The Stepped Reckoner of Gottfried Leibniz
The great polymath Gottfried Leibniz (see biography of Leibniz) was one of the first men (after Raymundus Lullus and Athanasius Kircher), who dreamed for a logical (thinking) device (see The Dreamer Leibniz). Even more—Leibniz tried to combine principles of arithmetic with the principles of logic and imagined the computer as something more of a calculator—as a logical or thinking machine. He discovered also that computing processes can be done much easier with a binary number coding (in his treatises De progressione Dyadica, March, 1679, and Explication de l'Arithmetique Binaire, 1703).
In the De progressione Dyadica Leibniz even describes a calculating machine which works via the binary system: a machine without wheels or cylinders—just using balls, holes, sticks and canals for the transport of the balls—This [binary] calculus could be implemented by a machine (without wheels)... provided with holes in such a way that they can be opened and closed. They are to be open at those places that correspond to a 1 and remain closed at those that correspond to a 0. Through the opened gates small cubes or marbles are to fall into tracks, through the others nothing. It [the gate array] is to be shifted from column to column as required...!
Leibniz dreamed of inventing the general problem-solver, as well as a universal language—I thought again about my early plan of a new language or writing-system of reason, which could serve as a communication tool for all different nations... If we had such an universal tool, we could discuss the problems of the metaphysical or the questions of ethics in the same way as the problems and questions of mathematics or geometry. That was my aim: Every misunderstanding should be nothing more than a miscalculation (...), easily corrected by the grammatical laws of that new language. Thus, in the case of a controversial discussion, two philosophers could sit down at a table and just calculating, like two mathematicians, they could say, 'Let us check it up ...'
Admittedly the impressive ideas and projects of Leibniz had to wait some centuries, to be fulfilled (the ideas of Leibniz will be used two and half centuries later by Norbert Wiener, the founder of Cybernetics). So, let's ground and examine his famous Stepped Reckoner.
Leibniz got the idea of a calculating machine most probably in 1670 or 1671, seeing a pedometer device. The breakthrough happened however in 1672, when he moved for several years to Paris, where he got access to the unpublished writings of the two greatest philosophers—Pascal and Descartes. Most probably in this year he became acquainted (reading Pascal's Pensees) with the calculating machine of Pascal (Pascaline), which he decided to improve in order to be possible to make not only addition and subtraction, but also multiplication and division.
At the beginning, Leibniz tried to use a mechanism, similar to Pascal's, but soon realized, that for multiplication and division it is necessary to create a completely new mechanism, which will make possible the multiplicand (dividend) to be entered once and then by a repeating action (rotating of a handle) to get the result. Trying to find a proper mechanical resolution of this task Leibniz made several projects, before to invent his famous stepped-drum mechanism.
The upper sketch is from a Leibniz's manuscript from 1685 (the full text is given below) and shows probably early Leibniz's designs for the calculating machine. There is an input mechanism, the lower circles, inscribed Rota multiplicantes, where must be entered the multiplier; there is a calculating mechanism, inscribed Rota multiplicanda, where must be entered the multiplicand; and there is a result mechanism, the top circles, inscribed Rota additionis, where can be seen the result of multiplication. The movement from the input wheels to the calculating wheels is transferred by means of chains. The calculating mechanism is based on pin-wheel, not on stepped drum.
The pin-wheel mechanism is described also on a sketch (see the nearby sketch) from another Leibniz's manuscript, which throw light on his initial idea for the calculating mechanism. The Leibniz' pin-wheel mechanism will be reinvented in 1709 by Giovanni Poleni, and improved later by Braun, Baldwin and Odhner. The undated sketch is inscribed "Dens mobile d'une roue de Multiplication" (the moving teeth of a multiplier wheel).
Obviously the prototype and first designs of the calculator were based on one of the above-mentioned pin-wheel mechanism, before the development of the stepped drum mechanism, which was successfully implemented into the survived to our time devices (the machine was under continuous development more than 40 years and several copies were manufactured).
Starting to create the first prototype, Leibniz soon faced the same obstacles that Pascal had experienced—poor workmanship, unable to create the fine mechanics, required for the machine. He complained: "If only a craftsman could execute the instrument as I had thought the model."
The first wooden 2-digital prototype of the Stepped Reckoner (this is a later name, actually Leibniz called his machine Instrumentum Arithmeticum), was ready soon and in the end of 1672 and beginning of 1673 it was demonstrated to some of his colleagues at French Academy of Sciences, as well as to the Minister of Finances Jean-Baptiste Colbert.
In January 1673 Leibniz was sent to London with a diplomatic mission, where he succeeded not only to met some English scientists and to present his treatise called The Theory of Concrete Motion, but also to demonstrate the prototype of his calculating machine to the Royal Society on 1st of February, 1673. The demonstration was probably not very successful, because the inventor admitted that the instrument wasn't good enough and promised to improve it after returning to Paris. Nevertheless, the impression of Leibniz must had been very positive, because he was elected as a member of Royal Society in April, 1673. It is known also, that during his trip to London, Leibniz met Samuel Morland and saw his arithmetic engine.
In a letter of 26th of March 1673 to one of his correspondents—Johann Friedrich, mentioning the presentation in London, Leibniz described the purpose of the arithmetic machine as making calculations easy, fast, and reliable. Leibniz also added that theoretically the numbers calculated might be as large as desired, if the size of the machine was adjusted: a number consisting of a series of figures, as long as it may be (in proportion to the size of the machine).
Back in Paris, Leibniz hired a skilful mechanician—the local clockmaker Olivier, who was a fine craftsman, and he made the first metal (brass) prototype of the machine. It seems the first working properly device was ready as late as in 1685 and didn't manage to survive to the present day, as well as the second device, made 1686-1694. (Olivier used to work for Leibniz up to 1694. Later on professor Rudolf Christian Wagner and the mechanic Levin from Helmstedt worked on the machine, and after 1715, the mathematician Gottfried Teuber and the mechanic Has in Leipzig did the same).
In 1675 the machine was presented to the French Academy of Sciences and was highly appreciated by the most prominent members of the Academy—Antoine Arnauld and Christian Huygens. Leibniz was so pleased by his invention, that he immediately informed some of his correspondents: e.g. Thomas Burnett, 1st Laird of Kemnay—I managed to build such arithmetic machine, which is completely different of the machine of Pascal, as it allows multiplication and division of huge numbers to be done momentarily, without using of consecutive adding or subtraction, and to other correspondent, the philosopher Gabriel Wagner—I managed to finish my arithmetical device. Nobody had seen such a device until now, because it is extremely original.
In 1676 Leibniz demonstrated the new machine again to the Royal Society in London. Let's clarify however, that this was a small device with several digital positions only. The full-scale workable machine will be ready as late as in 1694.
One of the old machines of Leibniz
It is unknown how many machines were manufactured by order of Leibniz. It is known however, that the great scientist was interested in this invention all his life and that he spent on his machine a very large sum at the time—some 24000 talers according to some historians, so it is supposed the number of the machines to be at least 10. One of the machines (probably third manufactured device), produced 1690-1720, was stored in an attic of a building of the University of Göttingen sometime late in the 1770s, where it was completely forgotten. It remained there, unknown, until 1879, when a work crew happened across it in a corner while attempting to fix a leak in the roof. In 1894-1896 Arthur Burkhardt restored it in Glashütte, and it has been kept at the Niedersächsische Landesbibliothek for some time. At the present time exist two old machines, which probably are manufactured during Leibniz's lifetime (around 1700) (in the Hannover Landesbibliothek and in the Deutsches Museum in München), and several replicas (see one of them in the photo below).
A replica of the Stepped Reckoner of Leibniz form 1923 (original is in the Hannover Landesbibliothek)
The mechanism of the machine is 67 cm long, 27 wide and 17 cm high and is housed in a big oak case with dimensions 97/30/25 cm. Let's examine what is the principle of the stepped-drum (see the lower sketch).
The stepped-drum mechanism
The stepped-drum (marked with S in the sketch) is attached to a four-sided axis (M), which is a teeth-strip. This strip can be engaged with a gear-wheel (E), linked with the input disk (D), on which surface are inscribed digits from 0 to 9. When the operator rotates the input wheel and the digits are shown in the openings of the lid, then the stepped drum will be moved parallel with the axis of the 10-teehth wheel (F) of the main counter. When the drum rotated to a full revolution, with the wheel (F) will be engaged different number of teeth, according to the value of the movement, which is defined by the input disk and the wheel (F) will be rotated to the appropriate angle. Together with the wheel (F) will be rotated linked to it digital disk (R), whose digits can be seen in the window (P) of the lid. During the next revolution of the drum to the counter will be transferred again the same number.
Stepped Reckoner without the cover (image from Leporello album on investigation in Glashütte, 1894
The input mechanism of the machine is 8-positional, i.e. it has 8 stepped drums, so after the input of the number by means of input wheels, rotating the front handle (which is connected to the main wheel (called by Leibniz Magna Rota), all digital drums will make 1 revolution each, adding the digits to the appropriate counters of the digital positions. The output (result) mechanism is 12-positional. The result (digits inscribed on the digital drums) can be seen in the 12 small windows in the upper unmovable part of the machine.
One of the main flaws of the Stepped Reckoner is that tens carry mechanism is not fully automatic (at least this of the survived until now machine). Let's see why. In the next sketch are shown mechanisms of two adjacent digital positions. The stepped drums are marked with 6, the parts, which formed the tens carry mechanism, are marked with 10, 11, 12, 13 and 14.
The tens carry mechanism (© Aspray, W., Computing Before Computers)
When a carry must be done, the rod (7) will be engaged with the star-wheel (8) and will rotate the axis in a way, that the bigger star-wheel (11) will rotate the pinion (10). On the axis of this pinion is attached a rod (12), which will be rotated and will transfer the motion to the star-wheel (10) of the next digital position, and will increase his value with 1. So the carry was done. The transfer of the carry however will be stopped at this point, i.e. if the receiving wheel was at the 9 position, and during the carry it have gone to 0 and another carry must be done, this will not happen. There is a workaround however, because the pentagonal disks (14) are attached to the axis in such way, that theirs upper sides are horizontal, when the carry has been done, and with the edge upwards, when the carry has not been done (which is the case with the right disk in the sketch). If the upper side of the pentagonal disk is horizontal, it cannot be seen over the surface of the lid, and cannot be noticed by the operator, so manual carry is not needed. I however the edge can be seen over the surface of the lid, this will mean that the operator must rotate manually this disk, performing a manual carry.
The mechanism of the machine can be divided to 2 parts. The upper part is unmovable and was called by Leibniz Pars immobilis. The lower part is movable and is called Pars mobilis (see the sketch below).
An outside sketch (based on the drawing from Theatrum arithmetico-geometricum of Leupold)
In the Pars mobilis is placed the 8-positional setting mechanism with stepped drums, which can be moved leftwards and rightwards, so to be engaged with different positions of the 12-positional unmovable calculating mechanism. Adding with the machine is simple—first addend is entered directly in the result wheels (windows) (there is a mechanism for zero setting and entering numbers in the result wheels), second addend is entered with the input wheels in the Pars mobilis, and then the forward handle (Magna rota) is rotated once. Subtraction can be made in a similar way, but all readings must be taken from the red subtractive digits of the result wheels, rather than the normal black additive digits. On multiplication, the multiplicand is entered by means of the input wheels in the Pars mobilis, then Magna Rota must be rotated to so many revolutions, which number depends on the appropriate digit of the multiplier. If the multiplier is multidigital, then Pars mobilis must shifted leftwards with the aid of a crank and this action to be repeated, until all digits of the multiplier will be entered. Division is done by setting the dividend in the result windows and the divisor on the setup dials, then a turn of Magna rota is performed and the quotient may be read from the central plate of the large dial.
Upper view sketch of Stepped Reckoner
There is also a counter for number of revolutions, placed in the lower part of the machine, which is necessary on multiplication and division—the large dial to the right of the small setting dials. This large dial consists of two wide rings and a central plate—the central plate and outer ring arc inscribed with digits, while the inner ring is colored black and perforated with ten holes. If for example we want to multiply a number on the setting mechanism to 358, a pin is inserted into hole 8 of the black ring and the crank is turned, this turns the black ring, until the pin strikes against a fixed stop between 0 and 9 positions. The result of the multiplication by 8 may then be seen in the windows. The next step requires that the setting mechanism to be shifted by one place by means of the crank (marked with K in the upper figure), the pin inserted into hole 5, and the crank turned, whereupon the multiplication by 58 is completed and may be read from the windows. Again the setting mechanism must be shifted by one place, the multiplication by 3 is carried out in the same manner, and now the result of the multiplication by 358 appears in the windows.
In 1685 Leibniz wrote a manuscript, describing his machine—Machina arithmetica in qua non aditio tantum et subtractio sed et multiplicatio nullo, divisio vero paene nullo animi labore peragantur. As the 1685 design was based on wheels with variable number of teeth, not on stepped drum, it is clear that survived to our time devices are later work. In english the manuscript sound like:
When, several years ago, I saw for the first time an Instrument which, when carried, automatically records the numbers of steps taken by a pedestrian, it occurred to me at once that the entire arithmetic could be subjected to a similar kind of machinery so that not only counting but also addition and subtraction, multiplication and division could be accomplished by a suitably arranged machine easily, promptly, and with sure results.
The calculating box of Pascal was not known to me at that time. I believe it has not gained sufficient publicity. When I noticed, however, the mere name of a calculating machine in the preface of his "posthumous thoughts" (his arithmetical triangle I saw first in Paris) I immediately inquired about it in a letter to a Parisian friend. When I learned from him that such a machine exists I requested the most distinguished Carcavius by letter to give me an explanation of the work which it is capable of performing. He replied that addition and subtraction are accomplished by it directly, the other [operations] in a round-about way by repeating additions and subtractions and performing still another calculation. I wrote back that I venture to promise something more, namely, that multiplication could be performed by the machine as well as addition, and with greatest speed and accuracy.
He replied that this would be desirable and encouraged me to present my plans before the illustrious King's Academy of that place.
In the first place it should be understood that there are two parts of the machine, one designed for addition (subtraction) the other for multiplication (division) and that they should fit together.
The adding (subtracting) machine coincides completely with the calculating box of Pascal. Something, however, must be added for the sake of multiplication so that several and even all the wheels of addition could rotate without disturbing each other, and nevertheless anyone of them should precede the other in such a manner that after a single complete turn unity would be transferred into the next following. If this is not performed by the calculating box of Pascal it may be added to it without difficulty.
The multiplying machine will consist of two rows of wheels, equal ones and unequal ones. Hence the whole machine will have three kinds of wheels: the wheels of addition, the wheels of the multiplicand and the wheels of the multiplier. The wheels of addition or the decadic wheels are now visible in Pascal's adding box and are designated in the accompanying figure by the numbers 1, 10, 1(X), etc. Each of these wheels has ten fixed teeth.
The wheels which represent the multiplicand are all of the same size, equal to that of the wheels of addition, and are also provided with ten teeth which, however, are movable so that at one time there should protrude 5, at another 6 teeth, etc., according to whether the multiplicand is to be represented five times or six times, etc. For example, the multiplicand 365 consists of three digits 3, 6 and 5. Hence the same number of wheels is to be used. On these wheels the multiplicand will be set, if from the right wheel there protrude 5 teeth, from the middle wheel 6, and from the left wheel 3 teeth.
In order that this could be performed quickly and easily a peculiar arrangement would be needed, the exposition of which would lead too far into details. The wheels of the multiplicand should now be adjoined to the wheels of addition in such a manner that the last corresponds to the last, the last but one to the last but one, and that before the last but one to that before the last but one, or 5 should correspond to 1, 6 to 10, and 3 to 100. In the addition box itself there should show through small openings the number set as 0, 0, 0, etc. or zero. If after making such an arrangement we suppose that 365 be multiplied by one, the wheels 3, 6, and 5 must make one complete turn (but while one is being rotated all are being rotated because they are equal and are connected by cords as it will be made apparent subsequently) and their teeth now protruding will turn the same number of fixed teeth of the wheels 100, 10, 1 and thus the number 365 will be transferred to the addition box.
Assuming, however, that the number 365 is to be multiplied by an arbitrary multiplier (124) there arises the need of a third kind of wheels, or the wheels of the multiplier. Let there be nine such wheels and while the wheels of the multiplicand are variable so that the same wheel can at one time represent 1 and at another time 9 according to whether there protrude less or more teeth, the wheels of the multiplier shall on the contrary be designated by fixed numbers, one for 9, one for 1, etc.
This is accomplished in the following manner: Everyone of the wheels of the multiplier is connected by means of a cord or a chain to a little pulley which is affixed to the corresponding wheel of the multiplicand: Thus the wheel of the multiplier will represent a number of units equal to the number of times the diameter of the multiplier-wheel contains the diameter of the corresponding pulley. The pulley will turn namely this number of times while the wheel turns but once. Hence if the diameter of the wheel contains the diameter of the pulley four times the wheel will represent 4.
Thus at a single turn of the multiplier-wheel to which there corresponds a pulley having a quarter of its diameter the pulley will turn four times and with it also the multiplicand-wheel to which it [the pulley] is affixed. When, however, the multiplicand-wheel is turned four times its teeth will meet the corresponding wheel of addition four times and hence the number of its units will be repeated as many times in the box of addition.
An example will clarify the matter best: Let 365 be multiplied by 124. In the first place the entire number 365 must be multiplied by four. Turn the multiplier-wheel 4 by hand once; at the same time the corresponding pulley will turn four times (being as many times smaller) and with it the wheel of the multiplicand 5, to which it is attached, will also turn four times. Since the wheel 5 has five teeth protruding at every turn 5 teeth of the corresponding wheel of addition will turn once and hence in the addition box there will be produced four times 5 or 20 units.
The multiplicand-wheel 6 is connected with the multiplicand wheel 5 by another cord or chain and the multiplicand-wheel 3 is connected with wheel 6. As they are equal, whenever wheel 5 turns four times, at the same time wheel 6 by turning four times will give 24 tens (it namely catches the decadic addition wheel 10) and wheel 3 catching the addition-wheel 100 will give twelve hundred so that the sum of 1460 will be produced.
In this way 365 is multiplied by 4, which is the first operation. In order that we may also multiply by 2 (or rather by 20) it is necessary to move the entire adding machine by one step so to say, SO that the multiplicand-wheel 5 and the multiplier-wheel 4 are under addition-wheel 10, while they were previously under 1, and in the same manner 6 and 2 under 100 and also 3 and 1 under 1000. After this is done let the multiplier-wheel 2 be turned once: at the same time 5 and 6 and 3 will turn twice and 5 catching twice [the addition-wheel] 10 will give 10 tens, 6 catching 100 will give twelve hundred and 3 catching 1000 will give six thousand, together 7300. This number is being added at the very same turn to the previous result of 1460. In order to perform as the third operation, the multiplication by 1 (or rather by 100), let the multiplication machine be moved again (of course the multiplicand-wheels together with the multiplier- wheels while the addition-wheels remain in their position) so that the wheels 5 and 4 be placed under 100 and in the same way 6 and 2 under 1000 and 3 and 1 under 10,000, If wheel 1 be turned once at the same time the wheels 3, 6, and 5 will turn once and thus add in the addition box that many units, namely, 36,500. As a product we obtain, therefore:
It should be noted here that for the sake of greater convenience the pulleys should be affixed to the multiplicand-wheels in such a manner that the wheels must move when the pulleys move but that the pulleys do not need to move while the wheels are turned. Otherwise when one multiplier-wheel (e. g., 1) be turned and thus all the multiplicand-wheels moved, all the other multiplier wheels (e. g., 2 and 4) would necessarily move, which would increase the difficulty and perturb the motion.
It should be also noted that it does not make any difference in what order the multiplier-wheels 1, 2, 4, etc. be arranged but they could very well be placed in numerical order 1, 2, 3, 4, 5. For even then one is at liberty to decide which one to turn first and which afterwards.
In order that the multiplier-wheel, e. g., the one representing 9 or whose diameter is nine times as great as the diameter of the corresponding pulley, should not be too large we can make the pulley so much smaller preserving the same proportion between the pulley and the wheel.
In order that no irregularity should follow the tension of the cords and the motion of pulleys tiny iron chains could be used in place of the cords and on the circumference of the wheels and pulleys where the chains would rest there should be put little brass teeth corresponding always to the individual links of the chain; or in place of cords there could be teeth affixed to both the pulleys and the wheels so that the teeth of the multiplier-wheel would immediately catch the teeth of the pulley.
If we wanted to produce a more admirable machine it could be so arranged that it would not be necessary for the human hand to turn the wheels or to move the multiplication machine from operation to operation: Things could be arranged in the beginning so that everything should be done by the machine itself. This, however, would render the machine more costly and complicated and perhaps in no way better for practical use.
It remains for me to describe the method of dividing on the machine, which [task] I think no one has accomplished by a machine alone and without any mental labor whatever, especially where great numbers are concerned.
But whatever labor remains to be done in [the case of] our machine it could not be compared with that intricate labyrinth of the common division which is in the case of large numbers the most tedious [procedure] and [the one] most abundant in errors that can be conceived. Behold our method of division! Let the number 45,260 be divided by 124. Begin as usual and ask for the first simple quotient or how many times 452 contains 124.
It is but very easy for anyone with mediocre ability to estimate the correct quotient at first sight. Hence let 452 contain 124 thrice. Multiply the entire divisor by this simple quotient which can be easily accomplished by one simple turn of the wheel. The product will be 372. Subtract this from 452. Combine the remainder 80 with the rest of the dividend 60. This gives 8060. (But that will be effected by itself in the machine during the multiplication if we arrange in it the dividend in such a manner that whatever shall be produced by multiplication will be automatically deducted. The subtraction also takes place in the machine if we arrange in it the dividend in the beginning; the performed multiplications are then deducted from it and a new dividend is given by the machine itself without any mental labor whatever.)
Again divide this  by 124 and ask how many times 806 contains 124. It will be clear to every beginner at first sight that it is contained six times. Multiply 124 by 6. (One turn of the multiplier wheel) gives 744. Subtract this result from 806, there remains 62. Combine this with the rest of the dividend, giving 620. Divide this third result again by 124. It is clear immediately that it is contained 5 times. Multiply 124 by 5; [this] gives 620. Deduct this from 620 and nothing remains; hence the quotient is 365.
The advantage of this division over the common division consists mostly in the fact (apart from infallibility) that in our method there are but few multiplications, namely as many as there are digits in the entire quotient or as many as there are simple quotients. In the common multiplication a far greater number is needed, namely, as many as [are given by] the product of the number of digits of the quotient by the number of the digits of the divisor. Thus in the preceding example our method required three multiplications, because the entire divisor, 124, had to be multiplied by the single digits of the quotient 365,—that is, three. In the common method, however, single digits of the divisor are multiplied by single digits of the quotient and hence there are nine multiplications in the given example.
It also does not make any difference whether the few multiplications are large, but in the common method there are more and smaller ones; similarly one could say that also in the common method few multiplications but large ones could be done if the entire divisor be multiplied by an arbitrary number of the quotient. But the answer is obvious, our single large multiplication being so easy, even easier than any of the other kind no matter how small. It is effected instantly by a simple turn of a single wheel and at that without any fear of error. On the other hand in the common method the larger the multiplication the more difficult it is and the more subject to errors. For that reason it seemed to the teachers of arithmetic that in division there should be used many and small multiplications rather than one large one. It should be added that the largest part of the work already so trifling consists in the setting of the number to be multiplied, or to change according to the circumstances the number of the variable teeth on the multiplicand-wheels. In dividing, however, the multiplicand (namely the divisor) remains always the same, and only the multiplier (namely the simple quotient) changes without the necessity of moving the machine. Finally, it is to be added that our method does not require any work of subtraction; for while multiplying in the machine the subtraction is done automatically. From the above it is apparent that the advantage of the machine becomes the more conspicuous the larger the divisor.
It is sufficiently clear how many applications will be found for this machine, as the elimination of all errors and of almost all work from the calculations with numbers is of great utility to the government and science. It is well known with what enthusiasm the calculating rods of Napier, were accepted, the use of which, however, in division is neither much quicker nor surer than the common calculation. For in his [Napier's] multiplication there is need of continual additions, but division is in no way faster than by the ordinary [method]. Hence the calculating rods soon fell into disuse. But in our [machine] there is no work when multiplying and very little when dividing.
Pascal's machine is an example of the most fortunate genius but while it facilitates only additions and subtractions, the difficulty of which is not very great in themselves, it commits the multiplication and division to a previous calculation so that it commended itself rather by refinement to the curious than as of practical use to people engaged in business affairs.
And now that we may give final praise to the machine we may say that it will be desirable to all who are engaged in computations which, it is well known, are the managers of financial affairs, the administrators of others' estates, merchants, surveyors, geographers, navigators, astronomers, and [those connected with] any of the crafts that use mathematics.
But limiting ourselves to scientific uses, the old geometric and astronomic tables could be corrected and new ones constructed by the help of which we could measure all kinds of curves and figures, whether composed or decomposed and unnamed, with no less certainty than we are now able to treat the angles according to the work of Regiomontanus and the circle according to that of Ludolphus of Cologne, in the same manner as straight lines. If this could take place at least for the curves and figures that are most important and used most often, then after the establishment of tables not only for lines and polygons but also for ellipses, parabolas, hyperbolas, and other figures of major importance, whether described by motion or by points, it could be assumed that geometry would then be perfect for practical use.
Furthermore, although optical demonstration or astronomical observation or the composition of motions will bring us new figures, it will be easy for anyone to construct tables for himself so that he may conduct his investigations with little toil and with great accuracy; for it is known from the failures [of those] who attempted the quadrature of the circle that arithmetic is the surest custodian of geometrical exactness. Hence it will pay to undertake the work of extending as far as possible the major Pythagorean tables; the table of squares, cubes, and other powers; and the tables of combinations, variations, and progressions of all kinds, so as to facilitate the labor.
Also the astronomers surely will not have to continue to exercise the patience which is required for computation. It is this that deters them from computing or correcting tables, from the construction of Ephemerides, from working on hypotheses, and from discussions of observations with each other. For it is unworthy of excellent men to lose hours like slaves in the labor of calculation, which could be safely relegated to anyone else if the machine were used.
What I have said about the construction and future use [of the machine], should be sufficient, and I believe will become absolutely clear to the observers [when completed].
The second description of Leibniz's stepped-drum calculator, made by Leibniz himself, appeared in 1710, in Miscellanea Berolinensia, the journal of the Berlin Academy of Sciences. It was a 3-pages short description (see the images bellow), entitled "Brevis descriptio Machinae Arithmeticae, cum Figura", and the internal mechanism of the machine is not described.
Brevis descriptio Machinae Arithmeticae, cum Figura
Leibniz did manage to create a machine, much better than the machine of Pascal. The Stepped Reckoner was not only suitable for multiplication and division, but also much easier to operate. In 1675 during the demonstration of the machine to the French Academy of Sciences, one of the scientists noticed that "...using the machine of Leibniz even a boy can perform the most complicate calculations!"
In 1764, forty-eight years after Leibniz's death, a Reckoner was turned over to a clockmaker in Göttingen for repairing. The job wasn't done, and the machine wound up in the attic of the University of Göttingen, where a leaky roof led to its rediscovery in 1879. Fourteen years later, the University gave the machine to the Arthur Burkhardt Company, one of the country's leading calculator manufacturers, for repair and analysis. Burkhardt reported that, while the gadget worked in general, it failed to carry tens when the multiplier was a two- or three-digit number. As it was mentioned earlier, the carrying mechanism had been improperly designed. It's unknown whether Leibniz has designed a machine without the abovementioned flaw.