To refresh your memory, the purpose of the workshops, as you will see in the UKPO's press release, was limited to trying to define terms, or as their spokesman phrased it, "to see if a definition which is clear to software developers can be found which continues to enable the patent system to protect technical inventions."

There was no question on this table as to whether or not to patent software, only how to carefully phrase it to clarify which types of software-related inventions are patentable and which are not. Currently, computer-implemented inventions can only be patented if they are novel, inventive and make a "technical contribution" but what is the line or clear definition of "technical contribution"? The workshop hoped to figure that out.

Here is an excerpt from the press release, explaining the purpose of the workshops:

Patent workshops for the software industry

Today the Patent Office is inviting software developers and patent professionals to a series of workshops to explore how to define the contentious border between patentable and non-patentable software-related inventions. With many thousands of patents granted in recent years for telecommunications, data processing and control systems the economic importance of this border could not be greater.

There is a currently a debate in Europe over a Directive concerning patents for "computer-implemented inventions" ("CII"). European Governments believe that a Directive is needed to clarify the law as to which types of software-related inventions are patentable and which are not. The current case law states that only CIIs which are novel, inventive and make a "technical contribution" are patentable. However, many software developers find this definition confusing, and ask "What is a 'technical contribution'"?

Peter Hayward, Divisional Director at the Patent Office, said:

"These workshops are to focus in detail on the issue of 'technical contribution'. We want to see if the definition evolving through numerous court precedents is the best one to formalise in the CII Directive. We will be very interested to see if a definition which is clear to software developers can be found which continues to enable the patent system to protect technical inventions."

"We need to hear opinions from a broad range of interests - not just from patent attorneys and patent-owning software companies. The views of those software developers who work without patents are just as important to us. Opposing views on the Directive have been expressed by different sectors of the software industry, and not just along the traditional division between the large and small firms."

Here is Sagitta's report on how it went. There will be more workshops to follow, and we expect to have more reports to share with you.

****************************

Report From One of the UK Patent Office's Technical Contribution Workshops (Bristol)

~ by Sagitta

The workshop took place in a conference centre in Bristol (although the building was many decades old and must have originally served a different purpose). The exercise was scheduled to start at one o'clock, with registration and tea available from half past twelve, so naturally I arrived at exactly 12:30. I found a handful of people already taking refreshments and debating the merits of software patents.

One guest was trenchantly (but politely) criticising software patentability, and in particular some existing European patents. A member of the UKPO staff expressed the opinion that some examiners in the European Patent Office had awarded a small number of patents they should not have done, and that an aim of the Directive was to stop this, although he didn't say whether the Directive should invalidate these patents. He also commented that making all software unpatentable would invalidate a large body of existing patents, which was (he implied) out of the question. I felt the implication of these remarks was that patent law could only ever progress in one direction: greater patentability.

I mention this conversation not just to show that UKPO staff have opinions like all other human beings, but because it explains a possible motive behind the reluctance to amend the directive. If it is deemed to merely codify existing case law (regardless of whether, in fairness, it actually does), there can be no objection to applying it retroactively and enabling a few dodgy claims to be weeded out. You may wonder, of course, whether this is sufficient reason to change the statutes of 25 nations so comprehensively.

I moved on to chat with perhaps the most genial and casually dressed person at the workshop. He turned out to be a patent lawyer, which proves the danger of judging by appearances. Another two other legal gentlemen standing close by were kind enough to reinforce the stereotype with guarded, incisive conversation, charcoal grey suits, and, to my eyes, a lean and hungry look.

When asked, I mentioned that I was in favour of patents, but against software patents, and the affable solicitor asked me why my industry was different from all others. A fair question, and I could have answered that companies already write innovative software without applying for patents, that a competitive market is probably a greater spur to productive research and certainly more beneficial overall to consumers, that the administrative costs and legal risks are especially burdensome where a small company or a private individual can create a new product quickly and easily, and that potentially there could be dangerous implications for research and free speech if a description of an algorithm becomes an infringement. However, on these occasions I tend to operate under the delusion that I'm channeling Socrates, and so instead the conversation proceeded something along these lines:

- Why should software be treated differently from other literary works, like films and books?
- Because software is technical, films and books aren't.
- Surely books can be technical, such as a technical manual or a textbook?
- Perhaps there should be patents for technical books, then.
- Why shouldn't software just be protected by copyright, as per the Directive on the legal protection of software programs? Or as an industrial design?
- Because you need an incentive to innovate.
- If we spend 12 months developing a product, then we're still 12 months ahead of the competition, and that's incentive enough.
- No, it would only take them 1 or 2 months, because they can look at what you've done and copy it.
- But if they look at the internal workings of my software and write something that does the same thing, isn't that a derivative work?
- Not necessarily. [An excellent lawyerly answer.] Copyright doesn't protect the idea, only the expression. You can't have a copyright on 'a girl goes into the woods with an apple for her grandmother...'
- Hmmm, I think I read a court suggested you can, although the case was decided for other reasons. ¹So you think software patents should cover the processes that the software uses to achieve its effect?
- No, a patent should cover the effect.
- So if you have a patent on a car, then a boat would infringe it, because they both have the effect of getting you from A to B?
- The boat wouldn't be novel.
- But if it were?
- Then the boat would infringe the patent.
- Even though it doesn't use any of the same methods?
- In that case the patent was written badly. It shouldn't say 'a car', it should say 'a device for getting from A to B'.

I decided my interlocutor could not be accused of inconsistency or clouded thinking. But it was now time to get into groups for discussion, and I found myself assigned to a table with four other people, apparently at random. I was introduced to another professional involved in software development (but I forget his precise job). There was also an academic who expressed a degree of pride in his role in developing a program comprising 5 million lines of FORTRAN. Beside me sat a patent lawyer, who indicated, with apparently unintentional irony, that he did not deal with software cases 'as such'.

I introduced myself, explaining that I'm a software engineer, and have spent the past 3 years working on embedded applications for a large multinational company. I added that I wasn't representing my employer (nor am I representing them here), my aim was to safeguard my own career as a software engineer, and more generally my interests as a consumer. The fifth gentleman explained that he had come as an observer from Intellect, and true to his word he sat calmly and silently throughout the session, as inscrutable as a CCTV camera but not so unnerving. During the group discussion he appeared to be interested mostly in our table, but he had positioned himself to view everyone in the room.

The persons present included 30 guests: young to middle-aged adults, of whom approximately 2 were female. Our hosts implied that probably most of us were software developers, but when I asked later they didn't know exactly how many attendees belonged to which professions. There were three gentlemen from the UKPO (all examiners I think), and a lady who was just as quiet as our Intellectual, and worked probably for the UKPO or possibly for the conference centre.

One of the UKPO staff addressed us to explain the background to the workshop. He emphasised that the UKPO was hosting this event on the instructions of the government, and that like any other civil servant his role was to follow the instructions of the government, not his own initiative. His modest stillness and humility was not marred when on one occasion he accidentally referred to himself as a 'minister of the government'.

He explained that following the public meeting last December, Lord Sainsbury had decided to explore alternative definitions of the term 'technical contribution'. He had directed the UKPO to undertake further consultation, and as a result the Patent Office had received more than 200 suggestions on the definition of the term, and many of these would be tested in this series of gatherings. We were told 'the purpose of these workshops is to see whether we can find a definition that gives broadly the same results as the current "technical contribution" test in European law but is clearer to apply.' The UKPO did not claim to be aiming for anything more satisfactory than current case law for the industry and the economy as a whole, but note the word 'broadly' permits plenty of room for manoeuvre. The original press release says much the same, but is more carefully spun. In both cases, the stated aim is clarity, not palatability.

The main exercise of the day began: we opened (and in some cases disassembled) our booklets, and began reading the questions set by the examiners, conscious of the strict time limit. But unlike most exams, we were encouraged to talk amongst ourselves as we attempted to solve the tricky problems presented. Given abstracts of five hypothetical patent applications, and four alternative definitions of the term "technical contribution", we had to decide whether each invention fulfilled each definition, and thus (given the assumption that they were novel and inventive) would be patentable. Our progress was slow at first, as we struggled to digest the complex phrasing and terminology of the definitions.

The UKPO's methodology is to test whether self-selected members of the public are able to apply these alternative definitions easily and consistently, given very limited time and no opportunity for background research. I'm not sure how relevant this is to ensuring that they give clear and consistent results in the courts: it appears that the workshops are seeking a definition that is clear and simple for the lay person, not (just) one that is most helpful to lawyers. These two criteria should coincide, of course. But I didn't have any confidence that a patent law specialist, equipped with a long history of precedents, proceeding methodically through a claim, would arrive at the same conclusions I did in a snap decision. We might end up with a definition that gives a clear and consistent result at first glance, but a contrary (though consistent) result on detailed examination.

For the benefit of future workshops I'm not going to reveal the precise phraseology we were given (which would also violate the UKPO's copyright). Other sessions will use different definitions for the most part, except that the wording currently proposed will be considered in all of them. One of the groups yesterday was forearmed with a legal opinion outlining the current interpretation of a term used in some of the definitions, but such cribbing rather defeats the aim of the exercise.

The sample patents are all computer-related, and mostly software patents in the sense that in inventive step consists of using new/different programs or data. Some were 'pure' software patents in the sense that they didn't mention any specific physical output. The examples will probably be reused in the future, so I shan't be specific about them now. The claims were much easier to understand than the definitions (if only all real patent claims were as comprehensible.) Because there were several competing definitions, and because the intention was clarity, the claims were not written to meet the specific terms. There were some cases where I suspected an effectively identical claim could have been made in different words, and passed instead of failing.

Although comparatively clear, the examples did require careful reading. We interpreted one of the patents as covering the use of an algorithm, but looking at it again today I think it only covers a particular hardware implementation. I hope others don't make the same mistake, because that distinction is important for some of the (IMO better) definitions. Another claim involved a sequence of calculations, but didn't say that they were performed by a computer. And I thought there was an example of a design technique trying to pass itself off as a manufacturing process. The easiest claims to decide were two very pure software patents covering internal features of applications.

Fortified with all the self-confidence of a legal ignoramus who has spent several hours researching a particularly dull field of EU and UK law, I took it upon myself to very gently guide the discussion at my table, and see if we could agree on a firm conclusion for each of the twenty combinations of claims and definitions. I hesitate to call myself a chairman; I prefer to think of us as a gang of four. Thus it's my fault if, as I suspect, my partners also misunderstood the algorithm/implementation patent. (Likewise, they share some of the responsibility for my error -- all very democratic.)

Our hosts hurried us along, afraid that we would not finish in time. Progress was slow at first, as we struggled to comprehend the subtle language in front of us, but once we had assimilated the information we were able to proceed at a steady pace, and finished with a couple of minutes to spare. The time constraint was a problem, but perhaps one of the goals was to simulate the pressures faced by a patent examiner. We treated each patent in turn, which I think was a methodological error. The definitions were more complex and subtle, and so it would have been easier to keep a single definition in mind while working through each of the patents.

When our time was up, the number of judgements for and against a technical contribution was recorded for each combination of patent and definition. I recorded these statistics, but unfortunately I jotted them on the green results sheet that had to be handed in at the end of the workshop, so I don't have them now. The full results of the whole series of workshops are to be published on the 8th of April, including all the sample patents and definitions. The organisers also asked a spokesperson for each table (who, me?) to give an opinion on the alternative definitions.

The first definition was the one currently proposed. All the groups agreed that it reduced to the question of whether the invention was in a field of technology and had technical features, and all felt that this was of little or no value in explaining the term 'technical contribution'. This was felt to be the most permissive, but we had been very uncertain about some of our decisions, and in two cases the table had split (no, not literally), with two of us expressing a tentative 'yes' and two a tentative 'no'. This definition also produced some approximately 15/15 splits across the workshop as a whole.

The next two definitions were very similar, almost word-by-word anagrams. Again they specified that a technical contribution must be in a field of technology, but added that controlling the forces of nature in a new way was technical, but data processing was not. If I understood them correctly, one said that only using of forces of nature could be technical, but the other left open the possibility that other things might be. Thus it would not explicitly allow or disallow business methods, if they count as a field of technology. For all the examples we had, however, our group arrived at the same conclusions in both cases. These definitions also tended to produce more consistent results across the whole workshop, although I don't remember the details.

All the groups found these definitions easier to work with, albeit some found the combination of positive and negative conditions to be awkward and possibly inconsistent. There was disagreement over whether the results were desirable, however. Some participants, probably the majority, were in favour, others were not. After a little prompting from one of the UKPO staff, one speaker described one of these definitions as 'unduly restrictive'. We were also asked whether we felt some of the words were ambiguous -- my group at least had seen there were multiple constructions, but were confident we understood the intent.

The last definition talked about physical processes, but in a more general way. As far as I can recall, it produced similar results to the previous two, but my group at least had much less confidence in our judgements. This was reflected in the general consensus of the workshop, which held that it was as bad, or almost as bad, as the first definition. I believe the overall results from applying this formula were intermediate between the original wording and the other two, but can't remember how consistent they were.

As the attendees prepared to depart, the organisers called for a show of hands over whether we thought each of the inventions ought to be patentable. I don't think there were any declared abstentions, but the results didn't always add up to thirty due to sampling errors and people leaving the room. In all cases the majority, including me, voted against, but the margins differed considerably. In one case there was a more or less equal split of 12 for and 13 against: this was the claim we had misread, and on reflection I should have voted the other way. The case of the design process/manufacturing process was equally contentious, with 11 votes in favour and 12 against. So the claims I found trickiest in themselves were also the most controversial.

Two very 'pure' software patents received votes of 6/19 and 2/26, and it would be interesting to hear why some people decided differently in these two cases - perhaps because the more popular one made reference to a the invention's ultimate effect in allowing certain hardware to be used more efficiently. The remaining patent used new software to carry out a new commercial activity, and this was rejected even more strongly, with only one vote in favour. (The solicitor I'd chatted with earlier was sticking firmly to his principles.)

(This may reflect an anomalous feature of current practice: my rather simplistic reading of the decisions of the EPO and various courts is that one can patent the use of a computer to perform an existing business process (running software on a computer has technical character), but not to perform a new business process (because the invention as a whole does not have technical character). For the benefit of skip-readers, I repeat that I am not a lawyer, and the casually held view I have just stated is not advice to anyone.)

Finally we were asked whether we were in favour of 'software patents'. Several people had left by this stage, but 9 remained in favour, 16 against, and at least one thought the question was too ambiguous to answer. By this point the participants in the afternoon's workshop were quietly dispersing with scarcely a backward glance, while those for the evening had begun to gather in the antechamber. On the way home, I pondered the results of the session. (Actually I didn't ponder that much: mainly I nursed a severe headache, and later read about dynastic turmoil in Spain circa 1500. But I need some link to my conclusions.)

*Each of our four definitions of 'technical contribution' attempted to do the entire job of defining an invention, specifying in one way or another that it must be technological, new, and inventive (which is the equivalent of 'non-obvious' in the USA, I believe). To me it seems clearer and easier to define 'invention', to save traversing the links from 'invention' to 'technical contribution' to 'technical effect' and onwards.

*Three formulae attempted to exclude software and algorithms, and in practice I think would usually forbid patents relating to file formats, communication protocols, and user interfaces too. But the proposal leaves the issue of software undefined, so on this point it modestly defers to existing precedent without attempting to clarify it.

*Although the sample applications were written comparatively clearly, at least one group misread at least one of the claims. In my case it caused two of the 'forces of nature' -- style definitions to reject a patent they should have accepted.

*In general, some proportion of the inconsistent results will be due to misunderstanding the claims, and not to different interpretations of the terms. Therefore the definitions will tend to look worse than they really are.

*Lord Sainsbury is described as being the prime mover behind this exercise, and more generally the endeavours of the British government and the UKPO on this issue. This is consistent with the UKPO's web site. I imagine he will direct the UK position in the Council of Ministers, and any governmental representations to the Commission and European Parliament.

*The workshop itself did not mention the current position of the directive, the parliamentary amendments, or the procedural stunts that it has performed in its eventful journey. There was some discussion beforehand on whether technically the second reading has begun yet.

*There has been no mention of coordinated exercises in other EU countries -- Lord Sainsbury and the UKPO are doing this alone.

*Those people who I knew were patent lawyers were consistently in favour of software patents, those I knew were software developers were consistently against. Some people had different opinions on the different sample applications - I don't know their backgrounds. The patent examiners preferred to maintain the status quo, regardless of whether it allowed any or all software patents, but they also wanted clarity.

*There was a clear majority in favour of pronouncing 'patent' with a short 'a'.

¹ I was thinking of Cantor Fitzgerald International v Tradition UK Ltd. , discussed here by Humphreys, (a law firm based coincidentally in Bristol), and see point 4.