Comptograph Co. v. Mechanical Accountant Co.

140 F. 136, 1905 U.S. App. LEXIS 4773

• Court: U.S. Circuit Court for the District of Rhode Island
• Decided: August 18, 1905
• Status: Published

Opinion

BROWN, District Judge.

Fetters patent No. 465,255, to D. E. Felt, dated September 15, 1891, are for improvements applicable to recording computing machines, and contain some claims applicable alike to calculating machines which print the result and to machines which do not print, but simply show the result on numeral wheels. The present controversy is now limited to claims 7 and 8, which relate solely to “subtraction cut-offs,” or devices whereby the operator may throw out of operation any one of the carrying mechanisms.

The opinion denying a preliminary injunction is reported in (C. C.) 129 Fed. 394, and it may be useful to refer thereto for explanation of matters dealt with in this opinion. The claims are:

“(7) The combination with numeral wheels, actuating devices therefor, carrying-pawls, and devices for actuating said carrying-pawls to carry the tens, of a device for preventing the operation of the carrying-pawls, whereby the operation of subtraction is accomplished, substantially as specified.

“(8) The combination with numeral wheels, 4, actuating mechanism therefor, and carrying-pawls, 27, of levers, 44, having arms, 46, substantially as and for the purpose specified.”

The important feature of the combination claimed is “a device for preventing the operation of the carrying-pawls,” “levers, 44, having arms, 46.” These will be referred to as “subtraction cut-offs.” Subtraction is performed on certain calculating machines by the addition of complementary numbers. This is because it is undesirable' to reverse the'movement of the numeral wheels. In “additive subtraction,” as it is called, the wheels move as in addition. For example, 6 — -3=3. The same result follows if we add the complement of 3, namely, 7, to 6=13, and mentally disregard or cross off the 1 in the tens column. This is an application of a familiar arithmetical principle. A complement of a number is the difference between thát number and a higher power of 10; for example, the complements of 25 are 75 (100 — 25), 975 (1000 — 25), 9975 (10000— 25), etc. As the arithmetical process of “additive subtraction” is in fact addition of a complementary number, it will, when completely carried out, either with pencil and paper or upon a calculating machine, give a result which, as recorded, is erroneous, the error being a “1” on the left, which must either be disregarded or struck out in order to give the correct result. The error may be avoided by omitting to carry the last “1” on the left.

The specification states:

“Another object of my invention is to prevent the carrying of tens from any column to the next higher whenever a subtraction is to be made by means of adding a complementary number. * * *”

This is done by “subtraction cut-offs,” so called — a series of levers each having a finger piece and an arm to lift a carrying-pawl. Any carrier may at will be thrown out of operation so that it will not move the wheel to the left. It was old to perform additive subtraction on a calculating machine, and to correct the error by mentally disregarding it, or by removing the “1” by adding a 9 to that wheel and all other numeral wheels at the left, a proceeding known as “running off the pines,” or, as suggested by Grant, by turning the wheel back one tooth by hand. The error was well known. It was an incident of the arithmetical principle,-and was not an error in the operation of the mechanism. The error in the result, however, involved the trouble of removing it by striking the nines. Felt devised the “subtraction cut-offs” to obviate the necessity of striking the nines to remove the undesired “1.” Instead of allowing the error to occur and then curing it, he provided means to prevent its occurrence. While, so far as appears, Felt was the first to prevent the ocurrence of the error by stopping the carrier which caused it, it hardly can be held that the claims are to be construed as broad claims for all means for the prevention of this error. The scope of the invention claimed is a narrow one. In any single problem in additive subtraction done on a calculating machine there was an error caused by the left-hand carrier. This was well known, and was provided for in one way. Felt provided for it in another way. At most, he is entitled to patent the means he devised for stopping a carrier, if the means involved invention. He cannot monopolize the principle that prevention is better than cure as applied to calculating machines. The means provided by Felt is a lever having a finger piece projecting on the outside of the machine, and an arm which engages and throws out the carrying-pawl. A single lever is all that is needed in any single problem in subtraction. As a calculating machine must be able to do a variety of sums, and as the subtrahends will contain a varying number of numerals or places, it is apparent that the location of the false carry will vary as the number of places in the sums varies. Therefore the means for correcting the error must correspond with the character of the sums to-be performed. Felt therefore provided a number of subtraction cut-offs — one for each carrier — for the simple reason that the sums to be performed varied, and not because several were necessary in doing a sum in additive subtraction. In doing a sum in subtraction, only one of the cut-offs is used, and, so far as the performance of that problem is concerned, all the other cut-offs might be removed from the machine.

Complainant’s counsel say that “the mechanical performance of additive subtraction is greatly facilitated and rendered mechanically complete by providing means for selectively throwing out of operation any one of the several carrying mechanisms,” etc. .This statement must be carefully scrutinized. The mechanical performance of additive subtraction requires only a single cut-off to correct a single carry. In any particular mechanical performance of additive subtraction the combination of operative parts would include only a single lever. It is doubtless a convenience to have a multiplicity of cut-offs, but a single cut-off is all that can be used in additive subtraction. A particular cut-off is selected by the operator, according to the sums he has to do, exactly as particular keys are selected to put on a given sum for addition. If, as complainant says, the invention is broadly mechanism whereby additive subtraction may be performed without requiring any correction of the remainder, all levers may be disregarded save the one in operation. If the machine were always operated to do the“same sum in subtraction, one cut-off would be enough. It is apparent, therefore, that these cut-offs have no mysterious relation to any arithmetical process, but that one acts separately to prevent the final carry in any sum in additive subtraction. It is claimed that “the invention is, broadly, mechanism whereby additive subtraction may be performed without requiring any correction of the remainders,” and that both the result and the means employed are broadly new with the patentee Felt. If the test of infringement is the performance of the function of preventing the final carry in additive subtraction, the defendant doubtless infringes. But this cannot be the test. It is open to all makers of calculating machines to use any arithmetical principle and to make mechanism operating on that principle. To omit the final carry in doing additive subtraction would be obvious to any mathematician who knew that if he wrote down the final carry he would have to cross it out again to arrive at the correct result.

In Mayo Knitting Machine Co. v. E. Jenckes Manufacturing Co. (C. C.) 121 Fed. 110, affirmed in 133 Fed. 527, 66 C. C. A. 503, we said :

“The broad idea óf perfecting a cam-cylinder machine so that it could perform all known knitting operations, and handle the needles in all known ways, was one that was open to all inventors of specific types of knitting machines.”

As it was well known that the error might be caused by any one of the carriers according to the sum to be done, it was obvious that means must be provided for throwing out the carrier which would make the undesired carry. Mechanically, all that Felt did to prevent the carry was to provide a finger piece to lift up the carrying-pawl. AVhen he did that with a single carrier, and thereby prevented the error in a single sum in additive subtraction, he had completely embodied his conception of preventing the false carry and of mechanism for doing it. The rest was mere repetition of the same device on other carries. There was no co-operation whatever between the additional cut-offs and the first one. The repetition had no relation to any arithmetical operation to be performed upon the machine, but merely to the fact that any of the wheels might be operated in doing such sums as the operator desired to do. It was well known in the art that a carrier could be thrown out of operation by lifting its carrying-pawl through a lever. Patents to Carroll, No. 176,833, Shattuck & Thorn, No. 268,135, Kelso, No. 58,347, and possibly others, show universal cut-offs for throwing out all the carriers simultaneously. As it was known that all the carriers could be thrown out simultaneously by a device which engaged the carrying-pawl of each carrier, it was obvious that each could be thrown out. The defendant may well cite these universal cut-offs to support the propositions that the mechanical problem of throwing out a carrier had been solved in the prior art by a device which threw out not only one, but all, the carriers, and that to provide a lever to engage a carrying-pawl involved no invention.

The defendant, however, devotes a considerable portion of its argument to the proposition that the prior art shows means intended for preventing the operation of carrying mechanisms where problems in subtraction are to be performed. Upon a review of the prior art, however, I am of the opinion that the defendant has failed to show that any one before Felt used or provided means for throwing out carrying-pawls in order to prevent the occurrence of the undesired carry in additive subtraction. It is true that a universal cutoff, or one which throws out all the carrying mechanism at once, is shown in the prior art, and that it is possible to use such a' universal cut-off in doing some sums in additive subtraction. But, on the other hand, it is also true that many sums in additive subtraction cannot be done if all the carriers are thrown out. It is necessary, in certain sums, to operate a part of the carriers, and to throw out only the one carrier which causes the undesired carry. In such case a universal cut-off cannot serve as a proper subtraction cut-off. If a universal cut-off was ever used for subtraction purposes, it was not a suitable device for such purposes, and was not an anticipation of the use of a single cut-off to throw out a single carrier. Nor was the operator able, with such a cut-off, to cut out any one of the carriers according to the sum to be performed.

The defendant’s chief reliance is on the patent to Kelso, No. 58,-347, dated September 25, 1866. In that is shown a rod “which serves to throw the carrying mechanism out of gear, so that each of the wheels can be turned independent of the others.” This is a universal cut-off, which at best could be used only in certain problems, and not in all. It is said, “By the application of the disengaging-slide, g, the operation of setting the wheels back to 0 is materially facilitated, and much time is saved in the operation of the machine.” There is not a word in this patent to inform a reader that this rod could be used in subtraction. On the contrary, however, directions are given for performing subtraction in a manner involving mental carrying from one place to another in the subtrahend, instead of mechanical operations for the correction of the error. Because no reference is made to its use in subtraction, because the patent says expressly what its purpose is, and because the device is not suitable for performing the function of cutting out a single carrier, we must hold that this patent does not show an intention to throw out a carrier to prevent an error in subtraction. But it is said the Kelso specification says that the machine can be used for subtraction. It does not follow, however, that the cut-off is to-be used. Such references as this are insufficient to anticipate. U. S. Peg Wood Co. v. Sturtevant, 125 Fed. 381, 60 C. C. A. 244.

The chief argument to show an intention by Kelso to use a universal- cut-off in subtraction is as follows: Kelso’s wheels have complemental marking for the purpose of subtraction. If the complemental marking is followed, the carrying mechanism should not be permitted to operate in any column. If it does, it will cause the result to be greater by one in each column than it should be. In such case it will be necessary to perform in each column the mental operation of adding 1 to the figure of the subtrahend before striking the key in that column, just as is done in performing the example in the Kelso patent. It is, further said that with all the carrying mechanism thrown out of operation no such mental operations will be necessary, because no errors will be produced by the carrying mechanisms. It is also said: “With all the carrying mechanisms thrown out, no error, of course, can be produced by the carrying mechanism in any column, and there will'be no error anywhere to be corrected, and the machine may be operated directly according to the complemental marking.” It is also said that a universal cutoff is the natural and obvious subtraction cut-off for a complementally marked machine. Reference is made to another patent of Turck’s (No. 679,348) to illustrate this fact, .said patent being, however, later than the patent in suit. This argument will not bear a critical examination. While it is obvious that, if the carrying mechanisms are thrown out, they cannot make a carry, either false or true, it is not true that there will be no error anywhere to be corrected if the machine be then operated exactly according to the complemental marking. In some examples this may be so, but in other examples an error will result, not because the carriers work when not desired, but because they are not at work when they should work in order to give a correct result. It is obviously not true that a universal cut-off is better in a complementally marked machine than an individual cut-off. This may be well illustrated by reading a ' portion of the specification of the Turck patent already referred to (page 7, lines 68 to 108). The operator will then be compelled’ to perform in his head a very much more difficult task than that referred to in the Kelso patent; i. e., adding one to the figure of the subtrahend before striking one in that column. Where all the wheels are deprived of their carriers simultaneously, the machine ceases to have one essential feature.; i. e., the ability to carry such l’s as are necessary for the addition of the complement. In my opinion, the Kelso patent does not directly or by any necessary or fair implication disclose an intention to use the universal cut-off for subtraction purposes, or for any other purpose than to enable the numeral wheels to be returned to zero. This patented device stands exactly like all the others employing universal cut-offs.

The elaborate discussion of complemental and codigital marking, and a discrimination of the uses of a cut-off in codigitally marked machines and in complementally marked machines, does not require extended discussion. It is all subordinate to the defendant’s •contention that a universal cut-off had been used before Felt to prevent carrying errors in subtraction. There is no direct evidence of this fact. The inferences drawn from the complemental marking of Kelso are not necessary inferences, and are offset, if not overcome, by the considerations that it is quite remarkable that Kelso did not refer to such use if he intended it, and that it is more probable that in subtraction he followed the directions for mental operations contained in the patent than that he adopted a method such as that shown in the second Turck patent.

The criticism of Mr. Wiles’ theory of mechanical subtraction is rather a verbal than a substantial matter.

In the arithmetical principle of additive subtraction, it is necessary to add the complement — that is, the difference between that number and a higher power of ten — as we have illustrated. This complement is found by markings on the keys. Neither codigital nor complemental marking, exactly followed, serves to enable one to put •on the true figures of the complement. If complementally marked, we follow only the right-hand figure, and make' a mental correction •of 1 in all the others. If codigitally marked, we make a correction of the right-hand figure, and follow all the others as written. If we make these mental corrections, slightly different in convenience, but similar in principle, an individual subtraction cut-off operated at the left is all that is necessary; and, whichever marking is used with individual cut-offs, a uniform rule for mental correction is possible. With a universal cut-off no uniform rule is possible, and, if all carriers are left out of operation, errors will result, unless mental compensation is made for them.

While the defendant’s showing of the prior art establishes the fact that Felt’s problem was the narrow one of preventing, as distinguished from correcting, a single carry in any sum that it might be desirable to do on the machine, it does not show that any one had tried to do this, or had furnished means whereby it could be done invariably. If it were true that “universal cut-offs” had been tried for this purpose, and if they were in the prior art of subtraction cutoffs, then it might be argued that there was invention in a series of individual cut-offs, since they involved the recognition of the fact that it is often desirable to continue the operation of some carriers while suspending one, and even if it were true that this was of value •only in a codigitally marked machine. But the fact is that, though there are cut-offs in the prior art which show how to throw out all the carriers, and therefore how to throw out a single one, there is no prior art of subtraction cut-offs, and there is no evidence that any •one had tried and failed to accomplish what can be accomplished with individual subtraction cut-offs.

Considering whether the claims show a patentable invention, we have this state of the controversy: Felt was the first to provide a fcut-off to prevent a carrying error. A single cut-off would be a full embodiment of this idea. A series of cut-offs is provided to enable the operator to do the same. thing in different columns. Felt says in his specification, “The complementary number is added, and the tens of the highest series are not carried.” There are, I think, two ideas evidenced by the presence of the series which are not involved in the single cut-off: First, that the place of the error will vary as the sums vary; second, that while one carrier is thrown out the others must be left in. The latter is a mere recognition of the arithmetical principle of additive subtraction. To an improver of a calculating machine familiar with the fact that in doing “additive subtraction” tlie machine is simply adding, it was obvious that, in correcting the error produced by addition, and the single error produced on a codigitally marked machine, the adding mechanism should not be disabled from performing the necessary addition. To attribute to him as an inventive thought the recognition of the necessity of leaving in the carriers at the right i's simply saying that he knew the arithmetical principle of complementary addition, and that he must not violate the principle to which his machine must conform. As, in a knitting machine, there is no invention in thinking of performing any well-known process of knitting, but only in’devising means for so doing, so in a numeral machine, or an adding machine, no one can monopolize the thought of doing additive subtraction upon it.

What te'st may be applied in this case in determining the question of validity? Felt is not entitled to a monopoly of means for correcting this error. The right to prevent the subtraction error is a common right. The right to use any arithmetical principle is a common right. The idea -of throwing out only the carrier that caused the error while leaving in the other carriers in operation was simply a following of the arithmetical principle which requires that in adding a complement all the carries must be made except the one at the left. The idea of providing a series of cut-offs so that any one might be cut off was merely a recognition of the fact that any wheel might commit the error. I am unable, therefore, to find. in the provision of a series of cut-offs any inventive thought, or any idea except of a mere aggregation of similar devices to be used successively. It is quite true that the machine as a whole is made capable of doing correctly any sum in subtraction that may be presented; but the mechanical problem involved in doing this was completely solved by providing a single cut-off and then putting on exact duplicates at all places. The repetition of the cut-offs was merely a repetition of the same mechanical devices, and of- exactly the same idea — a repetition made for obvious reasons, not involving additional inventive thought. Save for the effort of the defendant to show that Kelso’s universal cut-off had been used as a subtraction cut-off, the idea hardly would have occurred that any one who understood the principle of additive subtraction should have thought-of throwing out all the carriers to prevent the operation of one. This contention of the defendant is the only one basis for the argument that there was invention in seeing that some carriers should be left in when one was thrown out. Disregarding the defendant’s contention, and looking at the matter in a natural way, as Felt did, as appears by his specification, it is clear that there was no -invention consisting in his having seen that he must not throw out all the carriers. That a universal cut-off throws out all the carriers is a strong indication that it was not designed for subtraction, in which this must not be done.

The complainant contends that the series of cut-offs is in each claim an element of the combination claimed. The combination claimed is therefore made up of a series of groups, each of which has a carrier and a cut-off for that carrier. Each cut-off co-operates only with the members of one group. There is no mechanical operation in which all carriers of the series operate, nor any arithmetical operation in which more than one cut-off operates.

The defendant contends that the claims are for mere aggregations, not for a true combination of elements. It is apparent, I think, that there is no co-operation or combination of all the elements during the performance of any one problem. I have difficulty in seeing in what way these separate groups may be regarded as combined. To enable the machine to perform whatever sums may be presented, seems to be the purpose served by the presence of the series of cut-offs. The machine thus stands ready for all coming problems, but it is ready for all because it is ready for each one separately, and when made ready for one problem separately by the provision of a single cut-off no more invention is necessary. If A. were the inventor of a single subtraction cut-off on a single wheel, surely B. could not, by using A.’s cut-off on all wheels, claim a new combination as a patentable invention, and support such claim by showing that now, instead of cutting off one error, he could cut off any that might occur.

I am inclined to think that the objection that the claims are not for true combinations is well taken. As the device is novel and useful, however, I am reluctant to place the decision on this ground. As a practical judgment, I should say that what Felt did was, in substance, merely to save the running off of the nines to correct an error, and that he did this naturally and incidentally as a mechanic seeking to improve his machine in details. A little more convenience, the saving of a few strikings of nines, was a result practical and useful, but it was not a remarkable discovery or step in additive subtraction. It amounted to this: Here is the familiar trouble of striking off the nines. It would be well to avoid it. That can be done by throwing out the carrier, and that can be done by lifting the carrying-pawl. This was not a pioneer invention. It was at best an achievement very close to the line between invention and mechanical skill.

But at all events, I am still of the view expressed in the former opinion — (C. C.) 129 Fed. 394 — that:

“It seems hardly reasonable to bold that all computing machines, however different in mechanical operation, shall be tributary to Felt, so far as eonceras correcting mechanically this well-known error, and that no one after Felt shall be permitted to prevent the occurrence of this error by stopping the carrier that causes it.”

I am of the opinion that it is permissible, despite the Felt patent, to prevent this error by throwing out the carrying-pawl, and to-do this for each carrier, and that the Felt claims cannot be construed so broadly as to cover these points; that to hold to the contrary “would block the path of invention to an extent that would be unreasonable.” Colt’s Patent Fire Arms Co. v. Wesson, 127 Fed. 333, 62 C. C. A. 167. The claims, therefore, if valid, which need not be definitely determined, must be restricted to the detailed construction shown.

The complainant’s levers, 44, have each a.finger piece projecting outside the machine, and an arm engaging the carrying-pawl. The defendant has disengaging levers, but they do not project outside the machine, and the defendant does not use any new and additional parts on the outside of his machine. To operate the levers, he connects them with the keys, a half stroke of which serves to throw out the carrier. He dispenses with a part employed by Felt. Upon a view of the case which requires the patent to be restricted to details of construction, this is not an infringement.

The bill will be dismissed.