters X, G, I, P, Y, B, U, I Q, Z, K, F,

W and V) were deemed aceptable," then the remainder would be considered unacceptable. Of course, if few- -er than 14 acceptable lines were required, then this number could be reduced and the number of unacceptable lines increased correspondingly. Electrical connections to these lines couldthen readily distinguish the acceptable set from the unacceptable set.

For example, each of the unacceptable lines could en- Although this arrangement tives, Where x is the number of keys (which is not necessarily five), it nevertheless will embody the important feature of the present invention, the avoidance of all one-finger errors, provided the sets of acceptable .and unacceptable lines are properly chosen in the manner hereinabove indicated.

Throughout the foregoing specification and in the appended claims reference is expressions may imply direct finger operation by an operator, but it is desired to hereby define these terms as includingwithin their scope the binary keyboard-like mechanisms that have been developed in recent years and which j are operated by signals from punched tape or otherstorage devices to setup a binary code designation. Apparatus is,

for example, known, in which the operator is provided with a keyboard like an ordinary "typewriter (one key for each character), but in which the intelligence fed into the machine by the depression of a Single key is automatically converted into a binary-type signal equivalent to the depression of a plurality of keys on a binary keyboard. Accordingly it is within the scope of the present invention'for the binary signal to be generated indirectly in this way by any means for generating a binary signal, suchmeans necessarily including a plurality of individually operable members, whether such members be keys for direct physical operation by the operatorsfingers or whetherdthey be other movable devices controlled in- I directly by some separate selecting instr'umentality.

Moreover, although the term matrix has been used herein torefer specifically to the'cross over arrangement illustrated in FIGURE 3a, it is to be understood that this is merely by way of example and that any suitable produced those whichfall into a first acceptable set from thosewhich fall into a second unacceptable set, thesets being so predetermined. that for any selected combination in the first set there is an associated group of combinations in the second set, such associated group including every combination of said members differing ll claim: 7 I v 7 1. Apparatus comprising means for "generating binary signals includinga plurality of individually operable members, a matrix having two series of elements, first switching means operated by said binary members for connecting an energization source to each one" of the first series of said elements exclusively, the element so connected corresponding uniquely to the combination of operated-binary members, second switching means operable 7 made to binary keys, key-, boards and keybo ar cl means. Without explanation such combination of operated binary members corresponding to the last mentione'd matrix element combination the elements of which are associated with each otherin the said first manner.

2. Apparatus comprising binary keyboard means 'hav- I ing a plurality of individually operable keys, amatrixhav ing two'scries of elements, first switching means operated by said keys for connecting an energizatiou source to each one of the first series of said elementsexclusiyely, the element so connected corresponding niniquelyto the combination of operated keys, second switching means operable by said keys for connecting an energization source to each one of the second series of said'elements exclusively, the element so connected corresponding uniquely to the combination of operated keys, each matrix element of said first series being associated with each matrix element'of said second series in a selected one of two different manners, and indicating means connected to said matrix elements for distinguishing between energization by said first and second switching means of a combination of saidelements associated with each other in the first of said manners and 'energization by said first and second switching meansof a combination of said elements associated with each otherin the second of said manners, the combinations of matrix elements being so chosen that for each combination in which said elements are associated with each other in said first manner, there I only in respect of a single said member"( a so-called onefinger or one-key error) from the selected-combinaton;

is an associated group of combinations of matrix elements in which said elements are associated witheach other in said second manner, said associated group including every combination of matrix elements that corresponds to a combination of operated keys differing only in respect of a single key from't he combination of operated keys corresponding to the last-mentioned matrix element combination the elements of which are associated with each other in the said first manner. I I

3 Apparatus comprising a first binary keyboard for opcrationlby an operators left hand, said keyboard comprising a plurality of individually operablekeys, a second binary keyboard for operation by an operators right hand, said keyboard comprising a plurality of individually operable keys, a matrix having two-series of elements, first switching means operated b the keys of said firstkeyboard for connecting an energization. source to each one of the first series of said elements exclusively, the element so connected corresponding uniquely to the combination by said binary members'ior connecting an energization source to each one of the second series of said elements exclusively, the element so connected corresponding uniquely to the combination of operated binary members,

connected to'said matrix elements for distinguishing between energization by said first and second switching means of a combination of said elements associated with each other in the first of said manners and energization fby said first and second switching means of a combination of each elements associated with each' other in the second of said manners, the combinations of matrix ele- ,ments being so chosen that for each combination in which each matrix element of said first series being associated I with'each'matrix element of said second series in a se- 7 lected one of two difierent manners, and indicating means of operated keys, second switching means operatedby the keys of said second keyboard for connecting an energization source to each one of the second series of saideles ments exclusively, the element so connected corresponding I uniquely, tothe combination of operated keys, each matrix element of'said first series being associated with each,

matrix element of said second series in a selected one of two different manners, and indicating means connected to said matrix elements for distinguishing between simul-. taneous energization by said first and second switching means ofa combination of said elements associated" with each other in the first of said mannersand simultaneous energization bysaid first land second/switching means of a conrbination'of said elements associated with each other J' in the second of said manners, the combinations of matrix 7 elements being so chosen that for each combination in which said elements are associated with each other in said first manner, there is an associated group of cornbinations of matrix elements in which said elementsj'are associated with each other in said second manner, said associatedgroup including every combination ofmat'rix said elements are associated with each other in said first manner, there is anas sociated group of combinations of matrix elements in which said elements are associated with each'other in said second manner, said associated group including every combination of matrix elements that corresponds to a combination of operated binary members differing onl yin respect of a single said member from the elements that correspondsto a combination of operated.

keys difiering only in respect of a single key from the combination of operatedkeys corresponding to the lastmentioned matrix element combination the elements of which are associated with each other in the. said first manner. t a r I 4. Apparatus comprising a first binary keyboard for 17 operation by an operators left hand, said keyboard comprising a plurality of individually operable keys, means for printing a firstcode pattern on each of a-plur-ality'of' articles presented seriatirn to said printing means, the nature of the code pattern so printed corresponding uniquely to the combination of operated keys, a second binary keyboard for operation by an operators right hand, said second keyboard comprising a plurality of individually operable keys, means for printing a second code pattern on each of said plurality of articles, the nature of said second code pattern so printed corresponding uniquely to the combination of operated keys of said second keyboard, a matrix having two series of elements, first switching means operatediby the keys of said first keyboard for connecting an energization source to each one of the first series of said elements exclusively, the element soconnected corresponding uniquely to the combination of operated keys, second switching means operated by the keys of said second keyboard for connecting an energization source to each one of the second series of'matrix elements exclusively, the element so connected corresponding uniquely to the combination of operated keys of said second keyboard, each matrix element of said first series being associated with each matrix element of said second series in a selected one of two difierent manners, and indicating means connected to said matrix elements for distinguishing between simultaneous energization by said first and second switching means of a combination ofsaid elements associated with each other in the first of said manners and simultaneous energization by i said first and second switching means of a combination of said elements associated with each other in the second I of said manners, the combinations of matrix elements being so chosen that for each combination in which said elements are associated with each other in said first manner, there is an associated group of combinations of matrix elements in which said elements are associated with each other in said secondv manner, said associated group including every combination of matrix elements that corresponds to a combination of operated keys differing only in respect of a single key from the combination of operated keys corresponding to the last-mentioned matrix element combination the elements of which are associated with each other in the said first manner.

5. Apparatus according to claim 4 including means connecting said indicating means to said printing means for p-reventing'operation of said printing means upon simultaneous energization by said first and second switchvancing means for operation thereof to advance a said article to receive a further code marking rupon detection by said discriminating means of energization of a said comblnatlon in said first category, means associated with said printing means for ejecting a said article therefrom,

means connecting said discriminating means to said ejecting means for operation thereof to eject a said article from said printing means upon detection by said discriminating means of energization of a said combination in said second category.

-'7. Apparatus as claimed in claim 6, including means for detecting the energization of a said combination in said second category immediately following the energi:

zation of a said combination in said first category.

8. Apparatus comprising means for generating binary signals including a plurality of individually operable members, a matrix having two series of elements, first switching means operated by said binary members for connecting an energization source to each one of the first series of said elements exclusively, the element so connected corresponding uniquely to the combination of operated binary members, second switching means operable by said binary members for connecting an energization source to each one of the second series of said elements exclusively, the element so connected corresponding uniquely to vthe combination of operated binary members, each matrix element of said first series being associated with each matrix element of said second series in a selected one of two different manners, the number of combinations of matrix elements associated with each other in the first said manner being small in relation to the number of vcornbinations of matrix elements associated with each i energization by said first and second switching means of first and second switching means of a combination of said elements of the second such category, said apparatus further including means associated with said printing a combination of said elements associated with each other in the first of said manners and energization by said first and second switching means of a combination of said elements associated with each other in the second of said manners.

9. Apparatus according to claim 8, including printing means for printing a code pattern on each of a plurality of articles, means connecting said printing means to said signal generating means to cause the code pattern so printed to correspond uniquely to the combination of operated binary members, and means for preventing operation of said printing means upon energization by said first and second switching means of a combination of said -matrix elements associated with each other in the second said manner.

References Cited in the file of this patent

Claims (1)

1. APPARATUS COMPRISING MEANS FOR GENERATING BINARY SIGNALS INCLUDING A PLURALITY OF INDIVIDUALLY OPERABLE MEMBERS, A MATRIX HAVING TWO SERIES OF ELEMENTS, FIRST SWITCHING MEANS OPERATED BY SAID BINARY MEMBERS FOR CONNECTING AN ENERGIZATION SOUCE TO EACH ONE OF THE FIRST SERIES OF SAID ELEMENTS EXCLUSIVELY, THE ELEMENT SO CONNECTED CORRESPONDING UNIQUELY TO THE COMBINATION OF OPERATED BINARY MEMBERS, SECOND SWITCHING MEANS OPERABLE BY SAID BINARY MEMBERS FOR CONNECTING AN ENERGIZATION SOURCE TO EACH ONE OF THE SECOND SERIES OF SAID ELEMENTS EXCLUSIVELY, THE ELEMENT SO CONNECTED CORRESPONDING UNIQUELY TO THE COMBINATION OF OPERATED BINARY MEMBERS, EACH MATRIX ELEMENT OF SAID FIRST SERIES BEING ASSOCIATED WITH EACH MATRIX ELEMENT OF SAID SECOND SERIES IN A SELECTED ONE OF TWO DIFFERENT MANNERS, AND INDICATING MEANS CONNECTED TO SAID MATRIX ELEMENTS FOR DISTINGUISHING BETWEEN ENERGIZATION BY SAID FIRST AND SECOND SWITCHING MEANS OF A COMBINATION OF SAID ELEMENTS ASSOCIATED WITH EACH OTHER IN THE FIRST OF SAID MANNERS AND ENERGIZATION BY SAID FIRST AND SECOND SWITCHING MEANS OF A COMBINATION OF EACH ELEMENTS ASSOCIATED WITH EACH OTHER IN THE SECOND OF SAID MANNERS, THE COMBINATIONS OF MATRIX ELEMENTS BEING SO CHOSEN THAT FOR EACH COMBINATION IN EACH SAID ELEMENTS ARE ASSOCIATED WITH EACH OTHER IN SAID FIRST MANNER, THERE IS AN ASSOCIATED GROUP OF COMBINATION IN WHICH MATRIX ELEMENTS IN WHICH SAID ELEMENTS ARE ASSOCIATED WITH EACH OTHER IN SAID SECOND MANNER, SAID ASSOCIATED GROUP INCLUDING EVERY COMBINATION OF MATRIX ELEMENTS THAT CORRESPONDS TO A COMBINATION OF OPERATED BINARY MEMBERS DIFFERING ONLY IN RESPECT OF A SINGLE SAID MEMBER FROM THE COMBINATION OF OPERATED BINARY MEMBERS CORRESPONDING TO THE LAST-MENTIONED MATRIX ELEMENT COMBINATION THE ELEMENTS OF WHICH ARE ASSOCIATED WITH EACH OTHER IN THE SAID FIRST MANNER.

Images