Digital Technology Law Journal
|
|
Home
| Databases
| WorldLII
| Search
| Feedback
Digital Technology Law Journal |
|
The Revolution That Never Was: A Cynic's Eye View of the Software, Business and E-Commerce Method Patenting Controversy in the Wake of State Street
Gary Lea*
In this piece, the author disputes the view that patentability of software related inventions or business methods is a new phenomenon and he also disputes the view that such patents, if granted, are harmful to the software industry and commerce. Starting with the US, brief accounts of the various supposed 'bars' to patentability of software and business method inventions are given in order to show them in their true and proper context whilst overviews of developments in the UK, Japan and the EPO are also given.
Introduction
1. In the wake of the decision of the Court of Appeals for the Federal Circuit ('CAFC') in State Street Bank v Signature Financial[1] (“State Street”) the US patent system has been wide open to applicants seeking to patent software related inventions ('SRIs'), business method related inventions ('BRIs') or their combination subset, e-commerce related inventions ('ERIs').[2] Some commentators[3] have reacted as though State Street represents an abandonment of long-cherished and well-grounded legal principles that SRIs, BRIs and ERIs are unpatentable with a result not far away from the legal equivalent of Armageddon. It is the intention of this author to show that State Street represented no real break with the past and that, instead, it merely puts patent law and practice back to the position it started from and, indeed, which it should have stayed at.
2. In light of the above, it is strongly arguable that, whilst there are undoubtedly some genuine concerns around SRI, BRI and ERI patents as granted, these should not relate to any deep philosophical concerns over patentable subject matter but the more practical, if mundane, issues of novelty and inventive step; in other words, it is this author's contention that many criticisms of the so-called 'new' line amount to no more than a rehash of that old avian cliché, "the sky is falling, the sky is falling".
The proper limits of patentability: initial considerations
3. Few would argue that, in any rational patent system, some things should not be patentable: at the very least, one needs some basic indication of what actually constitutes an 'invention' for the purposes of applying for a patent.
4. In the US, the courts, in particular the Supreme Court, have been responsible for drawing the line between patentable and non-patentable subject matter: the sequence of Patents Acts, culminating with the current Patents Act 1952 as amended ('USPA'), merely draws a broad framework for the classification of subject matter. At present, therefore, section 101 USPA simply states that patentable subject matter comprises any " ... [new and useful] process, machine, manufacture, or composition of matter, or any ... improvement thereof".
5. How have the US courts drawn their lines traditionally? In a long line of decisions stretching from O'Reilly v Morse[4] to Funk Brother Seed v Kalo Innoculant[5] the Supreme Court repeatedly declared that laws of nature, natural phenomena, mathematical formulae and mere abstract ideas were not patentable per se. This was a very limited list and it was not entirely condemnatory: the Court also went on to emphasise that if an inventor devised something which utilised or applied any item(s) on the barred list then, subject to meeting the other patentability criteria, that thing was patentable[6] The baseline standard for 'invention', therefore, appeared to be the identification of something artificial[7] and actively functional.[8]
6. If the Funk standard had been carried forward without any gloss or extrinsic pressure, where would it have left SRI, BRI and, later, ERI patent applications? The mere fact that they almost certainly would have included ideas or mathematical formulae would not have automatically barred them from patentability. So why was there a problem and a later reaction? In the case of SRIs, four factors are considered: the mutual misunderstanding of judges and the USPTO, a widespread failure to understand software technology, an unnecessary judicial obsession with the correct classification under the USPA section 101 framework and overly-ambitious claiming by applicants. In the case of BRIs and ERIs, the last two factors also apply but, in addition, we see specific factors such as the relationship of business to technology also coming into play.
7. What of the world outside the US? One can point to specific statutory bars on patentability of subject matter comprising part or whole of SRI, BRI and ERI patent applications or patents in the laws of a number of countries; surely this proves there is something 'wrong' with patenting such 'inventions'. Below, the author hopes to indicate that the origin and nature of these bars has been misunderstood and that, in fact, that there is often little substantial difference between the position of the US and other industrially significant jurisdictions.[9]
8. In the course of this piece, consideration will be given to the patentability of SRIs, BRIs and ERIs; this will be primarily in relation to the US but with some consideration of matters arising in the UK, in Japan and under the EPC régime: mention will also be made of various EC and TRIPS issues arising. However, before examining the various twists of patent law and policy dealing with the inventions in question, it is necessary to describe some of the technical issues at stake with a view to setting benchmarks against which the progress made can be assessed.
An appreciation of SRIs
Introduction
9. The term 'software related invention' is, by comparison to the others under investigation, both broad and vague. Furthermore, a lack of proper understanding of software technologies on the part of lawyers and judges in particular has had serious repercussions for the proper development of patent protection[10]
Start at the beginning: what is software?
10. In the computer industry, a distinction is frequently drawn between hardware and software: hardware refers the physical parts of a computer whilst software is 'everything else'; within that miscellaneous category, one finds computer programs, instruction materials, design materials and the like. However, software engineers will tell us that, strictly speaking, the miscellany does not cover physical items such as disks for storing programs or paper for instruction manuals, but the intangible elements of instruction, data and information: for present purposes, therefore, software will take a more restrictive meaning synonymous with computer program.
11. Let us enter an imaginary patent system in which we leave aside what we know of the development of patent law for a moment, and we know only the Funk standard described above: any mention of instruction/data/information by a software engineer might raise a suspicion that, perhaps, this sort of subject matter does cross the line into unpatentable idea or mathematical formula. However, the point we must bear in mind is this: a computer program is a polymorphic entity i.e. it is capable of taking on different forms.[11] In source code form, it is informational and static in nature: in binary object code form, it is dynamic and functional - in plain English, it makes a computer do something.
12. Therefore, if we take active functionality as the touchstone for patentable subject matter in our imaginary patent system, it stands to reason that some elements of some manifestations of computer programs should, hypothetically, be patentable; however, on the other hand, it is equally clear something that is, or only amounts, to the idea of doing something - a mere plan, a mere scheme without more - should not be patentable. The difficult part thereafter lies in the drawing of the dividing line. In order to make the required distinction, it is worth considering what the object code version of a computer program actually does when it is loaded into a computer and run: the short answer is that, in directing it to perform a task or task it causes the logic circuits of that machine to be reconfigured in a particular way. The reconfiguration is such that, in effect, the programs are standing in for hard-wired links between each set of logic circuits in the machine; this means that programs 'stand in' for hardware and, indeed, the two are, in fact, interchangeable.[12]
13. This software/hardware equivalence gives rise to another intriguing possibility: if hardware can be simulated in software and vice versa, might is not be possible to claim the same program functionality either as a process or, if embodied, as apparatus? In principle, yes and, furthermore, we might also entertain claims to the 'computer-as-programmed' as a 'new' machine. What, then, is the cut-off point? Under our hypothetical patentability standard, programming techniques, algorithms, design elements or the like but then only to the extent that these things do not make a computer do something. The key distinction, therefore, is between the purely informational and the functional.
A ‘natural’ approach to SRIs?
14. If we now leave our imaginary patent system, it is important to note that that the patent system of a major industrialised nation, left largely untouched by external pressure, did indeed manage to reach exactly the same conclusions set out above without any sign of the ill consequences currently being trumpeted: the UK patent system prior to entry into force of the Patents Act 1977. Under the 'old' Patents Act 1949, the UK system functioned somewhat like the US: although, unlike the US, there was no rigid categorisation as between process, product and so on, it was left to the courts to determine what constituted patentable subject matter within the general rubric of 'manner of new manufacture'.[13] How did this work in the case of computer programs? The distinction between the informational and the functional was recognised in Slee & Harris' Appn.[14] where the applicants were seeking a patent for a linear function solution program, claimed in just those terms. The application was refused on the basis that it did not disclose a physical effect which amounted to manner of new manufacture[15] and went to an Office Hearing at which it was held that, although the claims to the program per se were unacceptable, it would be possible to submit new claims drafted in terms of "methods of controlling data processing machinery".
15. Following Slee, in February 1969, the UK Patent Office issued a practice note in which it stated that whilst it held that computer programs as such were not patentable, it would accept applications for methods of programming a computer to operate in a certain way, for computers so programmed or media bearing programs which controlled the operation of a computer in a stated way. Furthermore, no subject matter objection would be raised in relation to applications relating to computer-controlled manufacturing or testing of computer equipment. After this point, the Patent Appeal Tribunal and the courts dealt with a series of decisions in which what appeared to be at stake was usually more a matter of drafting than substantive issues of lack of 'physical effect'.[16]
o In Badger Company Inc.'s Appn.[17] a patent application for a method of computer-assisted plant design, originally claimed partly by relation to the program and partly by the method of use, was rejected by the Office for lack of patentable subject matter, use of known apparatus and use of known technique. On appeal to the PAT, the applications was remitted to the Office for consideration with revised claims directed to a "process for conditioning the operation of a computer and plotter of known types to produce the layout ..."; this technologically sophisticated form of process claim became standard.
o In Gever's Appn.[18] a patent application for a trade mark indexing program, originally claimed partly in relation to the program and partly in relation to the computer upon which used, was refused on grounds of lack of patentable subject matter, use of known apparatus and use of a known technique. On appeal to the PAT, it was held that the claims were worded and directed properly, and that, more importantly, confirmed that a program on a carrier[19] was patentable in the same way that a machine part which could condition the operation of a known machine might be.
o In Borough's Corp. (Perkins') Appn.[20] a slight but significant shift occurred; this application involved a method of communicating data between a central computer and intelligent[21] terminals and, on appeal to the PAT, it was held that the Patent Office's distinctions between the (unpatentable) programming of a computer and (patentable) controlling its operations were artificial. The claims were, therefore, acceptable as they were. Furthermore, it was held that any residual reliance on older, restrictive test for assessing patentable subject matter was to cease: once a computer program had physical form, either on a carrier or in a computer-as-programmed, such that "... it could be realised in practice, it should no longer be regarded as a naked conception for it has found a practical embodiment."[22]
16. The real, and last, breakthrough under the 'old' 1949 Act came in IBM Corp.'s Appn.[23] in this action, IBM applied for revocation of a patent relating to a data handling system for market price matching in fungibles; this was done on a number of grounds relating to subject matter, known method and known machinery. The court rejected all the grounds for revocation, holding that the mode of claiming used - pure computer program claims - was wholly irrelevant; the substance of the invention was patentable and, although a combination of known technique, machinery and other elements, it was, when taken together and as claimed, novel.
17. An additional point of interest in IBM is the evident admixture of a method of doing business; this did not cause a problem because, when implemented by the computer program, it was not just mere paper scheme or rule or method, but an active, functional component of a virtual market.
An appreciation of BRIs and ERIs
Introduction
18. Although conceptually simpler than SRIs, the range of inventions covered by the acronyms BRI and ERI does warrant some further investigation. In the case of BRIs in particular, much will turn on the issue of which business related activity is the subject of the application or patent in question. For present purposes, BRIs are taken to cover 'methods for doing business'[24] in the broad areas of advertising, marketing and accounting whilst ERIs are taken to cover the same things over computer networks.
Information v. function revisited
19. Let us return to our hypothetical patents system for a moment and consider the following: applicant Smith wishes to obtain a patent for some "method of market management". If, as before, we take the requirements of artificiality and active functionality as our starting point then it might be possible for Smith to obtain her patent provided that an examiner could be persuaded that the actual method in question was more than a mere idea or formula: the dividing line would lie, for example, between a method for conducting share sales so as to lower transaction costs (patentable) and a method for the quick calculation of transaction costs within the share system (unpatentable).[25]
20. Therefore, in our hypothetical patent system, as with SRIs, we find that there is no brightline rule excluding BRIs or, by extension from both, ERIs: it will all depend on determining the precise nature of the invention in question. Indeed, in the case of ERIs there is one interesting additional point to consider: under our hypothetical system of patentability determination, an invention that is an unpatentable BRI may well acquire potentially patentable status as an ERI if automated: the transaction cost formula could be used as part of an automated system for tallying and deducting transaction costs (patentable). It is vital to remember, however, that this only relates to subject matter status: it may equally be the case that it would be "obvious" to automate transaction cost processing as described above, in which case a patent should not be granted.
21. Returning to the real world, one of the great difficulties for BRI and ERI patents to overcome is the sentiment, often translated into patent policy, that the patent system is concerned with matters scientific and not matters commercial; in truth, the first part of this is correct only in so far as it relates to applied science and, more crucially, the exclusion cannot be axiomatically derived from the inclusion. Fuelling[26] rightly points out that the service industries involve as much 'technological art' as the traditional manufacturing industries[27] and, in light of that, there is no more reason to reject BRIs and ERIs as unpatentable per se than SRIs.
SRIs: the U.S. experience
Initial SRI patenting problems in the U.S: 1966-1970
22. If the UK patent system could adapt to software technology, what went wrong in the US? The trouble began not in the courts but at the USPTO and for purely administrative reasons. Until the first half of 1966, the Office had accepted a few SRI patent applications and, in August, released its first set of draft "Guidelines to the Examination of Programs" which allowed programs to be claimed if expressed as devices for controlling a computer. Worryingly, perhaps, the draft went on to say that process-types claims would not be accepted since these were analogous to mere mental steps or ideas;[28] this, as we have seen, above was illogical nonsense and a sad triumph of claim form over inventive substance - if, as they were, hardware and software were interchangeable, it did not matter how the claims were expressed since they related to the same thing.
23. In the event, the USPTO did not even get a chance to make the device/process mistake because, by December 1966, the draft had been withdrawn; the main reason for this was the publication of the Commission Report on the Patents System[29] which called for a bar on patentability of computer programs. The reasons offered up in the Report had little to do with subject matter arguments but, instead, pointed out that the USPTO was already up to 4 years behind with patents applications, that there were few experienced staff and that there was very little prior art to search against that was held within or available to the Office.
24. Shortly thereafter, the USPTO underwent a Pauline conversion to a restrictive approach, seeking a statutory bar on patentability in a 1967 amending Bill and, when that was dropped, issued a set of extremely restrictive Guidelines in 1968; these excluded any computer process that could be carried out by mental steps; however, the 'mental steps' test as described by the Office, was so wide that it covered any operation of a computer program that could be tracked by the human mind.[30]
25. In the background, though, was the start of a simmering institutional argument: the Court of Customs and Patent Appeals ('CCPA'), a body with special jurisdiction within the patent system[31] had already committed itself to a relatively liberal approach to SRI patenting; starting with Prater & Wei[32] and Bernhart[33], it upheld SRI patent applications containing machine and, also, process claims, pointedly stating that the mental steps doctrine did not extend to everything that could be tracked by mental steps and, ultimately, forcing the USPTO to withdraw its 1968 Guidelines[34].
The Benson application and its aftermath: 1970-1972
26. The argument between CCPA and USPTO took on a new dimension with the involvement of the Supreme Court in the case of Gottschalk v Benson[35]. Benson's application related to an algorithm for transforming binary coded decimal numbers (BCD) into pure binary; the computing significance of this was the enhanced ability to convert quickly from decimal, used by humans, to an intermediate stage and then to binary used in computers such as 53 to 0101 0011 (5 and 3 coded in binary) to 110101 in pure binary. The two main claims were to the algorithm (a) applied using a piece of hardware called a re-entrant shift register[36] and (b) recited as a method alone[37].
27. The USPTO rejected the application but, on appeal, the CCPA upheld it as disclosing patentable subject matter: whilst the applicant was prepared to admit that the hardware-based claim did not really relate to any specific apparatus, the court took the view that the claims could be construed as being limited to use on computer devices since they could only be deployed practically in that way[38].
28. In an unusual move, the Supreme Court allowed an appeal and quashed CCPA's decision; in the part of the judgment most often cited, it said that the main claims amounted to nothing more than a sham pre-emption of a mathematical formula that was unpatentable per se[39]. These words were taken to mean by many that SRIs were not patentable and that other forms of program protection had to be sought. However, what is usually missed is that the primary concern of the Supreme Court was the scope of the claims as drafted: they were so " ... so abstract and sweeping as to cover both known and unknown uses of BCD to pure binary conversion"[40]. Secondly, the court itself was at pains to say that the decision was not a denial of the patentability of program elements in all cases[41].
29. One immediate and regrettable side-effect of Benson was a diversion of effort by many lawyers and lobbyists into seeking 'strong copyright' protection rather than looking properly at the effect of the Benson decision and its progeny: following alterations to the Copyright Act 1976 in 1980 to accommodate software products, the race was on to try to stretch copyright to cover, first, program structure[42] and, later, functional aspects of program code and user interface[43].
30. In the event, as charted elsewhere[44], copyright was not up to the task set for it; although admirably suited to protecting static, informational material such as that found in source code, instruction manuals and the like, it showed itself to be manifestly unsuitable for protecting functionality[45], the thing which software producers prized above all else. However, it was only with the 2nd Circuit Court of Appeals decision in Computer Associates v Altai[46] and the later 1st Circuit Court of Appeals decision in Lotus v Borland International[47] that the courts finally halted the distortion of copyright, recognising the problem and restricting scope accordingly; in the former case, it was memorably and aptly observed that putting program functionality under copyright's protection was like trying to put a square peg in a round hole.
31. In the wake of Benson, CCPA did not hesitate from playing 'tug-of-war' over SRI patents: in between Benson and the two Supreme Court decisions listed below, CCPA kept taking a liberal line on patentability issues with the result that, in many cases, those contested SRI patent applications that reached CCPA were granted[48] because of the incredible rarity of Supreme Court intervention[49]. From a practical viewpoint, therefore, SRI patenting continued much as before, albeit at a lower level, not so much because of any real bar but the perception of one generated by overly-broad readings of the Benson decision; nevertheless, it is noteworthy that, even in the 'low ebb' period of 1972-1981, more than 2,000 SRI patents were granted, with 4,000 pending in 1980 alone[50].
32. In 1978, the Supreme Court handed down another decision which seemed, at first glance, to extend the 'bar' which Benson had supposedly set up: in Parker v Flook[51], it rejected a method for calculating and updating alarm limits in hydrocarbon chemical processing as unpatentable in that, when dissected, the invention supposedly embedded in the claims was revealed as nothing more than a mathematical formula dressed up in an algorithm (confusingly described as a 'mathematical algorithm') with post-solution activity for that algorithm being merely 'token' for lack of definition and substantiality.
33. This decision was immediately taken to mean that computer-tested or computer-controlled apparatus and processes were unpatentable: again, however, the court had been at pains to explain that it was not shutting the door entirely[52]. Moreover, when looked at from the point of the scope of the claims as drafted and presented, the application in question was undoubtedly ripe for attack; there was no detail at all on what alarm systems would be used, how they would be linked to the hardware on which the program ran and, most significantly, what range of trigger values would be used.
34. In 1981, the judicial tide in the Supreme Court finally turned; naturally, this was heralded as significant but, given what the CCPA had been doing with USPTO Board of Appeal decisions for the last ten years, it did not have as much impact as might be thought. In Diamond v Diehr[53], the applicant had used a known mathematical formula, the so-called 'Ahrrenius equation', as the basis for automated moulding and curing of rubber. The program which the applicant had written improved on existing manual uses of the equation in that it took information from sensors in the mould to provide a real-time update of numeric variables; the equation was repeatedly cycled through with revised variables to determine the precise moment at which the mould should be released.
35. In its decision, the court acknowledged the flaws in the Flook 'dissective' approach and, having properly taken the claims as a whole, acknowledged them to represent patentable subject matter within section 101; the mere inclusion of program and hardware elements did not render the subject matter unpatentable. The court also limited the meaning of mathematical algorithm by analogy to mathematical formula. In broad factual terms, it is difficult to separate Diehr from Flook but the answer lies in the wording of the claims: the thumbnail description of the invention given above, drawn from the decision, shows that US patent attorneys had learnt their drafting lessons as the actual claims were much more narrowly drawn than in Flook and they were clearly directed to the physical to ensure that there was an easily identifiable functional element to the process[54].
36. From 1981 onwards, the Supreme Court did not intervene in the subject matter conflict and USPTO was left to preserve the crumbling remains of the SRI patentability 'bar' in the face of an ever-growing body of liberal CCPA decisions. Immediately prior to its dissolution in 1982, CCPA gave judgement in a quartet of cases that marked out what it regarded as the boundaries for SRI patentability: In re Pardo[55], In re Taner[56], In re Abele[57] and In re Meyer[58].
37. CCPA's successor, the Court of Appeals for the Federal Circuit ('CAFC'), was established on 1st October 1982 but did not deal with any significant SRI patentability issues until the Grams decision[59]. Shortly before Grams, the USPTO had put promulgated a new set of Guidelines for examinations of applications involving SRIs; they attempted to bring together the whole of CCPA's and the Supreme Court's reasoning in the area and they relied on a two step approach called the Freeman/Walter/Abele test[60].
38. The first part of the test relied on seeing whether the claims recited a mathematical algorithm[61], cautioning that although it would be easy to detect these in process form, it would be more difficult where claims were cast in apparatus form and, in any event, claims should be treated carefully in light of supporting disclosure. If the claims did recite a mathematical algorithm, the second question asked was whether it was applied in any manner to physical elements or process steps; if so, it was patentable subject matter.
39. However, just after the USPTO created its friendlier Guidelines, CAFC promptly extended the basis upon which claims could be framed, thereby widening the gateway of opportunity for SRI patent applicants still further. In Iwahashi[62], the validity of 'means for' claiming was highlighted. Hitherto, SRI patent claims had usually been framed in terms of method or process but this, both inevitably and wrongly, had tainted many of them with the whiff of mathematical formula or mere idea. Here, the applicant had learnt a valuable historical lesson about hardware/software equivalence and claimed its invention for signal autocorrelation not as a method but as 'virtual' machine as follows:
(a) use of 'means for' apparatus claims[63];
(b) a token hardware element comprising a Read Only Memory with a look-up table of square values; and
(c) limitation on claims linked to detailed description such that it was clear that the invention was "for use in a computer for improved voice recognition".
40. The CAFC enthusiastically endorsed the approach, saying that although an algorithm was plainly involved, it had been physically implemented and this gave it proper subject matter status. Subsequently, 'means for' was used to railroad through SRI patents relating to heart monitoring software in Arrhythmia Research v Corazonix[64] and even software for removing jagged edges from digital oscilloscope traces ('anti-aliasing') in Alappat[65].
41. Until the State Street decision, the Alappat decision was treated as the 'gold standard' for patentability of SRIs[66]. The main independent claim in the application, Claim 15, read as follows:
"A rasterizer for converting vector list data representing sample magnitudes of an input waveform into anti-aliased pixel illumination intensity data to be displayed on a display means comprising:
(a) means for determining the vertical distance between the endpoints of each of the vectors in the data list;
(b) means for determining the elevation of a row of pixels that is spanned by the vector;
(c) means for nomalising the vertical distance and elevation; and
(d) means for outputting illumination intensity data as a predetermined function of the normalised vertical distance and elevation".
42. The specification disclosed an electronic circuit but the claims 'means for' format alluded to deployment in software. The Examiner rejected the application for disclosing nothing more than an algorithm whilst a Board of Appeal accepted that the claim as a whole referred to a machine (whether real or virtual). A Full Board of Appeal upheld the Examiner's decision on the narrower ground that, because section 112(6) meant that 'means for claims' covered any and every means, it extended to the interpretation of this claim as a disguised process claim for an unpatentable mathematical algorithm. On appeal to CAFC, it was held that the language used in the specification clearly disclosed a machine (albeit one which could be constructed in software) and, in light of this, the 'means for' claims were suitably conditioned. Moreover, the invention was not unpatentable in any guise because it represented the application of an idea and not just the idea itself.
43. The only slight pauses in the opening up of SRI patenting came in 1994 and 1995 when CAFC was faced with claims which, because of the invention, could not be directed to any kind of post-solution activity i.e. pure program claims. In Schrader[67], the court rejected a method of calculating competitive bids on a series of related items; this was done on the back of CCPA's older tests but, interestingly, the dissenting judgment by Newman J. formed the basis for the new reasoning in Alappat.
44. Bizarrely, with a somewhat more reticent panel than usual, in Trovato[68], the CAFC rejected its much-loved 'means for' claims, this time covering an SRI which calculated the optimum path between two given points by reference to stated parameters such as time, cost, capacity and distance; the post-solution activity here was confined to on-screen display and, as such, the claims were transparently directed towards the program. In language eerily reminiscent of Benson and Flook, the three judges on the initial appeal panel held the apparatus claims to be a mere 'guise' covering an unpatentable mathematical algorithm[69].
45. The next push forward, albeit on a different front, came in 1995 with the decisions in Lowry[70] and Beauregard[71]. Lowry related to a data structure for organisation and use of data stored in computer memory and the USPTO rejected by analogy to a minor exception to patentability relating to printed material; on appeal to CAFC, the applicability of this exclusion was rejected. However, the full significance of the decision did not become clear until Beauregard; here, the USPTO had rejected claims to a computer program on a carrier on the same basis as Lowry and, having learnt, late of the CAFC decision, promptly withdrew. Thus a computer program on a carrier could be patentable as apparatus[72].
46. By 1996, therefore, according to CCPA rulings starting with Alappat and ending with Beauregard[73], most SRIs were easily patentable; where they related to otherwise unpatentable subject matter, if claims were directed towards production of a useful, tangible or concrete result then that was sufficient. Claims could be framed in terms of (virtual) machines, programs stored on media or, if necessary, by reference to processes. The USPTO was left in considerable difficulties as it produced draft after draft of Guidelines to keep up[74] before finally publishing in February 1996.
47. The 1996 Guidelines were drafted with Freeman/Walter/Abele still in mind but with the overall emphasis on finding utility rather than trying to find some physical transformation or effect; this effectively widened protection to the extent demanded by Alappat. In addition, when considering classification of data structures, a form of SRI recognised in Lowry, and computer programs as machine, process, etc., it was permissible to look at storage medium as grounding for a machine claim per Beauregard.
BRIs and ERIs: the U.S. experience
48. Having discussed the bias against BRIs above, it is interesting to note that, in the US, cases cited against patentability were often decided on other grounds, a point which was emphasised strongly in State Street[75]: in other words, there seemed to be misdirection rather than a 'bar'.
49. In some very early decisions such as Munson v City of New York[76] and US Credit System Co. v American Credit Indemnity[77], the issue of patentability was clouded by the issue of refusal of patents to 'printed materials' to avoid reduplication of protection with copyright. However, even in the two cases most often cited for the proposition that "methods of doing business are unpatentable", the decisions were actually made on other grounds.
o In Hotel Security Checking Co. v Lorraine Co.[78], it was held that means for avoiding employee theft by use of a specially-printed receipt slip was not patentable: although the court stated that "... a system of transacting business disconnected from the means for carrying out the system is not [patentable]" was a specific comment on the means for claims in question rather than a general ground for disposal, it was treated as such; in reality, the court decided the case on lack of novelty.
o In Wait[79], the claims related to a process for selling commodities by offering them for sale at remote points. Again, the court decided the issue on the question of novelty and, indeed, specifically stated that it did not have to decide on the issue to dispose of subject matter patentability to dispose of the appeal.
50. Before the emphasis shifted to ERI patent applications, there were some indications that the courts would consider BRI claims: in Deutsch[80], an application to patent a method of operating a system of factories which produced products from raw materials having a plurality of market costs was allowed to proceed by CCPA; in its judgment, it avoided the issue at hand by saying that the invention was not a method of doing business.
51. Even to the extent that the business method exception ever had any substance, it seemed that automation of unpatentable subject matter to create a functional marketing, accounting or advertising system could disclose patentable subject matter just as in the UK IBM decision and within our hypothetical patent system: in Johnston[81], a system for automated cheque clearance was held to be patentable subject matter (though the patent application failed for obviousness) whilst in Paine, Webber, Jackson & Curtis v Merrill Lynch, Pierce, Fenner and Smith[82], a trading system incorporating automated cash management functions was held to be the subject of a valid patent on the basis that that the claims were directed to methods of operating a computer for actual market activities rather than to doing business per se.
State Street: the final push
52. Not content with its previous handiwork, CCPA moved again in 1998. Two years before, two mutual fund companies, State Street and Signature Financial Group, had been negotiating over use of business management software; the latter owned a patent on certain aspects of the software in question. Relations broke down and, seeking to pre-empt an infringement action, State Street sought a declaratory judgement of non-infringement based on lack of patentable subject matter through the District Court for the District of Massachusetts.
53. The patent, US patent number 5,193,056 was entitled 'Data processing system for Hub and Spoke Financial Services Configuration' and was issued on 9 March 1993; its main independent claim was as follows:
"A data processing system for managing a financial services configuration of a portfolio established as a partnership, each partner being one of a plurality of funds comprising:
(a) computer processor means for processing data;
(b) storage means for storing data on a storage medium;
(c) first means for initializing the storage medium;
(d) second means for processing data regarding assets in the portfolio and each of the funds from previous day and data regarding increases or decreases in each of the fund's assets and for allocating the percentage share that each fund holds in the portfolio;
(e) third means for processing data regarding daily incremental income, expenses, and net realized gain or loss of the portfolio, and for allocating data among each fund; and
(f) fourth means for processing data regarding daily net unrealized gain or loss for the portfolio and allocating such data among each fund; and
(g) fifth means for processing data regarding aggregate year-end income, expenses and capital gain or loss for the portfolio and each of the funds."
54. At first instance, the District court held that the patent was invalid for lack of patentable subject matter, under both the mathematical algorithm and business method exceptions. The case went on appeal to CAFC: here was the perfect chance to resolve SRI, BRI and ERI matters for once and for all.
The ‘mathematical algorithm’ exclusion
55. The District court had treated the 'means for' claims as process claims and then promptly declared them poisoned by the algorithmic taint. CAFC stated that the claims had to be viewed as machine claims, directed, as per the specification, towards a personal computer: more importantly, in any event it did not matter which of the five categories at section 101 the inventions fell into since the test for patentability was the same for each. The test to be deployed was a narrow one: given that Congress had intended "everything under the sun made by man to be patentable", the language of section 101 was wide; any exclusion to patentability had to be clearly justified in light of this and also construed narrowly. Accordingly, classification did not matter: if one could identify practical utility and satisfy the other conditions of the Act, that was enough. Mathematical algorithm and the search for post-solution activity were no more[83].
The ‘business method’ exclusion
56. The court held that the origin of the business method exception could be traced in its modern form to Hotel Security Checking v Lorraine[84] in which a system for preventing tip fraud by waiters was rejected as unpatentable[85]. However, as CAFC noted, neither of these decisions relied on the purported exception in making a finding of invalidity and, in any event, such discussion as there was related to mere ideas: there was no disclosed separate exception to patentability. Therefore, unless a business method claim fell within the limited boundaries of mere ideas, it could disclose patentable subject matter.
57. The United States Supreme Court denied certiorari proceedings, thereby tacitly accepting CAFC's approach. In a later case, AT&T v Excel[86], the court emphasised that, in addition to categorisation, it no longer mattered what claim format was used, the wide test applied and was the same for all formats; here, too, Supreme Court certiorari was denied. In short, despite the softly whispered reservations of the USPTO[87], the system had changed: whilst others may regard it as a revolution, from the history set out above, the author would argue that it was nothing more than year ride in full circle.
International, regional and national approaches
International and regional issues
58. Some clues as to the true nature of bars to patentability relevant to SRIs, BRIs and ERIs can be found in international laws and agreements, specifically TRIPS, EC law and the Patent Co-operation Treaty 1970 ('PCT'). In addition, it should be remembered that there is ever increasing co-ordination of approach as encouraged through mechanisms such as the Trilateral Review[88] and the meetings of the heads of the patent offices of certain developed countries[89].
59. In the case of TRIPS, attention has been directed towards Article 27(1) which provides that "... patents shall be available for all inventions, whether products or processes, in all fields of technology, provided they are new, involve and inventive step and are capable of industrial application". Articles 27(2) and (3) provide for limited exceptions but computer programs are not listed therein and the fact that the exclusions are so limited is intriguing: is it possible that, in taking up TRIPS obligations, signatories risk breach simply by virtue of having restrictions on patentability of programs? Bearing in mind that countries which have signed up to the EPC have included such restrictions, this is a worrying possibility. Schiuma has examined this problem by trying to find the true international law meaning of the provisions and seeing how they work in relation to the German Patent Act, one of the EPC-modelled laws[90]. Having concluded that, on any proper interpretation, software must be regarded as an invention in a field of technology, it would seem that German law is, indeed, at variance with TRIPs requirements[91], the country's declaration to the contrary notwithstanding[92]: the more complicated position of the EPO is considered below.
60. Following on from TRIPs, the European Commission recognised the problem with SRI protection and proposed a harmonisation Directive for patent protection within the Member States[93]; this is likely to be implemented in 2001/2002 and it is widely predicted will approximate the threshold of protection to that available in the US as a result of State Street[94]. Indeed, to the extent that Member States form the majority of EPC signatories, they will be obliged to push for harmonisation as between the national, EPC and projected Community patent systems.
61. Finally, in the case of the PCT, it is interesting to note that, whilst the Treaty does not deal directly with the scope of patentable subject matter, in relation to computer programs, Rules 39 and 67 only bar search or examination to the extent that the International Searching Authority or, where relevant, International Preliminary Examining Authority are not equipped to search the prior art[95]; this adds weight to the contention that, having been born out of administrative inconvenience rather than any great principle, restrictions on patentability of programs should be limited to the maximum possible extent.
Law and practice under the EPC
62. Unlike USPA, Article 52 of the European Patent Convention 1973 expressly lists certain things that are not to be regarded as patentable inventions: in ad+dition to discoveries and mathematical formulae, schemes, rules or methods for doing business, performing a mental act and computer programs are mentioned, as is the presentation of information. However, the restrictions are said to apply only to those things "as such".
63. In the early years of the European Patent Office's operation, much more attention was paid to the exclusion than the proviso: as a result, the original 1978 examination Guidelines set a threshold that SRIs, BRIs and ERIs would find difficult to overcome[96] with the result that the EPO acquired a continuing reputation as a place through which such inventions could not be patented. In 1985, however, the Guidelines were changed to take account of industry concerns; although tied to claiming the program when loaded in the computer, they did at least allow that if a technical (in the sense of "technological") contribution was made to the known art by the invention, patentability should not be denied merely because it was implemented in a computer program.
64. Starting with Vicom/Image processing[97], the EPO's Boards of Appeal slowly began to sketch the boundaries between, on the one hand, a computer program as such and, on the other, an invention making a technical contribution albeit that it was implemented in software; however, the general trend was to place the marker ever further in favour of the latter at the expense of the former[98]. In Vicom itself, a revised claim to a method of "digitally processing images in the form of a two dimensional array" was allowed because the image was a physical entity and it was directed towards a technical process affecting it. Following on from the straight physical effect test applied in Vicom, the EPO also began to accept applications where the invention overcame a technical problem; this might involve maintaining or improving the operation of hardware or software as in IBM/Data Processor Network Appn.[99].
65. However, although whittled down, significant obstacles remained: claims other than to a machine as programmed were severely limited and, by virtue of unfortunately similar terminology and concepts, issues of technical contribution tended to shade off into issues of novelty and obviousness. In a pair of decisions handed down by a Technical Board of Appeal in 1998 and 1999, a determined effort was made to bring clarity to the matter and to assimilate practice as far as possible with international norms (including TRIPS): IBM/Computer programs[100] and IBM/Computer programs II[101] were to set the 'gold standard' for the EPO.
66. Both cases revolved around software operations: in the first, the claims, both to methods and apparatus, related to a system for detecting overlaps of windows on computer screens and preventing the obscuring of information by causing display elsewhere whilst, in the second, they related to a system for recovery of memory and other resources including procedural failsafes in the event of non-recovery. In addition to the acceptable claim forms, the applicant had also included claims to program stored 'on medium'.
67. The Board explained in both instances that the mere fact of alteration of current flow in the circuits by loading the program was insufficient to show 'technical character', the replacement test requirement for 'technical contribution', of the sort to avoid the exclusion of programs as such. Instead, the point at which invention should be considered should be that at which the program was executed to produce 'further effects'; whether inside the machine or out was irrelevant, whether in hardware or software was irrelevant provided such further effects could be found. The effect was to limit the computer program as such limitation to the lowest possible level.
68. However, despite the change in standard, the Board recognised that its new approach would still leave 'on media' claims of the Beauregard type high and dry; recognising the illogicality of having something with the potential to have further technical effect left as unpatentable, it accepted that such claims should be assimilated and accepted. In addition, in IBM/Computer programs II in particular, the Board acknowledged that Article 27 TRIPS was an issue. However, it held that since the EPO was neither a party to nor otherwise bound by TRIPS, all that it could and should do was to acknowledge it as something to take into account when interpreting Article 52.
69. Turning to BRIs and ERIs, the EPO has had some dealings with the latter but, in view of requirements of 'technical contribution' and, latterly, 'technical character', some applications have fallen. However, there are two successful applications which show that, where suitable technical character can be identified, the business method bar is not absolute.
o In Sohei/General purpose management system[102], apparatus and method claims were made for a system of administrative and management task handling via a common booking slip for data storage and processing; as amended, the Board of Appeal characterised as relating to a user interface rather than a computer program as such, a method of doing business as such and/or a presentation of information as such.
o In Pettersson/Queuing system[103], a system for determining the queue sequence for customer service at a plurality of service points was said to be a method for doing business. However, the Board held that the main claim, Claim 1, related to apparatus, specifically a turn-number allocation unit, and not any method of use as such.
70. Ultimately, as with the discussion of the US situation, the point should be made that, in practice, many SRI and ERI applications are being accepted without subject matter objections; Cohen[104] notes that by 1999, approximately 100 out of 15,000 applications had been rejected in this field or, put another way, 99.33% overcame the subject matter threshold. Indeed, following concerns in the 1990s[105], practitioners are taking the view that difference between the EPO and USPTO on SRI patentability is negligible but, with ERIs, care still must be taken in drafting to identify some genuine and relevant 'technical effect'[106].
71. In the case of the UK, it is ironic to note that after the entry into force of the Patents Act 1977, working from the same exclusions as the EPO[107], the courts and, perforce, the Patent Office have developed limits on patentability of SRIs, BRIs and ERIs which are recognised as being stricter than their Continental counterparts[108]; this is so despite the relaxed approach previously taken under the 1949 Act[109].
72. The first hint of the problem came with Merrill Lynch Inc's. Appn.[110] in which an application directly corresponding to the US trading system application noted above[111] was rejected. Although the Court of Appeal rejected a dissective approach to claims analysis, it held that the invention as characterised by the claims was a method of doing business. Although the reasoning of the analogous IBM decision under the 1949 Act was raised, the court declined to follow it, failing to recognise the difference between business schemes, rules or plans and their realisation and the technical effect inherent in the virtual marketplace that the software created. However, it is worth noting that in the event, a limited set of apparatus claims was allowed to stand[112].
73. In the case of SRIs, whilst few doubted the correct decision had been reached by the Court of Appeal in Gale's Appn.[113] in which the mode of storage of a program was held to be irrelevant to the patentability of the invention as claimed, a quartet of later applications[114] were refused on the basis that they related to computer programs as such and/or schemes, rules or methods for performing a mental act as such; the latter objection is highly questionable on the basis that the judges and examiners concerned seem to have entirely overlooked the "as such" qualifier to section 1(2) in relation to it and the point that the subject matter as claimed could not be carried out mentally[115].
74. The overly strict approach which the UK Patent Office was obliged to follow was a source of concern during the 1990s[116] though, when actual output of the Office was compared in terms of applications proceeding to grant, the comparisons with the US, Germany and Japan were less invidious[117]. However, with the two pronged assault of TRIPS and the EC harmonisation Directive on the horizon, there should be no problem in future and, even in the interim, the Patent Office has issued a examination Guideline indicating its recognition of the recent IBM decisions in the EPO and its allowance of Beauregard-style claims[118]; this has particular advantages since the protection afforded to processes under the 1977 Act is much more limited than that offered to products[119].
75. In the case of BRIs and ERIs, the situation is still not so clear: under the 1949 Act, some BRI applications were rejected as disclosing either mere schemes, rules or plans[120] and/or on the basis of the printed matter exception[121]. Although the transition to the new Act meant that many old authorities were treated as being of doubtful relevance[122], BRI patent applications have continued to be treated with suspicion: on a broad construction of business, a corporal punishment scheme was rejected[123]. In light of the change in relation to SRIs and the shift to ERI patent applications, in the immediate future, the UK Patent Office is likely to take its cue from decisions such as Sohei and Pettersson.
76. In contrast to both US and EPC rules, Japanese patent law is marked out by a positive and relatively specific approach to the definition of invention: Article 2, Patent Law states that an invention is "the highly advanced creation of technical ideas by which a law of nature is utilised"[124]. However, from this starting point, there is rapid convergence.
77. As far as SRIs are concerned, the JPO's Guidelines[125] state that:
(a) claims should be taken as a whole when considering the issue of invention;
(b) utilising a law of nature in information processing by software is patentable;
(c) utilising a law of nature in hardware resources is patentable; and
(d) in determining patentability, it is irrelevant whether the invention is claimed as a process or product.
78. This seems to set the patentability threshold at approximately the same level as the U.S.[126] and it is considered unlikely that the Tokyo High Court will set any different standard in deciding future cases[127], particularly in view of the fact that Japan has one of the largest sets of SRI patents and applications in the world.
79. Turning to BRIs and ERIs, although the JPO's Guidelines, currently under review[128], are relatively generous towards BRIs and, by extension, to ERIs, there is, by contrast, a distinct shortage of patents granted when compared to the US; the reason for this appears to be a reluctance on the part of the relevant commercial players to apply[129].
The real problem with SRIs, BRIs and ERIs
Novelty, obviousness and utility
80. The greatest irony in this all is that the practical reasons proffered by the 1966 Commission Report on the US patent system for restricting SRI patenting has continued to hold true: a lack of experienced examining staff, time and prior art resources has meant that the quality of many US patents granted has been very suspect indeed.
81. Perhaps the most notorious example in recent time has been Compton's multimedia patent, number 5,242,671; this related to a database search system which used multiple graphical and textual entry paths and the width of the claims was such as to make any multimedia database system infringing, something which became clear when Compton's began demanding 1-3% royalties on multimedia database sales from the industry at large[130].
82. One slight problem with Compton's patent was that since the time of the development of hypertext in the 1960s, such systems had been known and there was relevant prior art in software engineering journals which showed virtually identical systems fully operational by 1986. Unfortunately, this had not been noticed by the USPTO which had proceeded to grant in August 1993.
83. The outcry from industry was so great that the then-Commissioner for Patents, Bruce Lehman, took the rare step of calling the patent back in for re-examination: on 23 March 1994, in light of the voluminous prior art, all 41 claims were rejected. In the aftermath, it was promised that examination would be improved by having tighter internal quality controls and more, better-qualified staff.
84. The only problem has been that this has not happened: in March 1999, the USPTO was forced to put a statement assuring all concerned that SRI, BRI and ERI handling would be improved and that new staff would be taken on[131]. Has this change of handling worked well? Judging by the state of US patent 6,041,310 for a system of selling cars via a computerised kiosk, probably not.
The chilling effect of patents
85. One aspect of the 'new' approach to SRI, BRI and ERI patents which disturbs some critics is the possibility of 'chilling' or retardation of innovation in the relevant technical or commercial field; a recent case, often cited in this regard, is Amazon.com Inc. v Barnesandnoble.com Inc. [132], in which Amazon's patent for a method and system of 'one click' Internet shopping once the product to be ordered is displayed to the consumer came under the microscope; it was found to be valid and infringed but, subsequently, the founder of Amazon, Mr Jeff Bezos, gave a speech in which he acknowledged that he was worried about the possible slowing down of e-commerce spread because of his company's patent[133].
86. The social concern exuded by Mr Bezos was pleasant but, it is submitted, not relevant: it was and, subject to exclusive dealing agreements, remains possible for Amazon to licence the technology out. Furthermore, assuming that the court was right to uphold the patent as valid, it also follows that there is no chilling issue: rather, the infringer was trying to get a free technical and/or commercial 'leg up' as all patent infringers seek to do. Similarly, although claims interpretation and infringement issues do arise in this case, this is equally true in any other field. However, it must be acknowledged that in the case of 'bad' patents i.e. those granted without proper, strict examination as to novelty, obviousness and utility/industrial applicability, a genuine chilling issue does arise.
87. In conclusion, Patent Offices around the world must sharpen up examination procedure as much as possible to see that rogue SRI, BRI and ERI patents are not granted; whilst it is possible to challenge for invalidity, the cost and uncertainty of doing so still hangs over others in the field and, here, the chilling effect is real. Examiners, particularly those in the US must use all information resources available to them - including the Internet - and not just rely on patent application or patent documents[134] However, with real effort, the quality of software patents can be improved and those critics who continue to object to the 'new' system on examination and fee-related issues[135] can ultimately be consoled.
* Research Fellow, Queen Mary Intellectual Property Research Institute, University of London.
[1] [1998] USCAFED 107; 149 F.3d 1368 (Fed. Cir., 1998), cert. denied 119 S. Ct. 85 (1999)
[2] See Kang et al., " A practitioner's approach to strategic enforcement and analysis of business method patents in the post-State Street era" 40 Idea 226 (2000).
[3] E.g. Raskind, "The State Street Bank decision: the bad business of unlimited patent protection for methods of doing business" 10 Fordham Intell. Prop., Media & Ent. LJ 61 (1999).
[4] 56 US 62 (1853) - Morse's telegraph patent. Claim 8 as filed claimed the concept of electronic communications without limitation as to apparatus; this was, unsurprisingly, rejected.
[5] [1948] USSC 22; 333 US 127 (1948) - Claims to a mixture of strains of bacteria used in plant inoculation.
[6] Thus, absent any novelty, obviousness or utility problems, whilst X-rays were not patentable, a machine for taking X-ray photographs would be.
[7] I.e. non-natural, the result of human intervention - later sonorously described by Congressional Reports accompanying USPA as 'anything under the sun made by the hand of man': see OTA, "Finding a balance: computer software, intellectual property and the challenge of technological change" US Govt. Printing Office, Washington DC, 1992 at p39.
[8] I.e. X-rays simply are whilst an x-ray machine does.
[9] A point examined in relation to the JPO and EPO rules examined below: for an Australian practitioner's view that the Australian, Japanese, US and EPO practice on software patenting was, essentially, harmonised by mid-1999, see McFarlane, "How to patent software" 113 PW 24 (1999).
[10] See Beresford, "Patenting software under the European Patent Convention", Sweet & Maxwell, London, 2000 at p8.
[11] See Appleton, "What is software?" 2CL&P 82 (1986).
[12] See von Hellfeld, "Hardware, firmware, software" 9 CL&P 18 (1993).
[13] Section 101, 1949 Act; this imported the traditional phrasing from the Statute of Monopolies 1623/4.
[15] i.e. under the broad test then taking root in the wake of NRDC's Appn. [1961] RPC 134 (Aus.).
[16] See Molyneaux, "A history of patenting software related inventions in the United Kingdom" [1996] CTLR 11.
[19] Here, cardboard punchcards.
[21] I.e. those with advanced communications functions.
[22] Judgment at 158.
[24] As opposed to 'methods used in business', a term which may encompass virtually any invention.
[25] See, further, for real world considerations, Fuelling, "Manufacturing, selling and accounting: patenting business methods" 76 JPTOS 471 (1994) at p477.
[26] See n25 above.
[27] Ibid. at p476; this point was made by the US Court of Customs and Patent Appeals in In re Johnston 502 F.2d 765 (CCPA, 1974), rev'd on other grounds sub nom Dann v Johnston 425 US 419 (1976).
[28] See Jacobs, "Computer technology (hardware and software): some legal implications for antitrust, copyrights and patents" 1 Rutgers J Comp Tech & L 50 (1970).
[29] See "To Promote the Useful Arts" S. Doc. No. 5, 90th Congress, 1st Sess. (1967).
[30] See Kayton, "Patent protectability of software: background and current law" 8 Jurimetrics 127 (1969).
[31] The CCPA handled not just appeals from inferior courts but also the Boards of Appeal of the USPTO.
[32] 415 F.2d 1378, aff'd. 415 F.2d 1393 (1969) (method of spectrographic analysis utilising linear equations solved by general purpose digital computer). On rehearing, CCPA upheld the 'computer-as-programmed' machine claims in the application but rejected the process claims as lacking limitation to machine implementation as described.
[33] 417 F.2d 1395 (1969) (method of plotting 3D objects on planar surface utilising transformation equations solved by general purpose digital computer). 'Computer-as-programmed' machine claims upheld.
[34] See, on similar grounds, reversals in Musgrave 431 F.2d 882 (CCPA, 1970) (method of signal generation and correction) and Mahoney 421 F.2d 742 (CCPA, 1970) (method of bit stream synchronisation).
[35] [1972] USSC 223; 409 US 63 (1972).
[36]
Claim 8.
[37]
Claim 13.
[38]
441 F.2d 682 (1971).
[39]
Judgment at 71.
[40]
Judgment at 68.
[41]
Judgment at 71.
[42]
Cf. Whelan
Associates Inc. v Jaslow Dental Laboratory Inc. [1986] USCA3 976; 797 F.2d
1222 (3rd Cir.,
1986).
[43] Cf. the first instance decision in Lotus Development Corp. v Paperback Software International 740 F.Supp. 37 (D.Mass., 1990)
[44] See Lea, "Software protection trends in the 1990s" 7 Ent. LR 276 (1995).
[45] Something, ironically, predicted by Commissioner Hershey when filing a dissent from the CONTU Report which paved the way for the 1980 copyright amending legislation.
[46] 20 USPQ 2d 1641 (1992).
[47]
34 USPQ 2d 1014
(1995).
[48]
See, for example, In
re Noll 545 F.2d 141 (CCPA, 1974) (apparatus claims for
graphics display
upheld) and In re Toma 575 F.2d 872 (CCPA, 1978)
(machine translation).
[49] See Peterson et al., "Patenting software" 31 Idea 181 (1990). Apart from the cases listed below, only in Dann v Johnston [1976] USSC 53; 425 US 219 (1976) did the Supreme Court encounter an SRI patent; interestingly, it dealt with the case on grounds of obviousness under s103 and did not deal with any of the patentable subject matter arguments in s101.
[50] See Soma & Smith, "Software trends: who's getting what of how many?" 69 JPTOS 415 (1987).
[52]
Judgment at 595.
[53]
[1981] USSC 40; 450 US 175 (1981).
[54]
See Schmidt,
"Legal proprietary interests in computer programs: the American
experience" 21 Jurimetrics 345 (1981).
[55] 684 F.2d 912 (CCPA, 1982) (method of converting computer to parallel processing under compiler control; this was characterised as an active control component of program execution rather than an unpatentable algorithm).
[56] 681 F.2d 787 (CCPA, 1982) (method of simulating seismic waves by summing signals of conventional waves; this was characterised as a physical conversion of summed signals and was, therefore, a patentable process).
[57] 684 F.2d 902 (CCPA, 1982) (improved method of CAT scanning; mere inclusion of computer program did not render invention unpatentable).
[58] 688 F.2d 789 (CCPA, 1982) (method and machine for storing and comparing medical test results; not applied to physical elements or process steps and, therefore, unpatentable).
[59] [1989] USCAFED 625; 888 F.2d 835 (Fed. Cir., 1989) (method for automated diagnosis from clinical test results; gathering and summation of data as disclosed did not amount to patentable process).
[60] Derived cumulatively from In re Freeman 573 F.2d 1237 (CCPA, 1978), In re Walter 618 F.2d 758 (CCPA, 1980) and In re Abele 684 F.2d 902 (CCPA, 1982): see Kunin, "Patentability of computer related inventions in the USPTO" 77 JPTOS 833 (1995).
[61] I.e. in the Diehr sense.
[62]
[1989] USCAFED 628; 888 F.2d 1370 (Fed.
Cir., 1989).
[63]
Permitted by s112,
USPA.
[64]
958 F.2d 1053
(1992).
[65]
[1994] USCAFED 665; 33 F.3d 1526 (1994).
[66]
See, though, for a
critical view, Stern, "Patenting algorithms: the Gordian Knot
retwisted" [1995] CTLR 13.
[67] [1994] USCAFED 373; 22 F.3d 290 (1994).
[68]
36 F.3d 1376 (1994).
[69]
Subsequently, some
amelioration was later gained in an en banc hearing.
[70] 32 F.3d 1579 (1994).
[71]
[1995] USCAFED 610; 53 F.3d 1583 (1995).
[72]
See Laurenson,
"Computer software 'article of manufacture' patents" 77 JPTOS 811
(1995).
[73] See Trzyna, "Patent protection for computer programs: scientific wisdom of US law turnabout tolls the bell for the world" 77 PW 13 (1995).
[74] See, for details of the various changes in policy, Stewart, "Patenting of software - proposed Guidelines and the magic dividing line that disappeared" 77 JPTOS 681 (1995) and Gregory et al., "The US Patent Office Guidelines for computer-implemented inventions" [1996] EIPR 56.
[75] See Section 6.02 below.
[76]
[1888] USSC 69; 124 US 601 (1888).
[77]
59 F. 139 (2nd Cir.,
1893).
[78]
160 F. 467 (2nd
Cir., 1908).
[79]
24 USPQ 88 (CCPA,
1934).
[80] 193 USPQ 645 (CCPA, 1977).
[81] See nn. 29 and 47 above.
[82]
218 USPQ 212 (D.
Del., 1983).
[83]
See King,
"Abort, retry, fail: protection for software-related inventions in the
wake of State Street Bank & Trust Co v Signature
Financial
Group, Inc"
85 Corn. L R 1118 (2000).
[84] See Section 5.01 above.
[85]
See Thomas, "An
epistemology of appropriation: patentable subject matter after State
Street
Bank" [2000] IPQ 27.
[87]
See Kunin,
"Patent eligibility in view of State Street and AT&T v Excel
Communication" 81 JPTOS 671 (1999).
[88] See details of this EPO/USPTO/JPO consortium at http://www.uspto.gov/web/tws/index.html.
[89] See notes of the Second Meeting at http://www.patent.gov.uk/snews/notices/meeting.html.
[90] "TRIPS and the exclusion of software "as such" from patentability" 31 IIC 36 (2000).
[91] Ibid., p45.
[92]
Ibid., p43.
[93]
See para. 3.3,
"Promoting innovation through patents: the follow up to the Green
paper" at http://www.europa.eu.int/comm/internal_market/en/intprop/indprop/8682en.pdf.
[94] Ibid.
[95]
See Beresford, op.
cit., p17.
[96]
In the case of SRIs,
for example, the Guidelines stated that claims should be rejected if
the
contribution to the known art resided
solely in a computer program,
regardless
as to how presented in the claims.
[97] T208/84 [1987] EPOR 74.
[98] To trace the progression over a relatively short time, see Newman, "The patentability of computer related inventions in Europe" [1997] EIPR 701, Davies, "Computer program claims: the final frontier for software inventions" [1998] EIPR 429 and Lakkonen & Whaite, "The EPO simplifies software patenting" [1999] IPQ 87.
[99] T6/83 [1990] OJ EPO 5
[100]
T935/97 [1999]
EPOR 301.
[101] T1173/97 [2000] EPOR 219.
[102] T769/92 [1996] EPOR 255.
[103] T1002/92 [1996] EPOR 1.
[104] "The patenting of software" [1999] EIPR 607.
[105] See Harris, "Report on Union Round Table at the EPO" CIPA Feb. 1998, p107.
[106] See Godar, "Who's been patenting your business method" 21 BLR 66 (2000) and Dagg, "How to recognise when a software innovation or method of doing business is patentable in Europe: Part I" (1999) 11 JIBFL 477.
[107] Compare s1(2) PA 1977 to Art. 52(2) and 52(3), EPC.
[108] See Davis, "Patentability of computer programs" 149(6916) NLJ 1872 (1999).
[109] A point noted by Blakemore, "Patent protection in UK and Europe for computer program related inventions" 77 PW 21 (1995).
[111] See Section 5.02.
[112]
See Molyneaux,
"Patenting software-related inventions in the UK: Part II" [1996]
CTLR 51 at p53.
[114] See Wang Labs. Inc's. Appn. [1991] RPC 463 (apparatus claims to expert system treated as claims to program as such), Hitachi Ltd's Appn. [1991] RPC 415 (process claims to mixed vector/scalar program compilation treated as claims to program as such), Raytheon's Appn. [1993] RPC 427 (method and apparatus claims to automated ship silhouette recognition system treated as claims to scheme, rule or method for performing mental act) and Fujitsu Ltd's Appn. [1997] EWCA Civ 1174; [1997] RPC 608 (method and apparatus for synthetic crystalline CAM unpatentable).
[115] I.e. the characterisation of the inventions was too abstract; restricting this would have meant the difference between patentability and unpatentability in the Raytheon and Fujitsu cases (n. above), in the latter case because the claim that the application did not relate to a program as such was upheld on appeal. Note that this latter point was welcomed as an advance by many practitioners: see Chahil, "Software patents: has Fujitsu taken us any further forward?" 95 PW 15 (1997).
[116] See Blakemore, "The UK Patent Office Forum: legal protection for software related invention" 69 PW 39 (1995): for a more sanguine view of slightly later vintage, see Beresford "The patenting of software in Europe and the UK" 91 PW 14 (1997).
[117] A point raised by Deputy Director Haselden, "Patents for software: the view from the British Patent Office" 105 PW 33 (1998).
[118] This is dated 19th April 1999 and it is available at http://www.patent.gov.uk.
[119] See, in particular, s60(1).
[120]
Cf. Stahl
& Larson's Appn. [1965] RPC 596 (method and
apparatus for passenger
journey tracking).
[121] Cf. Johnsons' Appn. 19 RPC 56 (1902) (apparatus for business correspondence).
[122] See para. 1.21, Patent Office Manual.
[123] Melia's Appn. (SRIS O/153/92).
[124]
Law No. 121 of
13th April 1959 as amended.
[125] Chapter I, published February 1997.
[126] See Mashima, "Examination of the interrelationship among Japanese IP protection for software, the software industry and keiretsu" 82 JPTOS 50 (2000).
[127] The court, in effect, operates a supervisory jurisdiction over the JPO roughly analogous to the CAFC's in respect of the USPTO.
[128] See press announcement for December 1999 at http://www.jpo-miti.go.jp.
[129] See Tennesohn and Yamamoto, "Exporting State Street" 94 MIP 3 (1999).
[130] See Rebeiro, "Compton's multimedia patent: should the patent system be used to protect software?" [1995] CTLR 81.
[131] See USPTO White Paper, "Automated financial or management data processing (business methods)" 19th July 2000 at http://www.uspto.gov/web/menu/busmethp/index.html.
[132] 53 USPQ 2d 1115 (WD Wa., 1999).
[133]
See Oakes,
"Another Amazon patent furor" at http://www.wirednews.com/news/print/0,1294,34670,00.html
[134] See, for more on examination practices in the US, Frome, "Prosecuting cyberspace patents" (1999) at http://www.uiowa.edu/~cyberlaw/cls99/sempaper/fitzg331.html
[135] See, for example, Grover,
"Software patents - damage limitation" 16 CLSR 39 (2000).
AustLII:
Copyright Policy
|
Disclaimers
|
Privacy Policy
|
Feedback
URL: http://www.austlii.edu.au/au/journals/DigTechLawJl/2000/4.html