Opinion
Patent Appeal No. 5021.
June 22, 1945.
Appeal from the Board of Appeals of the United States Patent Office, Serial No. 412,275.
Proceeding in the matter of the application of Edward F. Flint for a patent relating to a device for determining the median of a group of quantitative measurements. From a decision of the Board of Appeals affirming the Examiner's rejection of certain claims, the applicant appeals.
Reversed.
G.A. Ellestad, of Rochester, N.Y., for appellant.
W.W. Cochran, of Washington, D.C. (E.L. Reynolds, of Washington, D.C., of counsel), for Commissioner of Patents.
Before GARRETT, Presiding Judge, and BLAND, HATFIELD, JACKSON, and O'CONNELL, Associate Judges.
The Board of Appeals of the United States Patent Office affirmed the decision of the Primary Examiner in finally rejecting, upon the disclosure of a prior art patent, claims 1, 2, 6, and 7 of appellant's application for a patent relating to a median device, and appeal has been taken here from the board's decision.
Fifteen claims covering the device have been allowed, but the instant claims are broader in scope than those allowed. Claim 1 is illustrative of the subject matter of the appealed claims, and it reads as follows: "1. A device for determining the median of a group of quantitative measurements comprising operating means movable for making said measurements, movable indicating means operatively connected to said operating means and seriately movable thereby in accordance with movement thereof in making each measurement of the group, said indicating means being constructed and arranged to indicate independently the value of each measurement of said group, and means for determining the median value of the group of measurements indicated by the relative positions of said indicating means." [Italics ours.]
The invention relates to a calculating mechanism or device for obtaining the median value of a group of quantitative measurements. In his specification, appellant states that his device may be used with many types of measuring instruments, or may even be used separately from any instrument. The particular embodiment disclosed, however, shows the device adapted to a navigational instrument, such as a bubble octant or bubble sextant. Instruments of that character are used in the art of navigation to determine the altitude of some celestial body. As pointed out in the specification, the customary practice is to make a number of altitude observations or measurements and to determine the median value of the group of measurements, which value is employed in computing latitude and longitude. It is clear from the specification that a device which determines the median value of a group of quantitative measurements performs a function wholly different from any device which merely ascertains the average of such a group.
We have before us one of the commercial instruments built in accordance with the application. Its label states that it is an "Aircraft Octant". It weighs 4 or 5 pounds, has various reflectors, lenses, numeral charts, and a series of disks, all of which function together when used with a battery to determine the median of a group of figures. It is emphasized in the application that for the purposes for which the device is to be used the ascertainment of the average of a group of figures would be useless, whereas the ascertainment of the median value leads to satisfactory results.
The appealed claims were rejected by the examiner upon the disclosure of the Walker patent, No. 1,638,973, issued August 16, 1927. The examiner was of the opinion that the claims did not distinguish over that patent because, in his view, they depend for patentability upon "functions vaguely implied in the word `seriately'". He was also of the opinion that the word "seriately" did not define structure and was therefore indefinite.
The board did not expressly disagree with the examiner, but it went into the whole question, discussed the invention of appellant and the Walker reference in more detail, and showed a thorough knowledge of appellant's contentions and the difference in meaning between the terms "median" and "average". However, it agreed with the examiner that the claims were broad enough to read upon Walker.
The Walker patent relates to an apparatus to be used to measure distances traveled by ships. We deem it advisable here to set out Fig. 1 of that patent and Fig. 9 of appellant's application.
It will be observed that Walker has two sets of three dials each, an upper set and a lower set. These sets operate alternately at intervals of one minute. The pointers of the left-hand dials are electrically connected to the port propeller of the ship, and those of the right-hand dials are similarly connected to the starboard propeller. In each set, the pointer of the center dial is connected with those of the left-hand and right-hand dials through a differential mechanism, so that it indicates the average of their movements. At the bottom of Fig. 1 three registers are shown. That at the left indicates the revolutions of the port propeller; that at the right, the revolutions of the starboard propeller; and that at the center, the average. The two sets of dials are operated alternately so as to permit the conversion of propeller rotation indication to an indication of actual distance traveled by the ship.
Walker was interested only in ascertaining the average of two measurements or indications, namely, those of the revolutions of the two propellers. The device of appellant works upon a wholly different principle. Appellant does not seek to ascertain the average between two movements or measurements. For reasons presently stated, we are of the opinion that the record as a whole shows that the whole purpose of appellant's device is to ascertain the median value of more than two measurements.
The Board of Appeals, in affirming the action of the examiner, stated:
"Appellant argues that the device of Walker operates on an entirely different principle than applicant's device and gains a different end result and that the device of the reference does not provide the median value of the two propeller speeds of a ship but the average of the two speeds as determined by a differential mechanism.
"It is not apparent that there is a difference in so far as the claims are concerned in this attempt to distinguish between median value and average. The Walker patent seems to disclose means for making measurements, that is, for counting the number of revolutions and other means for counting another number of revolutions made in the same length of time as the first but different in number and the means for indicating the average of the two is shown in Fig. 1 of the patent. There is also means for indicating the number of revolutions made by each measurement and it is not apparent that merely stating that these means are movable in series or seriately amounts to a patentable structural distinction. The claims do not seem to be limited to any particular number of measurements so that two different measurements as made by the Walker patent and the mean average of these two as shown by the mechanism disclosed in Walker appears to satisfy the terms of the claims."
In support of the decision of the board, the Solicitor for the Patent Office in this court states, in substance, that the Walker patent does determine a median and is not confined solely to determining an average. He states that one of the definitions of the word "median" given by Webster is "being in the middle; occupying an intermediate position". Then he points to Walker and states that the average reading on the Walker device occupies the middle and therefore is the median. He furthermore argues, in substance, that appellant's device, as disclosed in the specification and claims, does not perform a function which is confined to ascertaining the median of more than two measurements, and that therefore the appealed claims are so broad as to read upon a structure which, admittedly, is designed to determine only averages. He seems to have some doubt that there is a difference between a median and an average, for he makes the following statement in his brief: "Finally, even if `median' were not satisfied by the average of two it is not thought that the broad use of this term in an apparatus claim imparts anything patentable. So far as apparatus is concerned there would seem to be no patentable distinction between indicating a median and an average, assuming them to be different things."
It is clear to us that ascertaining the median of a series or group of more than two numbers or measurements is wholly distinct and entirely different from ascertaining the average; and we think the application of this principle in a composite, workable device, which corrects error and performs a new and useful function, involves invention and is entitled to patent protection commensurate with the scope of the invention.
Appellant, of course, is entitled to such patent protection as he claims in the appealed claims only if they define structure patentably different from the Walker device. The Patent Office tribunals have not complained about the form of the claims. They have not urged that the term "A device for determining the median of a group of quantitative measurements" is merely an introductory clause and not a limitation which may be considered in connection with other features of the claims in determining the patentability thereof. So, we shall spend no time here in discussing the status of the above-quoted clause with reference to its being introductory only, or otherwise.
Appellant urges, and we think properly so, that the claims, calling as they do for a "device for determining the median of a group of quantitative measurements", for "movable indicating means operatively connected to said operating means and seriately movable thereby in accordance with the movement thereof", and for "means for determining the median value of the group of measurements indicated by the relative positions of said indicating means", definitely limit the claimed invention to a determination of the median and eliminate or exclude the ascertainment of merely an average.
It seems to us, therefore, that the first and most important consideration in deciding the instant issue is a determination of what meaning is to be given to the three quoted provisions of the claims. It will be necessary hereinafter to describe the device more definitely in order that the meaning of said provisions may be clearly understood. We shall first consider the meaning of the first and last-quoted provisions.
We here set out the following information from the sources indicated after each quotation:
"median * * * 3. Statistics. Designating a point so chosen in a series, that half of the individuals in the series are on one side of it, and half on the other. To illustrate the distinction between median and average, suppose five persons have wages respectively of $3, $4, $5, $7, and $11. The average wage is $6; the median wage is $5." Webster's New International Dictionary, 2d Ed., 1939.
"median * * * (3) Math. Of or pertaining to that number of a series which has as many numbers preceding as following it." Funk Wagnalls New Standard Dictionary, 1942.
"Median * * * median of a group of measurements.
"The middle measurement, or if there is no middle one, then the one interpolated between the two middle ones. If five students make the grades 60%, 75%, 80%, 95%, and 100%, the median is 80%. [The average would be 82%.]" Mathematics Dictionary, Glenn James, 1943.
"Median * * * 3. Statistics. a. Used to designate that quantity which is so related to the quantities occurring in a given set of instances that exactly as many of them exceed it as fall short of it.
Thus, 6 is the median number of the set 1, 1, 2, 6, 20, 20, 27.
"* * * 1900 Boston (Mass.) Transcript Mar., The average age of the population of the United States * * * is twenty-five years; the median age is twenty-one years. The latter means the point at which there are as many people above as below." Oxford, A New English Dictionary on Historical Principles, 1908.
"Median, the point on a statistical scale of the distribution of cases, above which and below which lie exactly 50% of the cases. The median is thus a measure of `central tendency.' It has the advantage over the arithmetical mean or `average' that it is not affected by unusually high or low values of the variable. For instance, given the values 3, 4, 5, 6, 7, 8, 100, the median is 6, and the arithmetic mean is 19. For some purposes the median value better describes the central tendency of such a series." Encyclopedia Britannica, 1942.
"The median. The median measure may be defined as the middle or central item when the values are arranged in order of magnitude. If there is an odd number of items it is definitely determined. But if the group has an even number of items an additional convention is needed: if the two centrally placed items are distinguishable, a value half-way between them is usually taken as the median. This measure can often be found without resort to arithmetic. For example, the median height of a class may be found by proceeding as follows. Arrange the students according to height, then march them off by twos, the tallest with the shortest, the next to the tallest with the next to the shortest, and so on. The last person or pair of persons will have the median height, which can then be marked on the blackboard.
"Unlike the arithmetic mean, the median is clearly unaffected by extreme variations at the ends of the range. If the tallest student in the room were replaced by a giant, or the shortest by a midget, no change would result in the median. For problems pertaining to wages, gifts, taxes, etc., it is usually more informative to use medians than arithmetic means. If a class of one thousand alumni gave a total of $14,997 to their Alma Mater, the class contributed $15 each `on the average.' But if one person gave $12,000 and the other 999 gave $3 each, the median donation ($3) would be more descriptive of individual contributions than the arithmetic mean." Mathematics in General Education, Prog. Education Ass'n, 1940, p. 124.
"1. Median * * *
"In addition to the arithmetic mean, which has been discussed, two other statistical constants are sometimes used to describe the characteristics of a distribution. If a number of individuals differ with respect to a measurable characteristic, they can be arranged in order corresponding to the order of magnitude of the characteristic observed. The median can then be defined as the value of the characteristic which corresponds to the individual in the middle of the series. There will thus be equal numbers of individuals having forms of characteristics higher and lower than the median. Where the individuals are so numerous that they are grouped in classes, the position of the midmost individual can generally be found by interpolation, if the limits of the characteristic for the group are known. When there is an even number of individuals, the value of the median is taken to be half-way between the values of the characteristic pertaining to the two individuals who stand nearest to the middle of the series." Mathematics for the Million, Hogben, 1943, p. 628.
From the foregoing, it is obvious that in many instances where more than two numerals are involved, a reliance upon the ascertainment of the average of the numerals would lead one far afield, whereas the ascertainment of the median would bring one closer to the desired result. For instance, aside from the various illustrations that are given in the above-quoted authorities, appellant has afforded what we regard as a very pertinent example — one which no doubt is in line with the use to which the device herein may be put. In his specification appellant states that his device may be adapted for use in the art of gunnery in connection with range finding. At oral argument, appellant gave the following illustration: Suppose it was desired to consider a number of shots, say seven, at a vessel on the water, all of which missed. Six of them were close. None were of exactly the same distance from the ship. The seventh was a wild shot, far away from the others. The one who directed the shots, in order to correct his errors, would not want to take the average because the wild shot would, by application of that principle, put him far away from his objective. But if he should apply the principles as have been announced in the foregoing quotations in determining the median, he would, to say the least, find himself fairly close to the objective.
We think that the provision "A device for determining the median of a group of quantitative measurements" in each of the appealed claims and the clause "means for determining the median value of the group of measurements indicated by the relative positions of said indicating means" (the same structural limitation or a similar one is found in each of the claims) imply that there must be more than two numerals or measurements, and that the use of the term "seriately" in the claims also strengthens this view.
The Solicitor for the Patent Office contended at oral argument that the word "group" may include two; that a group may consist of two quantitative measurements; and that therefore the median between two would be the average, as is shown in Walker. It is our view that ascertaining the median of a group of quantitative measurements necessarily involves the ascertainment of that relation when more numerals or measurements than two are involved. In not one of the above-quoted authorities, nor in any others that we have examined, is there any example or anything else suggesting that the median of a set of two may be determined. We probably could stop here, because Walker is concerned only with two measurements, and he determines the average between the two.
The definition of the word "group" may be broad enough to include two, but appellant's specification and other considerations, and inferences to be drawn from the context of the claims themselves, are all convincing that the claims should be so construed as to cover groups of more than two.
According to the dictionaries, "seriately" means in series, or following one after another. The particular part of appellant's delicate device to which this term is directed consists of a series of seven disks, operable one after another. The Solicitor for the Patent Office has definitely and, we think, properly described this feature of the invention. We quote the following from his brief:
"* * * For the purposes of this appeal it is sufficient to refer to Fig. 9 of the application * * *. This figure shows a series of seven disks, 43, which are mounted side by side on the shaft of the instrument. Each disk has a mark 45 on its circumference and, initially, these marks are all in alignment on the so-called fiducial line 45'; and the disks are locked in this position, so that rotation of the shaft does not affect them.
"When the first reading is taken the shaft moves to a position corresponding to the reading and, at this time, the operator manually releases one of the disks which, thereafter always moves with the shaft throughout the series of readings. When the second reading is taken, the second disk is similarly released and so on until the series is completed. Since each disk remains fixed to the shaft at a position corresponding to the reading at which it is released, the marks 45 will indicate the relative values of the readings. If all readings were exactly alike, the marks would be in perfect alignment and for different readings, the marks are spaced by amounts corresponding to the differences. If seven readings are taken there will of course, be one mark 45 which lies in the middle. In Fig. 9 this is the mark designated M, since three of the other marks are above it and three below. This is what appellant calls the median mark * * *."
It will be noticed from this description and from Fig. 9, supra, that the disks of the indicating means are seriately movable; that is to say, the disks move in series, one after another. This feature obviously, if it had no other value in the claims, aids in determining the application to be made to the term "A device for determining the median of a group of quantitative measurements". Nothing in Walker is seriately movable. His two measurements are made simultaneously, not one after the other. The structure defined by the claims requires that the measurements be indicated successively, that is, seriately.
There are other limitations in the various claims to which attention could be called, and which we think might be considered in connection with the instant device as distinguishing over the prior art, but we choose to place our holding upon the broader proposition that in the instant appeal the Patent Office has cited no reference which in any sense performs the function or intended purpose of the instant device as is defined by the appealed claims according to our construction of them. We think the claims are allowable.
In summary, we give the word "group" the meaning which applies to more than two; to "seriately", its natural meaning — that is, the movement of the indicating means must be in series or succession, following one after another. As to the word "group", we think it may properly be said here that if it were not for unduly lengthening the opinion, we would freely quote much from appellant's very elaborate specification which would justify our conclusion that when he employed the word "group" in each of the appealed claims, he meant the term to apply to groups of more than two. Before the board, on petition for reconsideration, appellant proposed by amendment to clarify the meaning of the word so as to limit it to more than two numerals. The board ruled that the proposed amendment was not properly before it and stated, in substance, that appellant's inference that the board would have allowed the claims if so amended was not justified by any language it used in its first decision. Whether or not the said ruling of the board was proper is a question not raised in appellant's reasons of appeal here.
The provisions "A device for determining the median of a group of quantitative measurements" and "means for determining the median value of the group of measurements indicated by the relative positions of said indicating means" necessarily imply that there must be more than two quantitative measurements.
So far as the instant record shows, this inventor may be regarded as a pioneer in the art. It discloses no other device or means of discovering the median as distinguished from the average; and we think the appealed claims, when construed as we have construed them, are allowable.
The decision of the Board of Appeals is reversed.
Reversed.