England and Wales Court of Appeal (Civil Division) Decisions

IN THE SUPREME COURT OF JUDICATURE
COURT OF APPEAL (CIVIL DIVISION)
ON APPEAL FROM CHANCERY DIVISION
PATENTS COURT
MR JUSTICE LEWISON
HC04C03993 – [2006]EWHC 225 (CH)
HC04C03993 - [2006]EWHC 225 (CH)

		Royal Courts of Justice Strand, London, WC2A 2LL
		23rd January 2007

Lord Justice Jacob:

1. This appeal is from a decision of Lewison J, [2006] EWHC 225 (Ch). Ferag sued Müller Martini ("MM") for infringement of its EP (UK) 0 367 715. MM make a machine the "NewsTrim", which it is alleged carries out the process of the process claims of the patent. The Judge held that the machine is outside the scope of those process claims of the patent, that in any event the patent was invalid and that it could not be saved by the amendment sought by Ferag. Ferag appeal these decisions. MM support the Judge's findings in their favour and, if necessary, support the result on grounds rejected by the Judge. The appeal is with the Judge's permission.
2. When things like magazines or books are printed, the sheets are put together and then bound (which includes stapled). The bound "signatures" as they are called have three rough edges which need to be trimmed. I shall call the edges "top", "bottom" and "front". The patent is for a method and apparatus for trimming signatures. Oddly the method claims are wider than the apparatus claims. There is no claim to "an apparatus for carrying out the method of claim 1".

The Law

Principles of claim interpretation

3. The Judge set these out at [24] in the form conveniently summarised by Pumfrey J in Halliburton v Smith [2006] RPC 25 at [68]. There was no dispute about the principles and I need not set them out here.

Anticipation

4. The word "anticipation" has long been used for the objection to validity that the invention is not "new" (the requirement of Arts. 52 and 54 of the European Patent Convention as implemented by the Patents Act 1977).
5. There was some dispute, at least as to the application of the principles. Mr James Mellor QC submitted that in relation to a cited document called Stobb the Judge had erred in his appreciation or application of the law as to novelty. I return to this in due course, for the present confining myself to citation of important principles from the two key authorities and some comments about them.
6. In SKB's Patent [2005] UKHL 59, [2006] RPC 323 Lord Hoffmann said:

"[22]… the matter relied upon as prior art must disclose subject-matter which, if performed, would necessarily result in an infringement of the patent. That may be because the prior art discloses the same invention. In that case there will be no question that performance of the earlier invention would infringe and usually it will be apparent to someone who is aware of both the prior art and the patent that it will do so.

….whether or not it would be apparent to anyone at the time, whenever subject-matter described in the prior disclosure is capable of being performed and is such that, if performed, it must result in the patent being infringed, the disclosure condition is satisfied."

7. Note particularly Lord Hoffmann's phrases "necessarily result" and "must result". This followed his citation of the judgment of the court given by Sachs LJ in General Tire [1972] RPC 457 at 485-486. This includes the following passage:

"To anticipate the patentee's claim the prior publication must contain clear and unmistakeable directions to do what the patentee claims to have invented ... A signpost, however clear, upon the road to the patentee's invention will not suffice. The prior inventor must be clearly shown to have planted his flag at the precise destination before the patentee."

8. Lord Hoffmann later re-emphasised the "necessarily result" test by the familiar test of asking whether, if the prior art were carried out, it would necessarily infringe:

"[24] … anticipation requires prior disclosure of subject-matter which, when performed, must necessarily infringe the patented invention."

9. Compliance with the "necessarily result" test is not in itself sufficient for a piece of prior art to be novelty destroying. The prior art must also be "enabling" – it must provide enough information to enable the skilled man to make or do that which is covered by the impugned claim – see per Lord Hoffmann at [26 – 27]. It is in relation to enablement that the skilled man is assumed to:

"be prepared to display a reasonable degree of skill and common knowledge of the art in making trials and to correct obvious errors in the specification if a means of correcting them can readily be found" per Buckley LJ in Valensi v British Radio [1973] RPC 337, 377 (quoted with approval by Lord Hoffmann at [27]).

10. It is important to keep the concepts of disclosure and enablement separate (see Lord Hoffmann at [28]. The distinction between the concepts is well illustrated by the facts of SKB itself. The attacked patent identified and claimed an identified crystalline chemical. The prior art contained both a description of such a product and a recipe for making it. If the skilled man tried to follow the recipe using his ordinary skill and knowledge he would have failed. The recipe as such was not enabling. But even without it, the skilled man would have been able, with a little trial and experiment, to make the described product. So the prior art satisfied both the "necessarily result" and "enablement" requirements.
11. Note particularly that the permissible "trial and error" concept of enablement is irrelevant to the "necessarily result" concept. To satisfy the latter it is not good enough to show that a skilled man, by making reasonable or sensible adjustments or additions to what is actually described in the prior art, would make or do something within the claim. That would be to move from the objection of anticipation to that of obviousness. This was made clear in General Tire in a passage immediately before the passage I have quoted:

"If, on the other hand, the prior publication contains a direction which is capable of being carried out in a manner which would infringe the patentee's claim, but would be at least as likely to be carried out in a way which would not do so, the patentee's claim will not have been anticipated, although it may fail on the ground of obviousness."

12. Sometimes it will not matter which objection applies. Only in the case of co-pending applications where the earlier application has not been published by the time of the later, is the objection of obviousness not available along with that of anticipation (see EPC Art. 56). But even where both objections are available, it is essential to remember that obviousness involves different considerations from novelty. There can be cases where once the novelty test has failed, an invention will not be obvious even though the differences between the prior art and what is claimed are small and would involve only reasonable trial and experiment or workshop modification. Once the differences are identified, it remains to be shown that what is claimed is obvious. It is not always the case that small variants from the prior art would be so.

Obviousness

13. There was no dispute as to the approach – that of the now familiar Windsurfing case, cited by the Judge at [35]. I need not set it out yet again. It is worth mentioning, however, that once one has identified the difference(s) between the prior art and what is claimed (Windsurfing step 3), the Court must always guard against any ex post facto analysis in considering whether the claim is obvious. As I said in Technip France's Patent [2004] RPC 919:

"[112] Fletcher-Moulton LJ's judgment in British Westinghouse v Braulik (1910) 27 RPC 209 is as true today as when it was first said:

'I confess that I view with suspicion arguments to the effect that a new combination, bringing with it new and important consequences in the shape of practical machines, is not an invention, because, when it has once been established, it is easy to show how it might be arrived at by starting from something known, and taking a series of apparently easy steps. This ex post facto analysis of invention is unfair to inventors'"

Obviousness on appeal

14. Again there was no dispute – the approach is that laid down in Biogen [1997] RPC 1 at 45. Again I will not set it out. I add however, that if the Judge has made an error in his starting point by wrongly construing the prior art so as to find anticipation, then rather less weight can be placed on any finding of obviousness. This was a point noticed in Technip (where the Judge had made an error as to anticipation):

"[114] Here the Judge was really in a difficult position to decide obviousness. For he had already concluded that the claims were anticipated. So, looking at Windsurfing step (3), he had decided there were no differences. His views on obviousness were therefore inherently conjectural."

In the nature of things less weight can be placed on a conjectural view than one which was decisive.

The Common General Knowledge ("CGK")

15. Mr Mellor did not challenge the Judge's findings as to cgk, though he submitted it was not quite complete and failed to follow through some of its consequences. I borrow with gratitude the Judge's summaries, starting with those about the cgk of cutting:

"[15] Both experts broadly agree that there are three kinds of cut: a knife cut, a shear cut and a burst cut. However, a knife may cut in one of two ways: either by itself slicing through the printed product; or by being in cutting engagement with another knife which slices through the product.

(i) In the simple knife cut (using one knife) the knife usually cuts down into a cutting strip or cutting stick set into a horizontal cutting table. The blade must penetrate the strip or stick in order to cut the lowest sheet.

(ii) In a shear cut the paper is subject to shearing load between a first and a second knife (called a "knife" and "counterknife" respectively). Ordinary domestic scissors perform a shear cut. The first knife typically has a narrow wedge angle to enable it to penetrate the stack, whereas the second knife can have a wedge angle between 80^o and 90^o. The first knife must pass close enough to the second knife to avoid the bottom sheet being creased or bent, but does not need to contact it. What is necessary is that the knife and counterknife must be in "cutting engagement". Cutting engagement means that they must be close enough (around 15-20 µm apart) for the product not to get caught between them. It does not mean that they must make actual contact. Indeed, if they are actually in contact, then the continual contact may cause excessive wear on the knives. In addition, it is not necessary for the counterknife to penetrate the product being cut.

[16] A burst cut is most suitable for flexible materials. The material is held between two clamping points. It requires no cutting table or counterknife.

[17] The forces exerted by a cutting knife will tend to move the printed product being cut. To resist this movement it is usual to clamp the product by clamping means located as close as possible to the cutting line of the knife.

[18] In addition a cut may be an "open" cut or a "closed" cut. Punching (which is a closed cut) is a sub-group of the "single-knife cutting principle". An example of a practical use of this punching mechanism is the punch-cutting (punching) of playing cards in order to make the corners rounded. Other examples include the punch-cutting of labels or tabbed, indented books, including dictionaries or large reference books which have the successive letters of the alphabet indented into the pages on the front edge. With punch-cutting, the cut occurs basically at one time, with the blade parallel or near parallel to the product. The cut does not normally occur progressively, with the blade at a large angle to the product. The product and the blade are brought together in parallel and perpendicularly to each other. It is normal for the product to be punch-cut to be fully supported during the cutting process; and that includes both the part of the product that will remain after cutting and also that part of the product that is to be cut away. The familiar hole punch, curiously, operates by means of a shear-cut: the circular opening acting as a counterknife to the punch itself.

[19] On trimming machinery there are two types of blade: plane knife blades and rotary knives. Rotary knives or rotary cutting disks can either have a continuous cutting line, a saw line or several milling cutters.

[21] ….

(i) In most cuts trimming is performed by arranging trimming knives so that cutting occurs progressively from one edge of the paper to the other so as to achieve a steady and not a sudden increase of the cutting force as cutting starts. A progressive cut is a characteristic of a shear cut. It is not a characteristic of a punch-cut.

(ii) Signatures are normally cut at right angles to their surface.

(iii) Where trimming arrangements use a shearing action, the cutting angle can change as the cut proceeds e.g. as in a hand controlled guillotine.

(iv) Trimming requires that the product to be cut is supported. In a shear cut the paper is supported by the combined action of the knife and counterknife. In a knife cut it can be supported by a surface on which the product is situated and in a burst cut by tension in the product.

(v) Clamping of the paper during the cutting motion is an important factor in obtaining a reasonable quality of cut. For example it was generally known that signatures conveyed in shingle streams could be clamped by using top belts and pressing rollers during trimming.

(vi) It is common to trim the head and foot edges simultaneously and the front edge either beforehand or afterwards. Sometimes this is done at different locations, and sometimes at the same location at different times.

(vii) A skilled engineer will arrange the knives in trimming machinery in such a way as to result in lower cutting forces (within the context of the machine he is designing).

(viii) A skilled engineer will choose a cutting angle suitable to his design, and be aware of the range of possibilities as to profile and cutting edge and direction and movement of knife. The ability to design and choose appropriate shear angles is part of the basic training of an engineer in this field.

(ix) In addition to using knife blades with cutting edges, print room trimming was also commonly performed with rotary cutting disks.

(x) A common feature in trimming machines cutting a shingle stream of signatures was for the knives to operate on more than one product at a time. The skilled engineer can adjust the spacing of the signatures to allow more than one to be cut at the same time.

(xi) Where trimming was carried out in a continuous operation (and in particular in in-line rotary trimming) the type of cut employed was a shear cut. It was not generally possible to cut along the full length of all three open sides of a signature at the same cutting station using a shear cut, because of the risk of conflict between blades.

(xii) Accordingly, where trimming of three sides of a signature took place at a single cutting station, the type of cut employed was usually a single knife cut.

(xiii) Knives used to trim signatures become blunt after sustained use and the normal practice was to sharpen them as required.

(xiv) Any steel component to be used as a knife or counterknife will be hardened to ensure low wear and ground to achieve a high quality surface and accurate edge.

(xv) A knife with a straight blade is made of one piece, rather than multiple pieces joined together. This is because the blade must be free of irregularities.

(xvi) It was normal practice to align printed products before trimming."

16. Later the Judge summarised the cgk of machines in use:

"[37] Professor Fritsch described a variety of machines that were in use in the 1980s. Professor Lüthi did not disagree. They were:

a. trimmers used in-line with gatherer-stitchers or as single machines, which cut with two knives per edge cut (i.e. a shear cut);

b. trimmers used in-line with perfect binders or as single machines, which cut with a single knife per edge cut;

c. different types of in-line rotary trimmers for print finishing inline behind web-fed printing presses;

d. sheet cutters (guillotines) for smaller circulations.

[38] Where products were cut in a continuous process, the type of cut used was a shear cut, as with the in-line rotary trimmers."

17. He described a conventional machine in-line rotary trimmer in more detail at [39-40]:

[39] … The printed products enter the machinery in shingle flow. They are aligned; and then pass between two rotary cutters, which cut the head and foot respectively. Each cutter consists of a knife and a counterknife; one of which rotates clockwise and the other anti-clockwise. At the moment of cutting the peripheral speed of the counterknife corresponds to the speed of advance of the shingle flow. The other rotary knife moves much faster. The same process would work equally well with products laid end-to-end or spaced, as opposed to being in shingle flow. Having been cut at both the head and foot, the product moves onto a second conveyor, at right angles to the first, thus exposing the front edge for cutting. The front edge is cut at a second cutting station, again by a rotary knife and counterknife. Although all the pairs of knives and counterknives rotate, they are fixed in position. They move together, rather than independently of one another. This kind of machine requires a "bump turn" between the first and second cutting stations.

[39] The typical layout of this kind of machine looks like this"



18. Mr Mellor pointed out some consequential matters of considerable significance which the Judge did not articulate expressly though they were part of the cgk and largely follow from the matters expressly identified.

i) Shear cutting when used for trimming always involved cutting the edge right off;

ii) This was because actually making a knife and counterknife to the high degree of precision required would pose technical problems. You do not want them actually to touch (unlike a pair of domestic scissors or garden shears) but you must get them closer than the thickness of a sheet of paper. Anything other than a straight edge is too difficult, save for a circular hole such as is made by a hole punch.

iii) It was also because a straight edge makes disposal of the trimmings easier.

iv) So in all prior-art two-knife (i.e. shear cut) systems cutting takes place at different locations whereas in all one-knife systems cutting takes place in the same location (but at different times).

19. Mr Mellor also pointed out that the Judge had made an error in relation to punch cutting when he said the blades were "parallel or near parallel." The blades can be arranged so as to cut progressively, as in a hole-punch. Mr Mellor showed us a standard text book picture showing the cutting of playing cards with a progressive punch-cut:



The point here (relevant to Stobb) is that a diagram showing an angled blade is not a necessary indication of the type of cut.

The Patent in Suit

20. A masterpiece of draftsmanship this is not, even allowing for the fact that we are working with a translation from the German. Nonetheless with reasonable study and with the knowledge of the cgk firmly in mind, it is possible to make sense of it.
21. It begins with a statement of what the problem is:

"[2] Multilayer, particularly printed products, e.g. journals folded one or more times, generally have to be cut on at least one, but usually two or three sides. In industrial in-line production of such printed products, which are generally conveyed away from the rotary press in the form of a scale flow, the aim is to integrate the cutting process into the dynamic production process, i.e. to cut at the full rotation capacity of e.g. 80,000 copies per hour."

22. This is not unimportant because of the speed mentioned. The printing presses and "binder" can turn out signatures at 80,000 per hour. The problem is to provide a trimming machine which can keep up. The "scale" or "shingle" flow on a conveyor is what comes out of machines running at this speed. How do you cut not only the sides but the front?
23. The Patent goes on to elaborate on the problem, describing first some well-known matters and then Ferag's own previous proposal.

[3] Numerous cutting mechanisms are already known. Thus, e.g. CH-A-650 967 and EP-A-17 878 in each case disclose a mechanism for the lateral cutting or trimming of paper sheets, in which the latter occurring in scale flow form are conveyed on a planar conveyor belt against one or more rotating cutting knives and are laterally cut by the latter. Although in this way generally acceptable cutting results are obtained, said mechanisms suffer from certain disadvantages. Thus, it proves to be difficult and costly to accurately orient the individual printed products in the scale flow. In addition, in the scale flow there is a gap between the individual products and the substrate, so that particularly when thicker material has to be cut the cut edges are torn, particularly the first sheets and irregular cutting tracks occur. If the products are not only to be cut on the sides parallel to the conveying direction, the scale flow must be turned by 90^o or the printed products must be individually turned, which is prejudicial to an optimum spatial arrangement of the production line and is technically complicated.

[4] The result is that a cutting or trimming in scale flow has a number of disadvantages and generally does not satisfy the requirements for high-quality cutting, so that the Applicant has already proposed solutions, in which the printed products are not cut in scale flow.

[5] A further apparatus according to CH-A-583 611, with which the printed products are individually cut, comprises a rotating cell or bucket wheel, whose individual cells or buckets are in each case intended for receiving a printed product and are provided with movable knives operable by means of common control cams and counterknives cooperating therewith. The construction of this apparatus is relatively complicated and it is therefore costly to maintain.

24. So you had the problem of a "bump" turn and other difficulties, or, by Ferag's 611 proposal you did away with that but in return had a complicated system of pairs of knives and counterknives associated with each cell of a bucket wheel.
25. The patent then summarises the position, again stressing that the problem arises with the kind of speeds with which the invention is concerned:

" [6] Thus, the known cutting technologies in the case of dynamic high-performance product processes do not meet the quality requirements or only do so at the cost of high constructional expenditure. However, the same cutting qualities as in static or alternating processing processes are not obtained, the latter having highly developed cutting technologies with which by a suitable guidance of the cutting forces excellent results are obtained. Thus, e.g. entire paper packs, which rest on a fixed substrate acting as the counterknife, are cut by a swinging or dropping knife, which cuts the product on an accurately defined path (e.g. logarithmated shape) with excellent accuracy."

26. It then says it has solved the problem – the same problem as Ferag's earlier proposal with complicated pairs of shear knives was aimed at:

[7] The aim of the present invention is to obviate this problem. On the basis of the prior art as known from CH-A-583 611, the set problem is to provide a reliable and accurate method and a simple, maintenance-friendly and inexpensive apparatus, with which a high-quality cutting of multilayer printed products in a continuous process and in particular in a high-performance production process is possible. This problem is solved by the characterizing features of the independent claims.

27. It then has a very important paragraph. The first part notices that all prior systems involve a knife and counterknife as a pair. It says:

"[8] In conventional dynamic cutting processes, as stated hereinbefore, either the printed products are supplied continuously or individually to a cutting mechanism, in which the knife and counterknife are fixed, or there are a plurality of cutting mechanisms at least temporarily moving with the printed product and also comprising a knife and a counterknife. These known cutting mechanisms, no matter whether they are fixed or moved together, consequently always comprise substantially two parts, namely a knife and a counterknife, which are associated with one another pairwise and in fixed manner.

28. Putting this another way, you either have a fixed station where in effect the pair of shear cutting knife and counterknife wait for the signature to be brought to the station, or you have pairs on the move which at least temporarily (obviously for the time of cutting) move with it.
29. The paragraph goes on:

"[8] The invention takes a new route, in that it breaks with the conventional knife – counterknife unit and the two knife parts are positioned and moved independently of one another, one knife part being fixed relative to the material being cut and is only temporarily brought into a clearly defined cutting engagement.

30. This is the heart of the invention: do not use the prior art "pair" of shear cutters system. Use independent knives and counterknives. One of these is "fixed" in relation to a particular signature and the other "temporarily" is brought into "cutting engagement."
31. The paragraph concludes:

"Thus, a cutting method comparable to the aforementioned static method is integrated into a dynamic process."

This is saying that at the moment of cutting you have in effect got what happens in the static processes.

32. The point is elaborated in the next paragraph:

"[9] The essential advantages of the invention are that on the one hand the throughput required for a dynamic high-performance process is obtained and simultaneously through the knife parts fixed relative to the material being cut and preferably the counterknife, a cutting quality comparable with that of static cutting methods is obtained. Through a suitable choice and arrangement of the second knife parts (preferably the cutting knives), it is possible, as a function of the characteristics, size and thickness of the printed products to be cut, to fix the cutting curve much as in the aforementioned static cutting mechanisms in such a way that an optimum distribution of the cutting forces occurs over the entire cut edge. Thus, the qualitative advantages of a static and the performance advantages of a dynamic cutting process are combined. In addition, through a corresponding arrangement of three cutting knives, the printed products can be cut on three sides in a single path, without requiring any turning or deflection of the product flow."

33. Later the patentee defines more fully what he means by "fixed":

"[22] The term "fixed" used hereinbefore in conjunction with cutting knives and rollers naturally does not exclude the possibility of the cutting mechanisms being adjustable with respect to the drum rotor, as a function of the printed product and the desired cut. It in fact means that there is no change to the position of the cutting mechanisms during operation.

It is true that this is said in the context of his description of the specific embodiment, but clearly the notion of "fixed" there given is intended to apply to the general concept of the invention. You get one knife part "fixed" in relation to the signature and the other part (which is not just a counterpart of a pair) then cuts. Always there is a shear cut (2-knife cutting) and the cut is from one side of the signature to the other so that the trimmings are in strips.

34. The patent illustrates the method of the invention by some very schematic diagrams, 1a and 1b:



It says:

"[11] Figs. 1a and 1b show the principle of the method according to the invention. In Fig. 1a the spatial sequence is considered from the side, whilst Fig. 1b, for the better understanding thereof, shows the sequence from above. Phase 1 e.g. represents the printed products 2 e.g. conveyed in the form of a scale flow 1, i.e. partly overlapping one another and such as are conventionally conveyed away from a rotary press. The scale flow is broken up and with each individual, or as shown in the drawing, plurality of printed products together is associated a first knife part 3 concomitantly moving with the printed products (phase II). Obviously the printed products conveyed in the scale flow 1 can in each case comprise several elements. The printed products designated 2 in the drawing can consequently comprise several individual products, e.g. superimposed journals or interinserted parts of a newspaper. When reference is subsequently made to a "printed product" this fundamentally covers one or more, possibly only temporarily unit-combined individual products.

[12] The printed products and associated first knife part 3 are oriented along a cut edge 4 and engaged (phase III) and moved past a second knife part 5. The first knife part 3 and second knife part 5 are brought into cutting engagement, so that they cut or trim the printed products 2 along the edge 4 (phase IV). The second knife part 5 is shown in Fig. 1 as a fixed, rotary cutting knife. "

35. It is worth dwelling on what is described. The "signatures" come from the press in shingle flow. There are unassociated knife parts 3 and 5. The latter is a fixed rotary knife. The signatures are broken into groups of three (2). For each group a knife part (3) comes along behind the group and catches it up by the time of arrival at the counterknife where there is "cutting engagement." Note particularly that it is not said that the knife part 3 must itself provide the clamping necessary for a shear cut.
36. This is unlike the prior art. There is no longer a pair of knives (knife and counterknife) waiting for the signature. Nor is there a one-on-one pair of knives, travelling or not. There is no fixed relationship between the knives.
37. The patent goes on to say:

"[13] It obviously falls within the scope of the invention to give a different shape and function to the second knife part. Thus, the second knife part can e.g. be constructed as a stationary, non-rotating blade or as a movable, non-fixed knife part."

What is meant by a "movable non-fixed knife", given that claim 1 calls for movement past "a second fixed knife part"? To this I must return when construing claim 1.

38. It adds other points about possible variants:

[13] ….. Although the rotating knife part 3 is preferably constructed as a counterknife and the knife part 5 as the actual cutting knife, their functions are naturally also interchangeable. Thus, it is unimportant for the purposes of the described method, whether the printed products are conveyed in as a scale flow or in any other arrangement. Thus, the situation can be such that the printed products are supplied individually or in groups from a preceding working process."

39. The patent then turns to its specific embodiment. I borrow Mr Mellor's description based on fig. 2:



"This shows that as they enter the cells 12 at location A, the printed products 2 are travelling at a completely different velocity to the first knife parts 31. Having entered the drum, the printed products 2 continue to move towards the centre of the drum, and therefore in a different direction to the knife parts 31. After being cut by the fixed blade 15, the printed products immediately start to move outwards (away from the centre of the drum, and therefore in a different direction to the knife parts 31) so as to be cut by blade 16. As they exit the cells at location B, the printed products 2 are travelling at a completely different velocity to the first knife parts 31."

40. In the embodiment the cell edges are knife edges. Inherently there must be a very small clearance between the edges of the cell and the fixed knives 14, 15 and 16 to achieve two-knife cutting. The operation is not linear as shown schematically in figs. 1 and 2, but that does not matter: essentially the same process is carried out, with each product meeting up with one knife part before cutting engagement with the other. In the picture one can see the products completing their fall into the "bucket" just before they are first cut.
41. In passing one may wonder what stops the products falling out. The answer for this embodiment is almost certainly unshown clamps – the patent contemplates the use of these – see e.g. claim 15 and the passage quoted at [54] below. This however makes no difference to the principle of what is being shown.

Claim 1 and Infringement

42. It is not necessary to say more about the specific embodiment at this stage. I can turn to claim 1 in its original, unamended, form. I borrow Mr Mellor's numbering:

"(1) Method for cutting or trimming continuously conveyed, multilayer printed products in a continuous process,

(2) in which at least one first knife part is jointly associated with each one or several printed products,

(3) in which at least one first knife part and the associated printed product are moved at substantially the same velocity and are engaged with one another along at least one cutting edge,

(4) characterized in that the first knife part and the associated printed product are moved past a second, fixed knife part,

(5) in order to be brought into cutting engagement therewith, so that the printed product is at least cut along an intended cutting edge."

43. A number of points of construction arise. They are to be answered by application of the Kirin-Amgen principles. Whilst of course the claim is not to be construed with a view to making it read on, or not read on, to the alleged infringement, it is helpful to understand the issue in the context of the alleged infringement. So I turn to that first.
44. The Judge described the NewsTrim as follows:

"[122] The following description is taken from the product and process description. A conveyor rotates around an axis in an anti-clockwise direction. The angular speed of the conveyor is steady. Each printed product is held between a pair of leading and trailing clamp pieces. The products move as the conveyor rotates. Each leading clamp piece is attached by a spring to a point inward and ahead of it on the conveyor. Accordingly, when the leading clamp piece follower rolls are free the printed products are clamped between each leading clamp piece and its corresponding trailing clamp piece. Each pair of leading and trailing clamp pieces is pivotally journalled on a pivoting bolt. The centres of these bolts are disposed at a constant radial distance from the centre of the conveyor and are equally spaced at intervals of 20°. Each trailing clamp piece has a follower roll at the end of a lever arm. The rolls are constrained by cam surfaces, which together comprise a guide track. This causes the leading and trailing clamp pieces (and the printed products held between them) to rotate into different angular positions with respect to their respective radial directions as the conveyor rotates.

[123] The trimmer comprises six rotating knife parts which are disposed on a circumference at equal 60 ⁰ spacings. The rotating knife parts of the trimmer rotate clockwise about an axis. Each printed product makes contact with a rotating knife part at about 10 o'clock. The trimmer includes a single fixed knife part. The inner edge of the rotating knife part passes the fixed knife part at approximately 11 o'clock at which position the printed product is trimmed at its end by being cut between the two knife parts. The drawing in the PPD is partially reproduced below."

45. The Judge then sets out an agreed drawing. It is actually in colour and will be difficult to follow if not so reproduced.



46. The Judge continued:

"[124] Thus if we follow the progress of a printed product, it is shown clamped in position 1. The clamp moves anti-clockwise through positions 5 and 9. Meanwhile the rotary counterknife is rotating clockwise. The fixed rotary knife is rotating anti-clockwise. The counterknife and the printed product come into contact just before position 13; and the cut takes place between positions 13 and 14. The speed of the rotating counterknives is constant. However, the speed of the printed product varies during this process. Between positions 1 and 10 it moves quickly. The purpose of this is to allow the printed products to fan out so that they can be placed more easily between the rotating knife parts. Then the printed products are slowed down suddenly, so that by the time of the cut itself they are moving at the same (or almost the same) speed as the rotating counterknives.

[125] A pair of side trimmers are placed at a different position on the circumference of the conveyor, usually upstream. One trimmer is in front of the drawing, the other behind it. As each printed product passes the side trimmers it is caused to move to the positions 1 to 17 as shown on the drawing. The side trimmers have their axes parallel to the plane of the drawing and to a radius corresponding to position l3 of the printed product, and so perpendicular to the axis of the conveyor. The distance between the fixed knives on the two side cutters is equal to the width of the product after trimming."

47. The Judge added this comment:

"[126] The essential difference between NewsTrim and the patent in suit is that in NewsTrim the counterknife does not travel with the printed product. The printed product is conveyed by a separate clamp. Of course the counterknife makes contact with the printed product at the time of the cut; otherwise the cut could not take place."

48. He had construed the claim at an earlier part of his judgment, giving a meaning to the words "jointly associated" in element (2) of the claim. He said:

"[78] The description of the process in the specification says that during Phase II the first knife part is "concomitantly moving" with the printed products. This is the phase in which the knife part "catches up" with the printed products. In Phase III (i.e. before cutting) the printed products and the first knife part "are … engaged". The use of the word "engaged" must mean (at least) that the first knife part and the nearest of the printed products are touching. Indeed the products must be clamped in the cell before being cut. In Phase IV both the printed product and the first knife part "are moved" past the second knife. The use of the passive voice both here and in claim 1 suggests that they move together, rather than independently. The basic idea behind the invention is that the two knife parts move independently of each other, and that the first knife part is "fixed relative to the material being cut"; and that the cutting engagement is temporary. This is in contrast to known machines, where the knife parts move temporarily with the printed product, but where they are fixed relative to each other. It is in my judgment clear that the word "associated" requires a physical association; and not merely predictability.

[79] The second area of dispute was whether the association must last longer than the mere moment of cutting. In my judgment it must. One of the purposes of the cell is to hold the printed products in a clamp. The product must be aligned before cutting and then re-aligned (still in the cell wall) before the final cut. The patent teaches that the first knife part must be "fixed relative to the material being cut" which itself suggests that the association is more than merely momentary. It contrasts this with the cutting engagement, which is only temporary. In addition the claim requires that the first knife part and the printed product "are moved at substantially the same velocity and are engaged with one another along at least one cutting edge". This is clearly a reference to Phase III, since it precedes the movement past the second knife. The requirement that the velocity of the knife part and the printed products be the same means that they must be moving at the same speed and in the same direction. The claim goes on to require that the first knife part and the associated product are "moved past" a second knife. Thus the physical association (same speed and direction) lasts from at least Phase III until the product has gone past the second knife, which is necessarily after it has been cut."

49. Mr Mellor's criticism of this is that it misses the key idea of the patent. He says that the association required is a "logical association": that all that is called for at this point in the claim is that the first knife be allocated to a particular product or group of products as being one of the knife parts destined to cut it. No actual physical contact between the knife part and the product is required by this element of the claim.
50. I think that is right. It is consistent with both the purely schematic figs 1 and 2 and the specific embodiments e.g. Figs 3 and 4. In both these cases the association between the first knife part (3 in figs 1 and 2) starts without physical contact. Purposive construction requires one to consider the purpose of the words used as they would be understood by a skilled man trying to find out what the inventor was intending to convey in defining his monopoly. The skilled reader would see no purpose in requiring physical (i.e. touching) association until the moment when cutting begins.
51. On this basis the question of the exact period of "association" becomes irrelevant. During the period of the association element 3 of the claim requires that the first knife part and product(s) at some point move at substantially the same velocity (i.e. speed and direction) and engage along at least one cutting edge. Elements 4 and 5 require that the associated knife part and product are moved past the second, fixed knife so as to cut. It is self-evident that during the cut the association must continue and the skilled man would surely also say that it is only during the cut that the engagement and substantially same velocity requirements are essential.
52. All this is consistent with the key idea of the patent, that identified as the "new route" in [8] (cited at [29] above). The statement there, that the fixed knife part "is only temporarily brought into a clearly defined cutting engagement" would surely guide the skilled man's approach to the meaning of the claim – one which focuses on the actual time of cutting.
53. It is also consistent with claim 3. This is to a method according to claim 1 "characterized in that the printed products by the at least one first knife part associated therewith are moved past at least one second knife part." The grammar is awful but the sense is clear: the claim introduces the requirement that it is the first knife part which moves the product past the second knife. So by implication claim 1 is not so limited, yet on the Judge's construction it is.
54. Mr Mellor also submitted that the Judge had focussed too closely on the specific embodiment. I think there is some force in that, though there is no need to pursue it in detail. For instance the Judge said in [79] "one of the purposes of the cell is to hold the printed products in a clamp". But claim 1 does not call for a cell at all. Or a clamp. And even the specific embodiment makes clamps optional:

"[18] ….The printed products can also be so fixed in their cells by a clamping mechanism, that they cannot be further displaced until they have passed the second cutting knife 15."

55. So I hold that:

i) the "association" between the "one knife" and "each one or several printed products" called for by the claim is that during the process a knife part which is to cut a product or group of products is allocated to the product or group;

ii) the association must continue until after the cut;

iii) there must be a period during which the knife part and product or products have substantially the same velocity;

iv) that period must cover the period of cutting;

v) but it need not be any longer.

56. With that construction in mind, I turn to the process carried out by the NewsTrim. The Judge describes it in [124] quoted above. I will put it another way. There is a fixed knife part "1" shown on the drawing, bottom right. It "awaits" the arrival of the printed product, brought to it in a clamp. By at least position 11 the product is associated with one of the rotating knife-parts. By the time the fixed knife part is reached it and the product are travelling not only at the same speed but in practically the same direction. There is cutting engagement just as the trim starts.
57. That seems to me to be a process essentially as described in the patent. And I think it falls within the meaning of the words of claim 1. The NewsTrim does away with the prior art pair or pairs of knife and counter-knife just as the patent proposes.
58. The Judge said at [126]:

"The essential difference between NewsTrim and the patent in suit is that in NewsTrim the counterknife does not travel with the printed product. The printed product is conveyed by a separate clamp. Of course the counterknife makes contact with the printed product at the time of the cut; otherwise the cut could not take place."

59. I do not accept that is right. Neither the inventive concept explicitly identified by the patentee as a "new route" nor claim 1 require that the counterknife "travel with" the product in the sense that it not only abuts but is responsible for carrying it. What matters for claim 1 is that both first knife part and the product move at substantially the same velocity and be engaged along a cutting edge when cutting engagement begins. It is claim 3 which calls for the product to be moved by the first knife part.
60. Mr Christopher Floyd QC for MM said that the passage at [13] about a movable second knife part supported the Judge's construction. He said:

"Clearly the parts of the cutting mechanism, if they are rollers or rotating or oscillating blades, will move. What is important is that there is no change to the overall position of the mechanism during the operation. The contrast that the claim is making is therefore between the first knife, which travels continuously with the product at the same velocity, engaging with the intended cutting edge, and the second knife which remains behind at the same position or station, even though it may move."

61. The Judge dealt with the point by considering that the phrase could cover an oscillating knife part (see [83]). After all a rotating knife was clearly treated as "fixed" so why not an oscillator. He rejected Prof Fritsch's suggestion that the passage was an error. I do not think it matters. The passage is simply not a useful aid to construction.
62. I should mention some additional points. The Judge said:

"[130] However, up to that point [i.e. that of cutting] they are not moving at substantially the same velocity. This can be seen most easily by comparing the position of the counterknife along the length of the printed product. Professor Lüthi carried out rather more sophisticated calculations which showed the same thing. The printed product is located between two counterknives at position 11. The trailing counterknife "catches up" with the printed product at position 12. From then on until the cut at position 13 they move at substantially the same speed. But between position 12 and position 13 the counterknife has moved further from the uncut edge of the printed product, so that they are not travelling in the same direction. Thus they are not travelling at the same velocity. At position 13 the printed product and the counterknife are moving at the same velocity for an instant of time, necessary to perform a cut. Immediately afterwards their directions of travel diverge. When dealing with the meaning of the word "displacement" in the patent in suit Mr Mellor argued forcefully that the skilled practical person would not be concerned with "infinitesimal" displacement. I think that the same approach is warranted when considering whether the counterknife and printed product in the NewsTrim machine are moving at the same velocity. The practical person would not be concerned only with the precise instant at which the cut takes place. If one lifts one's eyes from the exact moment of cutting, the counterknife and the printed product do not move at the same velocity."

63. I think the Judge was mistaken here. The claim calls for the first knife part and the printed product to move "at substantially the same velocity" and to be "engaged with one another along at least one cutting edge". That is exactly what happens in the NewsTrim just before the moment of cutting.
64. The Judge also said this:

"[131] In addition, I agree with MM that the two rotating knife parts of the NewsTrim machine are fixed (in the sense that they remain in the same overall location during the trimming process) and are associated with each other as a pair (in pairwise manner). Thus, in my judgment, they more closely fit the patent's description of conventional cutting machines than the new route that the patent claims."

65. The Judge erred here also. The NewsTrim has one fixed knife "l" and six rotating knives "k". The rotating knives operate in turn with the fixed knife. They are not in the acknowledged prior art sense "pairwise" at all. The "pairwise and in fixed manner" acknowledgement in [8] of the specification is of a knife and counterknife which form a pair which always operate together and only with each other.
66. No other point of construction relevant to the issue of infringement arises. I therefore conclude that claim 1 is infringed. Given my conclusion that claim 1 is valid, it is not necessary to spell out which other claims are infringed.

Validity – Stobb (US Patent No. 4,496,140)

The disclosure of Stobb

67. Stobb is the principal citation against validity. The Judge held (on the basis of his construction of the concluding words of claim 1 to which I will come) that it did not anticipate claim 1. But he held that apart from that last feature, all the other features of the claim were present. And he went on to hold that the claim (and all other claims) were invalid for obviousness over Stobb. Ferag appeal the finding of obviousness, MM that of no anticipation.
68. I turn to the detail of Stobb, noting at the outset that it does not ever seem to have been implemented – it is what is sometimes called "a mere paper" proposal. It was published in 1985, about 3 years before the priority date of the patent in suit.
69. I can gratefully borrow the Judge's description of Stobb, omitting, at this stage, anything controversial:

"[46] … In Stobb, a trimming station is integrated into a stacker (a "compensated stacker"). Stobb takes three fixed blades arranged in a "U-shape", and the signatures are moved past those blades by means of a rotary motion. Two of the blades are adjustable.



…

[47] The operation of the Stobb process can be seen from Figure 1 of Stobb (partially reproduced).

[48] A conveying means (12) and (13) drops a signature (10) into a clamp, formed by members 18 and 19, pointing vertically upwards, and then rotated anti-clockwise past two fixed side cutters (16) and a fixed front cutter (17) and cut. The figure shows four clamps in a cruciform arrangement. The blades on the side cutters (16) are shown as angled or sloping blades. Once cut, the clamped signature continues to rotate anti-clockwise until the clamp points vertically downwards. The clamp releases the signature, and it falls into the compensated stacker (not shown). The stacker sits on a rotating table, which is rotated periodically by 180^o so that the signatures stack evenly. The patent explains that although the embodiments show the signatures being rotated past the cutters, the cutters could be rotated past the signatures.

[49] Figure 4 explains how the clamp and the side cutters can be adjusted.

"A cross bar 47 is connected at the extending ends of the bars 46, and that determines the spacing between the outer edges 48 of the bars 46 and that spacing is the width of the signature after it is trimmed. When it is desired to change the width of the trimmed signature, then the cross member 47 is changed and a member of a different length can be inserted between the bars 46 and connected thereto, and that will change the spacing between the edges 48 and thus change the width of the signatures to be trimmed, and of course it will be recognised that the side cutters 16 will also be adjusted to have their cutting edges 49 in the plane of the edges 48 for trimming the signature overhang designated 51 in FIG 5."

[50] The final paragraph of the specification in Stobb says:

"It will be seen that the arrangement is such that the holder or clamp member 19 is fixed with the shaft 15 to rotate therewith in the direction of the arrow B, and that rotation is such that member 19 becomes a bed knife relative to the three cutting blades 16 and 17. Thus, the holder 18 is used for pressing the signature against the member 19 which then serves as the main member against which the cutting is accomplished".



Figure 5 shows the side cutters (16) in the plane of the leading and trailing edge of the clamp (46). The clamp is shown as symmetrical. The view is a head on view; and the front of the clamp (corresponding to member 19 in figure 1) is shown as being made of three pieces (46, 47, 46). It is this member that "becomes a bed knife".

70. I draw particular attention to the fact that the left and right edges of the clamp do not show a cutting edge. The signature is just clamped between two identical members. This cannot be an illustration of a shear cutter, for which it would be necessary to make one or other of the clamps slightly smaller than the other. And it would be the other which was provided with a knife edge.
71. Although the Judge does not actually reproduce it, fig. 4 of Stobb is important too. It, along with fig. 5 show "arrangements of the clamp." The clamps have two sides, "members 18, 19". The reason for the importance is this: that it was common ground that fig. 4 cannot be showing a knife edge. Knife edges are not made in parts and are not adjustable. So much was agreed between the experts. And as I have pointed out fig. 5 is consistent with this.

Anticipation by Stobb?

72. So if one actually makes Stobb as described, the edges of the clamps are not knife edges. That in turn implies that the cutters 16, 17 must be single blades. Moreover, apart from the last paragraph (as to which more) Stobb does not say that the edges of the clamps are knife edges. And when he wants to refer to cutters he does so expressly e.g.:

"two side cutters 16 and one end cutter 17 are operative on the signature 10 when it is in the trimmer 14" (col. 1_59-60)

This is not the language one would use if 16 and 17 were only parts of a 2-knife cutting system, the other parts being the edges of the clamp.

73. There is another reason why what is specifically described cannot be a two-knife system. As shown the edge of 17 (the front edge cutter) and the edge of 16 (the side edge cutter) are touching at right angles. No skilled man would attempt to make a 2-knife system in such a way – see para 18 above.
74. Accordingly, in my judgment (leaving aside the difference about "intended cutting edge") what is actually described in Stobb has a number of differences from claim 1:

i) There are not two knife parts;

ii) There is no cutting engagement between two knife parts;

iii) There is no first knife part moving at substantially the same velocity as the product along a cutting edge.

75. The Judge thought this was not so, that apart from the fact that Stobb did not cut right across (which is what he held was required by "cut along an intended cutting edge") Stobb described a 2-knife system. He did so for two reasons. First he accepted the evidence of Prof. Lüthi that what was shown in Stobb was purely schematic and second by reliance on the final paragraph. I think he was in error.
76. Above I have tried to set out the specific two requirements for anticipation, namely disclosure and enablement. What Stobb actually discloses in detail is simply not a 2-knife system. One has to do something different from what he describes specifically to arrive at such a system – you must depart from what he actually describes. That is not sufficient to satisfy the disclosure requirement.
77. Nor do I think it satisfactory to accept evidence that the drawing was purely schematic. What the Judge said was:

"[100] In my judgment Professor Lüthi had the answer to this point. He said that Stobb was not a blueprint for the manufacture of a working machine. It was a schematic presentation. A skilled addressee would construct member (19) as a counterknife and would realise that there needed to be marginal clearance between the leading edge and the trailing edge of the clamp. The kind of clearance required would be no more than a few microns. The details were not difficult to achieve. The edges of member (19) could be mitred corners, which any skilled person would adopt in constructing a working model of Stobb. To build in a measure of clearance was straightforward engineering. They do not fall outside the scope of trial and error experiments to make the machine work. I accept Professor Lüthi's view ("if it moves like a duck, and quacks like a duck …"). For practical purposes member 19 is a counterknife in a two-knife shear cut. In my judgment Stobb discloses a counterknife (or, in the language of the claim a "first knife part"). "

78. True it is that Stobb is not a blueprint. Patent drawings seldom, if ever, are. But that does not mean that one can regard a prior disclosure in a patent as something different from what it is, however obvious it may be to make something different. The Judge was wrong to accept this evidence as amounting to proof of disclosure of a 2-knife system by Stobb.
79. And he wrongly thought that trial and error to make the machine work was relevant. Whether the trial and error is obvious or not, there is a departure from the actual disclosure. As I have said, trial and error really only forms part of the second requirement of anticipation, namely enablement.
80. What then of the final paragraph and particularly its statement that:

"rotation is such that member 19 becomes a bed knife relative to the three cutting blades 16 and 17."?

81. This to my mind is far from a "clear and unmistakeable direction" to make knife edges on all three sides of 19. I say that because:

i) The expression "bed knife" is not a recognised technical term. It is true that Prof. Fritsch rather unsatisfactorily tried to establish otherwise, but his failure in that regard merely leaves the court in the position that the skilled reader would have a conundrum as to what Stobb meant;

ii) It is wholly improbable that Stobb meant 19 actually to be constructed as a knife part. It is not merely that he has not said so, it is that what he has actually described in detail about the adjustability of the clamp means it cannot be a knife part;

iii) The language used – "becomes a bed knife" – is not that which one would use if 19 simply had a knife edge: actual knife edges do not "become" so. They are always so.

82. Mr Floyd strongly relied on the cgk about single-knife cutting to interpret Stobb. He raised a syllogism thus:

i) only two kinds of cutting are known, single and 2-knife;

ii) single knife cutting always requires a cutting strip - you need a chopping board to chop an onion;

iii) no cutting strip is shown in Stobb and indeed you could not have one because you need the rotation of the clamp past the knives;

iv) So Stobb must be 2-knife cutting.

He added that this is reinforced by the fact that knives 16 and 23 are shown sloping so that they make a progressive cut.

83. Mr Mellor said not so, for the reasons I have already mentioned (Stobb does not say so and describes 18 and 19 as being multipart and adjustable, which are incompatible with having a knife edge). He further submits that what Stobb had in mind was a variant of punch cutting. Stobb's clamps pass at speed close to (but not in cutting engagement with) single knives. The fact that the product is clamped, coupled with the speed of the operation, the inertia of the product and the closeness to the knife means that product is cut rather than moves out of the way. In argument there was an analogy with a rotary blade mower, whose blades go so fast that the grass has no time to get out of the way.
84. Professor Fritsch spoke of this in technical language, talking of "visco-elastic" principles. The Judge said that:

"[98] However, I do not consider that a knowledge of the principles of visco-elasticity would have been part of the mental toolbox of the skilled addressee of the patent at the time."

But you do not have even to be skilled in the art of trimming to know that if you slash at an unsupported membrane (e.g. a falling piece of paper) with a knife you may cut it, whereas if you only push gently against it, it will just move away. It is fair to say, however, that no-one had actually used an unsupported single-knife to trim in this way.

85. Mr Mellor said that was no reason for the skilled man not to see what Stobb was proposing. He might not think much of it, but that is another matter.
86. Mr Mellor in further support of his argument, relied on a model of Stobb made by Ferag. It was substantially in accordance with Stobb. It had two single-knife side cutters and a horizontal single-knife front cutter. It did not have knife edges on the clamp and provided more clearance than would be consistent with two-knife cutting. It worked to produce a U-shaped trim.
87. Mr Mellor also refuted the suggestion that because Stobb shows sloping blades it cannot be a punch cut by pointing to the fact that progressive punch cuts were known as part of the cgk. I agree that the sloping nature of Stobb's two side-cutting knives is no indication in itself of a particular type of cutting.
88. This model demonstrates to my mind clearly that one can make a trimmer in accordance with Stobb which is not a 2-knife machine – it indeed demonstrates the no "inevitable result" point.
89. I accept Mr Mellor's answers to Mr Floyd's syllogism. The latter depends on the assumption that the skilled reader would take it as a given that there are only two types of cutting. I think he would read Stobb's specific description, take it that he is proposing single knife cutting, and probably not think much of the idea. After all this is a patent which is supposed to be proposing a new idea.
90. Mr Floyd also pressed us heavily with what Stobb says in describing his specific apparatus. He refers to the sentence saying:

"When it is desired to change the width of the trimmed signature, then [specified adjustments can be made], and of course it will be recognised that the side cutters 16 will also be adjusted to have their cutting edges 49 be in the plane of the edges 48 for trimming the signature overhang…(Col 3_52-61)"

He seized on the reference to in the plane of saying that it is only in 2-knife cutting that you have cutters in the same plane. But that is not right. Even for 2-knife cutting you must have a small difference in plane, as was well-recognised. And even more significantly the passage is in the specific context of Stobb's adjustability. It is not possible to spell out of this phrase in this context that side cutters 16 each form part of a pair of shear cutters, the other member of the pair being the sides of the clamp.

91. There are two other matters which reinforce my conclusion about Stobb being a single-knife system. First there are Stobb's patent claims. I need only go to the first, for the point is the same on the others. What this says is:

"Apparatus for handling a signature, comprising a conveyor for moving a signature and having a terminal end for releasing the signature, a rotatably mounted signature holder (movably) disposed adjacent said terminal end for receiving the signature from said conveyor and with an edge of the signature extending beyond said holder, a stationary cutter disposed adjacent said holder and in the path of movement of the extending edge of the signature for trimming the extending edge of the signature when the signature is rotationally moved past said cutter."

This is simply not the language one would use if the holder itself was to be provided with a knife-edge. It is the "stationary cutter" which is to be in the path of the "extending edge" of the signature. There is nothing requiring a cutting engagement between the cutter and the holder. And nothing about the "holder" having knife edges or being a "cutter."

92. The second, upon which I do not place great reliance, is the reference to "becomes a bed knife". Of the various interpretations of this the one which seems to me most plausible (or at least least implausible) is that Stobb is trying to say that the rotation makes the single knife act as though it were pressing into a "bed" i.e. the rotation causes the cutter to act as if there were a cutting strip.
93. I therefore conclude that the Judge was wrong about Stobb disclosing a 2-knife system. What then of his holding that Stobb did not disclose the last part of claim 1 about the "intended edge?" It was Mr Floyd who attacked this. What Stobb discloses on any view is a machine which cuts a 3-sided trim rather than three strip trims. The Judge considered that claim 1, by the language "at least cut along an intended edge" was limited to strip trims. The "intended edge" meant the whole edge of the signature. He said:

"[85] Mr Floyd submits that the "intended cutting edge" is no more than the edge along which it is intended to cut. If it is intended to cut along the whole of the (uncut) edge of the product, then that is the intended cutting edge. But if it is intended to cut only half way along the (uncut) edge, then that is the intended cutting edge. Mr Floyd also drew attention to claim 18 (an apparatus claim) in which it is claimed that the apparatus is so constructed that "the edges of the printed products are cut from one end to the other". This, he says, shows that the patentee knew how to describe a cut along the whole length of the (uncut) printed product. He had not done so in claim 1.

[86] Mr Mellor, on the other hand, points to the embodiments of the method. The figures consistently show that the printed product is cut along the whole of its (uncut) edge. Moreover, it is clear that the patent is concerned with two-knife cutting: its whole point is the separation of knife and counterknife. In any two-knife cut the line of engagement of the two knives must extend beyond either side of the product to be cut. This is a general principle of two-knife cutting. Consequently the skilled addressee, reading the patent in context would understand the expression "intended cutting edge" to mean the whole edge of the (uncut) product.

[87] I do not consider that claim 4 sheds any light on this debate. It is true that it refers to cutting from "one end" to "the other". But from one end of what? From one end "of the intended cutting edge" to the other. So the precise meaning of the intended cutting edge is still at large. All that you can say about it is that it has two ends. So I put claim 4 on one side.

[88] Although I see the linguistic force of Mr Floyd's submission, I consider that it places insufficient weight on the context of the patent; and the common general knowledge that the skilled addressee would bring to bear on his reading of it. The argument is the same sort of argument that the House of Lords rejected in Catnic ("vertical" contrasted with "substantially horizontal"). The skilled addressee would know that the trimming of products is typically performed by arranging trimming knives so that cutting occurs progressively from one edge of the paper to the other. This was common ground between the experts. So if the product was not to be cut from end to end the skilled addressee would, I think, have expected to be told that in terms. He would also notice from the description of the preferred embodiment that as the process unfolds "the observer-facing lateral edge of the printed product is trimmed"; and that the final cut is that of "the top edge of the printed product". He would also know that if the printed product was not cut along the whole of its (uncut) edge, there would be waste still partially attached to the partially trimmed product, with no clear means of disposing of it. Read in context, I conclude that the intended cutting edge means the whole length of the uncut product.

94. Mr Floyd submitted that this was a non-purposive construction. What the skilled man was aiming at was a product trimmed on all three edges. Provided he got that, he got what he wanted and could not care less whether the cut went further than was necessary. Suppose, Mr Floyd asked, that Stobb was later than the patent and was accused of infringement. Could he escape by saying "I do not cut the part that does not matter?" Surely purposive construction would catch him.
95. At first I had much sympathy with Mr Floyd's argument. But on reflection I have concluded that the Judge reached the right answer. The expression "intended cutting edge" must be construed in the context of the patent in suit. There is no doubt that in that context only 2-knife cutting is contemplated. And it is common general knowledge that you always cut from one side to another when doing a shear cut – unlike a punch cut. It is simply not within the nature of the technology of shear cutting to cut U-shaped pieces. So the skilled reader would not read the closing words as contemplating that.
96. And Mr Floyd's point about Stobb viewed as an anticipation is not realistic – a man making Stobb as shown would not be using a 2-knife system anyway, so he would not infringe in any event.
97. Accordingly I think the Judge was right in his conclusion as to non-anticipation by Stobb, but wrong in that he failed to appreciate there were more differences between claim 1 and Stobb than he found.

Obviousness over Stobb

98. The upshot of my reading of Stobb and the patent claim is that there are the differences I have identified above plus the fact that he cuts U-shaped pieces of trim. Is altering Stobb so as to remove those differences obvious? As I have said the Judge necessarily considered this only conjecturally.
99. I think the answer is no. My reasons are as follows:

i) Stobb is not dealing with the same problem as Ferag, keeping the trimmer up with the speed of delivery from the printer and binder – speed of 80,000 copies an hour. Stobb as shown has a feed which would operate at much slower speeds (as was common ground). So it is not likely to be of interest to a man concerned with the problem of high speed trimming – how to keep up.

ii) Stobb is a single-knife cutter with three blades;

iii) As a single-knife cutter depending on a punch the skilled man would have caution as to Stobb's approach. His use of a phrase which was unknown to him ("becomes a bed knife") would, I think, regard Stobb as perhaps a man who has not thought his idea through very clearly.

iv) Stobb produces U-shaped trimmings which at best would be a nuisance to get rid of.

100. In short, I think a skilled man would pick Stobb up, read his idea, and just put it down again as a not very practical proposal. Putting it another way, I think it clearly involves a lot of hindsight to start from Stobb and make such variations as would bring it within the claim.
101. In fact MM made two models intended to demonstrate how, in practice, a skilled man would make variations from Stobb. We were shown them. Each departed from Stobb in the following ways:

i) The trailing part of the clamp (Stobb's 19) was provided with knife edges;

ii) The leading part of the clamp was made of slightly lesser dimension than the trailing edge so as not to interfere with a shear cut.

iii) The front edge of the clamp was non-adjustable;

iv) The slope of at least one side cutter was reversed in comparison with the printed product during cutting;

v) The front and side knife edges were displaced. Stobb shows the side cutters 16 at their front end level with the front cutter 17 and sloping away downwards;

vi) The front knife edge extended beyond the side cutters at each end so that there was no simple U-cut – the "vertical" arms of the U each were partially cut by the front knife extension.

The two models varied in that in one the front knife part was displaced up above one side cutter and in the other down below both side cutters. Neither were as shown in Stobb where the knives are level.

102. To my mind these models clearly show just how much of a departure from Stobb one would have to go to get within the claim. These are not mere workshop variants. Who actually devised them was never explained. It was not Prof. Luthi. One is entitled to infer they were devised by someone with the aim of demonstrating obviousness and knowing the target. They are hindsight models.
103. What then of the Judge's conjectural view? I think it was in error in two ways. First he did not really get the inventive concept of the patent accurately. He said:

"[113] In my judgment the heart of the invention (and hence the inventive concept) is the separation of knife and counterknife: the one travelling with the printed product (but fixed relative to the printed product) and the other remaining fixed (in the sense I have described). This means that there can be many counterknives per knife (i.e. the knife and counterknife are not associated with each other "pairwise"). Of equal importance is the feature of the counterknife that it is (or is part of) the cell wall, and therefore also acts as a part of the clamp. The invention is limited to cutting products from end to end."

104. On my construction of claim 1 this has a number of errors, particularly the requirement of prolonged physical contact between product and first knife and second importation of a requirement that the cell wall acts as part of a clamp. However this does not really matter – the Judge's view is a narrower concept than mine.
105. Where I think the Judge went wrong is in failure to guard against hindsight and the danger of a step-wise approach to obviousness. Once one understands that Stobb would be read as a slow single knife puncher offering no hint of a solution to the problem tackled by Ferag, I just cannot see how a skilled man would be induced to make all the necessary changes. There is no motive for change unless and until one sees that it can provide the solution to the high speed problem.
106. One can see just how many changes there would have to be from what I have set out above concerning MM's models. I do not set out all the Judge's reasons, but I note particularly that he relied on the "bed knife" point (where I think he was wrong as a matter of construction). As to the problem of the U-shape waste, the Judge went on to hold that the obvious way round that was to use a "double-Stobb" – two varied Stobb machines in series, one cutting the sides and the other the front. Added to all the other variants from Stobb, I think the collection of them is far too much to expect of the unimaginative notional person skilled in art.
107. Accordingly I think the Judge erred in principle and would hold claim 1 valid over Stobb.

Rösner (GDR Patent 112,380)

108. This is cited only for anticipation. The Judge rejected that. MM say he was wrong, but only if the claim covers the NewsTrim. The Judge's description of Rösner was accepted as accurate:

"[41] Rösner dates from the mid-1970s. It describes a process for trimming paper stacks, principally book blocks, by one or more straight knives while the product is moving. This is achieved by trimming the book blocks by means of a cutting knife as the book blocks are moving, and by then pivoting the book blocks about 90°, again as they are moving, before they are trimmed further by another similar cutting knife. The blocks are held on a trolley or carriage. Rösner describes the cutting knife as follows:

"One advantage is that the knife movement can be chosen at will, anything from a straight-line knife movement in a guide cam to a circular movement if the knife represents the coupling link between the cranks of a parallel crank drive fixed to the frame."

[42] The embodiments show both single knife cutting and also two knife cutting. It is the latter that is relevant for present purposes. As Rösner states:

"The so-called bottom cutting tool which may be a bottom knife for shear-cutting or a cutter board for knife-edge cutting is more effective if fixed to the guiding arrangement although it can be in a fixed position in the machine's cutting station."

[43] In the two-knife cutting, the cutting knife engages the cutting edge of the book block, and a shear cut is achieved against the edge of a counterknife, also associated with the cutting station. As Rösner states:

"It goes without saying that the movement of the knife is synchronised with the movement of the carriage in such a way that the cutting procedure always takes place when the carriage has run into the guiding arrangements and the guiding arrangement briefly moves at the speed of the carriage for the duration of the cutting procedure."

[44] The upper, or cutting, knife moves, guided by a cam, so that it both follows the book block horizontally as it moves through the cutting station, and descends on the book block so as to cut it in cooperation with the lower knife. However, as Professor Lüthi demonstrated, when the cutting process takes place the book block (with the counter-knife attached) moves past the cutting knife, although not all the way past.

109. Ferag argued that there were two reasons why Rösner did not anticipate, first that Rösner's top knife was not "fixed" which the judge rejected and second that his lower knife and book block did not "move past" the upper knife, which the judge accepted.
110. It is not necessary to go into both reasons. The second is sufficient. In claim 1 of Ferag's patent there is a requirement that "the first knife part and associated printed product are moved past a second, fixed knife part in order to be brought into cutting engagement therewith so that the printed product is cut at least along an intended cutting edge." That simply does not happen in Rösner. The Judge was entirely accurate when he said:

"[105] As I have said, Professor Lüthi demonstrated that the book block and the lower knife moved part of the way past the upper knife. They did not move the whole of the way past. Rather the blade of the upper knife retracted, and then the book block moved on. In addition, the movement of the book block and the lower knife was the conveying process, rather than part of the cutting process itself. In other words the movement did not take place "in order" to bring the two knives into cutting engagement. In my judgment Rösner did not disclose this integer."

111. Mr Iain Purvis QC for MM said that there was a "squeeze", that the conceptual similarities between the NewsTrim and Rösner were great. But his own submission shows the difference:

"Just as with Müller Martini the counterknife here does not move with the paper throughout the conveying process. It is associated with a particular cutting station and it only moves briefly with the paper as it is brought into engagement with the cutting knife. So paper comes in, and just at the point where you are going to cut you introduce the counterknife, you slice through it, and it moves on, just as with Mueller Martini. That is why it is an important squeeze document. The only real difference from the Müller Martini system which can be identified is that the cutting knife in Rösner, although it is fixed to the cutting station, has an element of movement relative to the frame of the station. We have seen how it swings into the path of the paper as it moves through. In the specific embodiment of Rösner it is rotating about two fixed points."

So the knife simply does not "move past etc."

Conclusion

112. Accordingly I have reached the conclusion that the NewsTrim is a machine which carries out the process of unamended claim 1 and that that claim is valid. All other attacks on validity (e.g. insufficiency) were abandoned.
113. Mr Mellor made it plain that if that conclusion was reached he would abandon his application for amendment of the patent. So I need not consider it or whether it took the claim further from the prior art. It is true the application was made unconditionally, but there was no concession that the unamended claim was invalid, so it is difficult to see what "unconditional" meant. Doubtless there may be a question of costs about the amendment proceedings. There is no more to be said about them now.
114. In the result I hold that the patent is valid and infringed. I would allow the appeal.

Lord Justice Tuckey: I agree.

Lord Justice Mummery: I also agree.