Current focus: EndSoftwarePatents.org: wiki, mailing list, news, donate

Portrait photo of Ciaran

Software Patents

Note: for a look at my current work, see this transcript of a presentation about End Software Patents, March 2009, and swpat.org - the software patents wiki.

Table of Headings


back to table of headings

Preface

A person writing an academic paper has asked me for information about the EU software patents Directive, so I've decided to publish the information that I send him here. So this is a work in progress, and it may have rough edges but hopefully I'll find the time to turn this into a coherent post-Directive summary.

Other pages you may find useful is a transcript of a talk I gave: "Preventing Software Patents: How & Why", and a transcript of Richard Stallman on "The Dangers of Software Patents". And more information can be found on FSFE's software patents web page or my blog entries on the subject (tagged "swpat").

"[ref-3]" references: at the end of this document is a numbered list of referenced documents. In this text, where you see a bracketed note like this: "[ref-3]", that means the source of what I'm quoting can be found in item number three in the list of references. Clicking on "[ref-3]" will take you to the specific item in the references section.


back to table of headings

First installment: General intro

My experience

I started working on the CII Directive in early 2003, in Ireland. In January 2004, myself and others set up Irish Free Software Organisation. I came to Brussels in August 2004, and continued my working on the directive in my spare time. In April 2005, Free Software Foundation Europe took me on full-time.

Basics

In case you haven't found out already, I'd better warn you that this is a meaty issue.

There's a choice between short and precise and this document goes for the latter. I'll try to be concise, but more importantly I'll be trying to avoid leaving anything unclear, up to interpretation, or requiring trust/faith.

I suppose there's also a warning there: beware of summaries that you see. Many summaries were written for politicians -- since convincing them was the goal, not making them understand every detail -- but they may not be a good basis for fully understanding the issue.

Where software patents stand in the EU today

They're still being granted by the European Patent Office, but national courts are mostly throwing them out whenever litigation actually gets to that stage. (I say "mostly" to be on the safe side, but as far as I know, it's actually "always" - but that's not something I've researched.)

For this reason, owners of software patent are generally afraid to take litigation to that stage, so for the most part, European software developers are safe from litigation.

Why the CII directive was first drafted

For harmonisation. This is the justification of most EU Directives.

But the battle/debate/issue/problem/whatever was nothing to do with harmonisation. Everyone's for harmonisation, but the way the Directive was written, it would have made software functionality patentable.

Another justification, was "to comply with our TRIPS commitments". TRIPS is a global treaty which discussed patents, among other things, however, there is no requirement in TRIPS for software innovations to be patentable. In my view, stating that TRIPS requires software patents is a bluff: how many politicians have the time and inclination to look for, read, and understand a copy of TRIPS?

Those who claimed that TRIPS requires software patents based their claim on article 27 of that agreement, but when you actually look at that article [ref-1], it says:

"...patents shall be available for any inventions, whether products or processes, in all fields of technology, provided that they are new, involve an inventive step and are capable of industrial application."

The key part there is "in all fields of technology". Is writing software, in legal terms, a field of technology? The answer is that it depends on the definition of "fields of technology", and there is no definition. It's also worth looking at another part of TRIPS: Article 10(1) [ref-2]:

"Computer programs, whether in source or object code, shall be protected as literary works under the Berne Convention (1971)"

This is interesting because nobody claims that TRIPS requires patents for ways of writing books. Stories can be inventive and technical.

There is also no definition for "capable of industrial application", so rather than there being a mandate to bring in software patents, there is only a mandate of TRIPS signatories that they themselves define "field of technology", etc. ...and then allow patents for innovations in fields of technology.

The pro-software-patent camp wanted to define it as "applied exact science" (which would include math) and the anti-software-patent camp wanted to define it as "applied natural science" (which would include computer controlled braking systems, but would exclude math and computer programs).

Another term used by the anti-software-patent camp was "forces of nature" - if an innovation was a new teaching in ways of using forces of nature, it would be patentable, if not, it wouldn't. In the end "applied natural science" was decided to be a better way of saying practically the same thing.

The pro-software-patent camp say "oh look, software is technical, so it must be patentable".

This made sense at first to politicians, since they were finding the issue to be complex, and one word that English speakers use to express "complex" is "technical".

On the term "CII" ("Computer-implemented inventions")

For a clear discussion, it's best to avoid the term "CII". That term is one of the terms used in the name of the Directive, but like "harmonisation", it's not what the big policy debate was about.

"CII" is an unclear and ambiguous term. What relationship to the computer does "implemented" imply? Is the invention implemented on the computer? in which case it is software? Or is the invention implemented by the computer? such as a computer-controlled braking system for vehicles?

(Nobody was arguing against the patenting of computer-controlled braking systems.)

Even the BSA couldn't decide on what definition they wanted to use for "CII". While pushing for "CIIs", and while repeating the line "this Directive won't introduce software patents", just before the vote that rejected the Directive, the BSA published a report which started off "....computer-implemented inventions (usually referred to as 'software patents' in United States)..." [ref-3]

That was pretty stupid of them, but such stupidity is probably unavoidable when making a self-contradictory argument.

...and anyway, SAP had already blown their cover by buying full page ads in prominent European newspapers, asking for "CIIs" despite the fact that they make nothing but software.

The term "software patents" was the primary term used because that is what the debate was about: software.

An even clearer term would have been "software idea patents" or "software functionality patents", but they're a mouthful. "logic patents" was anther term, but few people used it.

In the closing stages, the terms "CIIs" and "software patents" became flags of either side of the debate. This is most likely because the Parliament took a very clear "anti-software-patent" position early on, so the pro-software-patent side needed to rebrand.

That's assuming good faith all-round. Another possible reason for switching to the confusing "CII" terminology is that clear-thinking politicians were not of benefit to those in favour of software patents.

...but the reasons can be put aside. "CII" is not a clear term for anything, so for clarity and understandability, it has to be avoided.

(Normally there would also be the option of noting what definition you are using, but in this case, whatever definition you use will contradict the definition used by one part of the pro-software-patent camp, so confusion can't be avoided.)

Why it failed

It was rejected in July of this year, in the European Parliament's second vote on it, the "second reading". The text handed to the EP was pro-software-patent. 21 amendments [ref-11] were proposed on behalf of the anti-software-patent campaign. For any amendment to be adopted, it need a majority of the EP (not a majority of those present, they needed a flat majority, including the MEPs who didn't turn up, so in effect each amendment needed support from ~65% of those present).

The pro-software-patent camp, despite having made progress, saw that there was too big a chance that all or enough of our amendments would indeed get this 65%. So they supported rejection.

We, the anti-software-patent camp, saw that the EP is not that powerful, and even if we got all our amendments adopted (like we did in the "first reading"), they could be discarded by the Council (as happened after the first reading). So rejection was the only sure power that the EP had. So we too went for rejection.

Where does it leave the EU software market

Mostly unchanged (see above).

What prompted the EU commission's first draft?

"Harmonisation".

...but it's worth noting that when the European Commission published the first draft - which is supposed to have been drafted by the Commission and only the Commission - they published it as a Microsoft Word document, and when you right click on it, and go to properties, and look at the "Author" tab/field, the file said the author was "Francisco Mingorance". [ref-4]

Mingorance works for the Business Software Alliance.

So from the start that there were strings being pulled.

My core objections to the directive

It would make software ideas patentable ("software patents").

What are my objections to software patents?

I think there are two core issues, from which all the other harm flows.

  1. They make software development legally risky and costly.
  2. They specifically inhibit the development of useful software, by blocking the use of standards.

Issue #1 happens because if I sit in a cave with a laptop and write some software, I can walk out and be handed a letter saying I'm a patent infringer. It doesn't matter if you have never read the patent. It doesn't matter if you didn't know the patent existed. It doesn't matter if you thought of the idea yourself (independently). The deal is: if you write some software, and you use an idea that is contained in any of the 50,000 existing software patents (each of which contains numerous claimed ideas), then you are an infringer.

This means you can be sued for damages, and you can be told to cease distributing your software - and whatever work you've put in was all a big waste.

...and how do I know if the patent holder is indeed correct in claiming I infringe that patent? How do I know that that patent is even valid? I can get a legal opinion and I can go to court. A legal opinion can cost thousands, and it doesn't get me anywhere solid. A court case can cost hundreds of thousands. So all options, other than submitting, are financially prohibitive.

Avoiding an infringement is practically impossible since there are 50,000 software patents today to check, and software ideas (being math) can be described in very abstract terms and are very difficult to recognise when written in patentese.

Avoiding an infringement is often also undesireable, for reason #2:

Issue #2 happens because in computing, due to "the network effect", standards are crucial (both public standards and de-facto standards). If I write a word processor, and it doesn't read and write the market leader's file format, and it doesn't look and behave like the market leader's word processor: it's not a word processor.

If the market leader patents an algorithm that is required to read or write their file format - I'm screwed.

And on the other side, if I patent an algorithm that is required to read the file format of my new word processor - it's worth nothing. Who wants to ready my innovative but unused file format? No one.

For these reasons, development of useful software becomes something that can only be done by very large companies, with cash, lawyers, and a patent portfolio for counter-suing or for cross-licensing.

The core of this issue for me, is that I think society as a whole should have the right to write software. Remember that making a website is "writing software", and that free software such as the GNU/Linux operating system exists today because every part of society, including the richest, but not just the richest, could write software.

There is also an element of indirect descrimination against free software ("free" as in free to use, study, modify, and redistributed), because free software relys on many development models, many of which do not entail large amounts of cash, lawyers, and patent portfolios. Free software relys on far more efficient models which rely on individuals, students, volunteers, SMES, companies for whom software is not their core business, all the way up to the very large companies.

Will the lack of commonality have a negative impact on the EU sofware market?

Maybe a little, maybe not at all. To date, there has never been commonality of patent law in the EU. This Directive was the first attempt at it, and nobody on either side was particularly interested in commonality.

But maybe the question was: Will the rejection of the Directive have a negative impact on the EU sofware market?

No. There will be almost no effect, if there is an effect it will be a slight positive one because the Directive, in it's original form would have introduced software patents, and by rejecting it, the Parliament has become yet another body (along with the national courts) who do not agree with the European Patent Office.

The effect will be positive because this gives software developers a little bit more assurance that they don't have to worry about litigation from patent owners - but the impact of the effect will be slight or zero because the EP, like the national courts, does not have power on it's own to discipline or fix the behaviour of the European Patent Office. So the EP's action has a symbolic effect only.

(Note that the European Patent Office is not answerable to anyone. It's not an EU institution, it's a completely independent body formed by an intergovernmental treaty: "The European Patent Convention")

If the Directive was adopted in the form that the anti-software-patent camp had pushed for, there would have been a positive effect on the EU software market.

If the Directive was adopted in the form that the pro-software-patent camp had pushed for, there would have been a severe negative effect on the EU software market.

Will the lack of commonality mean that the EPO can continue issuing patents for software (contra to Munich)?

The EPO will indeed continue to grant software patents, but it has nothing to do with any lack of commonality. The lack of commonality is at the national level.

The EPO will also claim that they are in line with the European Patent Convention (and of course we will disagree and call their interpretation ludicrous, and thankfully the national courts continue to agree with us).

Should there be a new directive?

There should indeed be a new Directive. The European Patent Convention is very clear, but astoundingly the EPO is granting software patents, and the only way to stop them is to be extremely explicit and maybe stamp it on the back of their hands. A Directive shouldn't be necessary, but it is.

Do software patents promote or hinder innovation within the software market

In software, they clearly hinder innovation. A prerequisite for writing innovative software is the right to write software.

The patent system takes the right to product development away everyone who doesn't have large amounts of money and lawyers.

This is not a big deal in many fields, since, for example, if your product is pharmaceuticals or mechanical engineering, it can be taken for granted that all potential developers have large amounts of money and lawyers because the requirement of having money and lawyers is already introduced by the need for manufacturing facilities, staff, safety tests, certification, etc.

However when the product is software, the vast majority of people who are capable of software development do not have the necessary degree of money and lawyers - so, unlike other industries, in software, patents prevent a huge block of bodies capable of innovating from actually innovating. For software development, the only thing you can take for granted is that potential developers have a computer and an Internet connection.

Also, patents encourage "lateral innovation". By blocking off one path, and forcing product developers to find a new path, they encourage the development of numerous solutions to the same problem - with the hope of finding ever better solutions.

...but for software, if I develop a word processor document format that is "better" than the market leader's format, my format is useless because compatibility, not technical merit, is the driving force of word processor technology.

On SMEs

I'm no SME representative, so my word mightn't stand for much on this, so I'll point you to a (or "the") real SME representative: UEAPME.

UEAPME is a union of european SME unions. It's members represent 11 million (yes) SMEs, with a total of 50 million employees. There's no bigger SME union, and there's nothing close. For Ireland, the Small Firms Association (SFA) and the Irish SMEs Association (ISME) are members.

UEAPME were one of the strongest anti-software-patent lobby groups. Here are two quotes from UEAPME press releases:

A press release [ref-5] said: "This directive will threaten the existence of many small businesses"

Another [ref-6] said:
"UEAPME is opposed to the introduction of an EU software patent, which would reinforce monopolisation in the software sector, damage interoperability and act as a barrier to innovation by SMEs. Small firms simply do not have the resources to engage in the costly and time-consuming process of patent application. This would enable dominant large firms in the sector to secure vast numbers of patents and result in crippling litigation costs, which would put small firms out of business."

Be wary of comments from small SME unions, from individual SMES, and from unions that claim that most of their members are SMEs, and on closer inspection those unions include some very large companies. SMEs are vunerable and needy (by definition, due to size). This makes them prone to coercion and enticement.

SFA and ISME, as far as I know, didn't make any individual statements on this. I don't know why.

SMEs and small unions are free from these weaknesses when they band together, so statements via UEAPME should be quite representative, while statements from small groups might actually represent a larger company whom they are dependent on or whom they are trying to curry favour with.

I add that warning because a handful of SMEs were shipped into the parliament before the vote and they lobbied for swpats (they probably said "CIIs"). This fooled some MEPs into thinking that there were SMEs on both side of the debate, but it was a sock puppet show, not a real lobby.

In response, UEAPME made a press release [ref-5] saying the following:
European small businesses have criticised a cynical campaign being run to misinform policy makers on the impact of the proposed software patent, currently being discussed in the European Parliament. The vast majority of small software firms in the EU are opposed to the proposal to make software patentable, according to an internal survey by UEAPME, the European association representing 11 million small and medium businesses.

However, a campaign claiming SMEs support the directive is being carried out with the funding of large software companies. [...]

Regarding the patent system in the US

It's an extremely valuable example. For software, they've sent a clear signal not to follow them.

Again, don't take my word for it, read the "software and internet patents" section of the US Federal Trade Commission's 2003 "Report on Innovation" [ref-7].

It's a ~350 page report on the entire US patent system. The software section is a very readable 13 pages. The conclusion of that section is pretty concise in listing 11 ways that software and internet patents are hurting US software users and industry, while giving zero redeeming qualities.

The US Federal Trade Commission doesn't represent any specific section of society. It's a government body (something like the Competition Authority here), it focuses on making a functional market for the US economy. Pages 153 to 165 of the pdf deal with software patents:

Even if you only have 2 minutes, read the last paragraph on page 165 of that PDF file.

For more info...

For more info, a guy well worth reading is Richard Stallman. He's been campaigning against software patents since the mid-80s in the USA and he also did tours of EU states helping the lobbying here. Here are two of his essays:

Free Software Foundation Europe's pages:

IFSO's pages:

FFII's anti-software-patent website:

(This one is the most complete, but not too organised.)


back to table of headings

Second installment: Nice quote from the Economist

"Bruce Chizen, Adobe's boss, sued Macromedia. There was no particular reason for the lawsuit, he recalls, but he had been irked for some time that Macromedia...had appeared to be embarassing the much bigger Adobe...Mr Chizen's people found an obscure patent that Macromedia was probably infringing upon and took it to court. The idea was 'to slow them down a bit' smirks Mr. Chizen."
The Economist, December 10th-16th 2005, p71 "Sue, kiss, marry"


back to table of headings

Third installment: Current state of play

Can a computer program be patented within the EU?

No. This is something everyone agrees on, and is something nobody is lobbying for/against.

Computer programs are covered by copyright. Patents cover ideas (detached from any implementation).

So the question is: If I think of a new, useful, unobvious way of displaying data on a computer screen, is that patentable?

(So the question assumes that the other three criteria (new, useful, unobvious) are fulfilled, and just leaves the forth in question: does the innovation occur in a patentable field? is the "subject matter" patentable?)

(Whether or not I write a program that actually does this is not relevant.)

What is patentable is laid out in Article 52 (1,2.3) of the EPC [ref-10].

The European Patent Office reads this as saying that

"programs for computers ... shall [be] exclude[d] ... as such"

(where "as such" is an apposition that could be replaced by "programs for computers")

And they deduce that "A method, using a cathode ray tube or LCD, conected to a processor, using copper wire or fibre optic connections, to cause a processor to represent data on a storage device, in such a way..." is not a "program for a computer" (since there is also a computer, wires, screens, etc. involved) and is therefore patentable. So the patent gets granted - which means nothing more than the EPO - which has zero enforcement abilities - says "we approve and record this application".

The anti-software-patent campaign, and the national courts, look at that same application and say that the innovation (the difference between the state of the art and the application) lies in the a new teaching of how to write a program for a computer, and since Art 52.2 says that programs for computers are excluded from patentable subject matter, this patent is not valid.

The interpretations are interesting, but I think it's more important that the policy be looked at. For example, if certain races were banned from the front of buses, would what's written in the law books be the most important thing? or would examining the policy?

I do think the law agrees with my position, but I'm also saying that what the law says is not what we should be looking at. If bus laws were being revised, then regardless of what they currently say, the new law should not discriminate anyone based on race.

So the interesting issue, IMO, is whether new patent legislation will/should clarify beyond all interpretation that software is not patentable subject matter, or will/should it clarify that software is patentable subject matter.

How can CII be seperated from from computer programs?

(I don't know how/if that'd be done.)

The debate was about whether innovations in ways of writing computer-executable sequences of instructions should be patentable.

Nobody wrote that on a banner, it's just not accessible, but that's what it's about.

Did the EU Directive fail?

It was rejected. The legislative process has a fixed set of outcomes, and I don't think any are called failure.


back to table of headings

Fourth installment: The other side of the debate

By the way, for the other side of the debate, here's a talk by a Director of the EPO:
Audio: http://mp3.ipag.info/forfas_ip_lecture_050610.mp3
Transcript: http://www.forfas.ie/news/intellectual_property/Lecture6_Transcript_june05.pdf
(found on: http://www.forfas.ie/news/intellectual_property/index.html )

It's actually quite interesting for the first 75%, but it is hard to believe that what he says in the conclusions are a good-faith attempt at accuracy about what the anti-software-patent campaign were working for.

He says that free software can co-exist with software patents, and his proof is that Linus Torvalds (the guy the wrote the kernel of GNU/Linux) lives in Europe and that most free software development happens in Europe. However, he doesn't mention that software patents are not enforceable in Europe, that they're just pieces of paper that the EPO hand out. I could quote his same proof for the counter-point: free software is so harmed by software patents that it doesn't happen in the US any more, it happens in Europe.

He also says that we wanted to abolish all patents for anything that can be assisted by a computer (so only pharmaceuticals and mechanical things like new wheels would be patentable). This is part of the "CII" smoke and mirrors.

It's a smart, but misleading, slogan for them to say "They're against CIIs" (here, "CII" would be inventions implemented by a computer, i.e. assisted by a computer, i.e. the invention would be in the subject matter of automotive braking, and a computer would be involved).

We were never against patents on computer-assisted inventions. We were against patents on ways of writing software.

He also says things like:

"We have stuff introduced with regard to, software can only be implemented if it has an effect on the forces of nature. Now I don't know what that means. But I do know that it will take more than a day in court to settle."

Which is a reference to one of our earlier proposals for a wording. "Forces of nature" is a wording taken from the German courts, so it has precedent and has proven there to be an effective delimiter of subject matter. As a Director of the EPO, he should know what that means (I find it difficult to believe that he doesn't).

I should back up what I say about our amendments, so you should look at the exact text of the 21 amendments the anti-software-patent camp got submitted for vote: [ref-11].

It's a readable and informative 17 pages. Amendments which I think are paricularly important are numbers 4, 7, and 9.

(For reading the amendments: the left-hand blocks of text are what was being handed to the EP, and the right-hand blocks were what they would be replaced by. "Articles" are the content of the legislation, "Recitals" are the preamble or transposers' guide.)

The proposal was for a "Directive". Directives are not law, but they are transposed into national law - so if the final Directive said "Technical means applied natural science, and non-technical innovations are not patentable", then the legislators in each member state would have to make any necessary modifications to their patent law, and patent office practice, to comply with what the Directive directs them to comply with.


back to table of headings

Fifth installment: Looking at our amendments

If the EPO is prepared to issue patents for computer programs then how can you say computer programs cannot be patented within EU?

Because the patents that the EPO are issuing are invalid (when tested in court).

If the EPO is prepared to issue patents for computer programs then how can they say computer programs cannot be patented within EU?

Interpretations of "as such", and how to read article 52 (1,2,3) of the EPC [ref-10]. This is detailed above.

Maybe the question is "Why do Europeans and Americans view software protection so differently?"

They don't. There's no line there. Europeans and Americans have the same or similar views on software patents (careful with the word "protection"). The difference is that software patents exist in the US, and they only exist in a near-meaningless form in Europe.

If you check the US Federal Trade Commission report (quoted & linked above), you'll see that the US sees software patents as purely harmful to software users and industry.

But undoing their mistake is not easy - how do you declare 100s of thousands of patents invalid after people have bought them and built the business models with the assumption that patents were something they had to deal with, or could use?

It's too difficult, it would be too upsetting (economically, not so much emotionally).

So while US bodies are wishing the US didn't have have swpats, the US diplomats are trying to tell everyone else that they must have swpats. (one way to stay ahead in a race after losing a foot is to require that your competitors take off one of their feet.)

The is no strong movement for changing US patent legislation to exclude software from patentability. It's seen as currently doomed. There are alternative projects, such as those started by the Public Patent Foundation, and the Electronic Frontier Foundation, and there are new "patent pools" which are mostly just PR stunts IMO - but everyone involved in these activities agrees that they're holding back the tide and not having software patents is the proper solution. Richard Stallman of the Free Software Foundation compares them to trying to stamp out malaria by swatting mosquitoes.

Those in the US who want to abolish swpats there see the clearest path as being to make sure other countries reject software patents, and eventually (20 years?) the US will fix local legislation for economic reasons.

The second reason for the US telling the rest of the countries that they should have swpats is that the US owns the software industry, and it owns the software patents. (72% of European software patents are owned by the US.) The US are quite good at convincing others to change their legislation. The "Free Trade Agreements" it makes with countries usually try to introduce software patents. ("We'll trade with everyone except you, if we can't protect our intellectual property" - those are their sort of words, for clear thinking I'd advise putting a field between yourself and the terms "protect" and "intellectual property").

Software patents stagnate development, so when the US is in the lead, stagnation isn't so bad for them.

...CIIs are not the same as software patents in the US...

The Business Software Alliance study says they are the same thing (as quoted and linked above). I disagree. I think they just couldn't maintain their own contradictory line.

"CII" covers two things: the first is computer-assisted inventions (new techniques for doing things in the physical word, all or part of which is controlled by a computer. One example is the ABS which is a way of using brakes, in response to data from sensors, controlled by a computer), and the second is ways of writing software.

The debate was only about the second of those things.

The term "CIIs" was used during the debate, but not for reasons of precision/correctness.

That much is fact. The speculation I'll add is that "CIIs" was used so as to broaden the debate - by them saying "CIIs", and us being against them, they could say "They're against CIIs". Which is not true. Their implication was "they're against CIIs, they're going to tear down the entire microelectronics industry".

What we said, in amendment #6 of our 21 amendments [ref-11] is:

"In order to be patentable, a computer-aided invention must make a technical contribution. The technical contribution must be new and not obvious to the person skilled in the art."

Our amendment #2 defines "computer-aided invention":

"(a) computer-aided invention means any invention in the sense of patent law the performance of which involves the use of programmable apparatus;"

Our amendment #3 defines "contribution" and "technical contribution":

"(b) "technical contribution" means a contribution to the state of the art in a field of technology. The contribution is the set of features by which the scope of the patent claim as a whole is considered to differ from the state of the art. The contribution must be technical, that is, comprise technical features and belong to a field of technology. [...]"

And our amendment #4 defines "field of technology":

"(ba) a "field of technology" is a field of applied natural science;"

So, piecing them together, word-for-word, replacing terms with definitions, that says:

"In order to be patentable, any invention, in the sense of patent law, the performance of which involves the use of programmable apparatus must make a contribution to the state of the art in a field of technology. The contribution is the set of features by which the scope of the patent claim as a whole is considered to differ from the state of the art. The contribution must be technical, that is, comprise technical features and belong to a field of applied natural science. The technical contribution must be new and not obvious to the person skilled in the art."

Now, lets check the ABS example against this.

There's a car, with brakes, computers, and sensors. To implement a new idea (ABS), nothing has to be changed other than the software. Is that a software patent? Is that excluded by our wording? No.

ABS involves the use of a programmable apparatus, so it has to fulfil criteria that we set out, and it does. The difference between the invention as a whole (car with ABS) and the state of the art (car) is ABS. ABS makes a contribution to the field of brakes usage, which is applied natural science.

A research paper on CIIs is perfectly valid, but half of it will be quite uninteresting since everyone was in agreement about computer-assisted inventions. It was the patenting of computer-executable instruction sequences that was debated, and is still being debated.

Software protection should promote innovation, yes?

The legal environment should promote ...something. "Progress"?

"innovation" can be useful, but only if it reaches the software users (in a useful way).

A bit broader: "The legal environment should benefit society"?

"protection" is a loaded term. Me, I want to be protected from litigation - I want software to be patent-free. Microsoft wants to be protected by litigation, it wants software patents to exist.

It's loaded because it generally gives the impression that that patents offer protection, when the other side of the coin is that they create threats and problems. When 50,000 patents exist, software developers are in danger, not "protected".

This issue has been going on for so long that the debate and the wording has become very precise, very analysed. Neutrality takes care.

Innovation within software is normally incremental

Yes. We build new on the old.

Other possible definitions

In the Amicus Brief submitted by End Software Patents for the "in re Bilski" court case, they suggested the wording "a process must involve significant physical activity" and "substantial physical manifestation", and excluding "an information processing algorithm with no physical manifestation of any sort", and they pointed out that "the Supreme Court repeatedly ruled that an information processing algorithm with 'insignificant postsolution activity' appended should still not be patentable".


back to table of headings

Sixth installment: "Protection"

If you spend years and millions on research, you want to be sure that you can produce that invention without someone else jumping in on your design and profitting from your research expenditure

For software, patents will damage such plans, not help them.

If I lock myself in my science lab basement, and come out ten years later with a cure for AIDS - I can then patent it, get investment for a manufacturing facility, and market it.

If I get frustrated with the spell-checker on my word processor, and I lock myself in my basement which has a comptuer, and come out ten years later with the perfect algorithm for checking spellings - what happens?

Do I patent it and buy www.spell-checker-algorithm.com and start sleeping on money? No, I've to write an implementation, and put it into a product. Lets say I put it in a word processor.

My new word processor starts with 0% market share. Rumors spread that it checks spelling like no other, but it can't open the word processor documents that people current use and that people have a lot of data already stored in, and people who try it say it doesn't feel "familiar". In a word, it's useless.

Software ideas don't exist in isolation, they have to be part of a product, and products must rival and be compatible with their competitors.

If the mad inventor does write a word processor that reads and writes the market leader's file format, and makes his word processor look and behave "familiar" to the word processor users, the market leader will come along and say "hey, we notice you use fonts."

"We've got twelve patents on rendering fonts. One patent on hierarchical menu structurs. Five on ways of clicking a mouse button (that's no joke). Several on our file format. And we probably have a load more but that's enough already so we won't even look."

The mad scientist says "yes, but your spell checking algorithm is inferior to mine, and I own mine." Big deal.

If they really want it, they'll cross license with him: "we won't sue you for your current product if you give us access to your patented algorithm".

The mad scientist's product can't be marketed unless he agrees to this (or unless he removes all their patented ideas, reverting to a useless piece of software).

So the scientist has not be "protected" at all. The market leader has access to his patents, just as the market leader gets access to everyone's patents - and the market leader was in control the whole time. If the market leader thought this guy would be a real threat, they'd simply tell him to keep his algorithm.

Most innovations in software happen during development, not during research, but what about the RSA encryption algorithm? It was developed during research. The researchers invested in it, spent time on it, and it does stand alone. Should it be patentable? No, because such a patent is too abusable.

What would happen if it was only patent in the word? The market leaders of various fields could license the patent, make it part of their file format, and competitors would then be prohibited from writing software that is compatible with the market leader. Competitors wouldn't have a hope. Such a patent is just too abusable.

Why do you imply that this Wikipedia quote does not apply to computer programs?

"A patent is a set of exclusive rights granted by a government to a person for a fixed period of time in exchange for the regulated, public disclosure of certain details of an invention."

What Wikipedia said is pretty accurate, but the exchange is a bad deal for software.

The assumption is that if there are no patents, people will keep their ideas as trade secrets, and if there are patents, the public will be restricted for 20 years, but at least then they get access to the idea. This doesn't fly for software: how many great software ideas from 1985 are of use today? (I can only think of two: public key encryption, and the RSA algorithm)

The disclosure is worthless for software. If the market leader could patent an algorithm needed to read their file format, rather than let their strangle-hold disappear after 20 years, they'd just extend their format to involve a new patented algorithm every 10 years.

Which shows that the value was never in the one idea, it was never in the specific file format, the value is that 90% of the market are using the file format created by the market leader's software, and the market leader is in control.

Groups of all sizes should be capable of legally protecting their investment (time, money, resources etc), no?

Having to work within the patent system means never being safe, never being independent, and never being able to safely invest time, money, or resources in software development.

There's a good audio and video recording of a recent talk by Stallman:
http://www.gnu.org/philosophy/audio/#CALG
on the topic of software patents. Some of it will be familiar from what I've written.


back to table of headings

Seventh installment: Corruption

During our lobbying, we saw a lot of dirty tactics and disingenuous statements. Starting now (August 2007), I will see what proof I can find.
Update July 2008: Unfortunately, I haven't found the time to give this topic the attention it deserves.

If it's any use, here's a story about the BSA having written the "Commission's" directive text: http://news.zdnet.co.uk/software/0,1000000121,2104807,00.htm
http://www.stacken.kth.se/~abo/tmp/proposal.png.

And here's a story ZDnet printed about the BSA publishing misleading figures: http://opinion.zdnet.co.uk/comment/0,1000002138,39205463,00.htm

Fake lobbying groups were set up during the campaign. These pretended to represent artists, creative people, and developers. One particularly dishonest group called itself "Campaign 4 Creativity" (C4C). This group misrepresented itself to MEPs. In truth, it was funded by Microsoft, SAP, and other pro-software-patent megacorporations. C4C won the 2005 award for the worst (most dishonest) lobby group. Corporate Europe Observatory also tracked their 2006 transformation into a general pro-patent lobby group. I remember that C4C had forums where people could post comments and ask questions, but when people asked difficult questions, C4C deleted those comments from the website. Here's a link to an archive.org copy of the C4C website from March 2005.

BoycottNovell.com has compile a list of Microsoft's sock-puppet lobby groups, with links to evidence and other such lists.

There are also possible conflicts of interests, such as those of MEPs Sharon Bowles and Klaus-Heiner Lehne, as described in SpinWatch's July 2008 report, Too Close for Comfort.

Another issue I remember is that an MEP has some pro-software-patent lobbyists registered as being his assistants. This gave them easier access to the parliament and facilities.


back to table of headings

Eighth installment: Two years later

It's now April 2008 and I'm getting to thinking seriously about software patent policy. I'll scribble my thoughts here for anyone interested, but don't expect them to be coherent. (Updated July 2008.)

Maybe patents should only exist for solutions to problems that people cannot solve on their own. For example, making games, deciding schedules, making jokes, writing books, and writing software and things that people can do for themselves. Those don't need patents. Society will do these things.

Manufacturing cars, TVs, and pharmacuticals are things that people can't do for themselves, so patents might be beneficial there. Or maybe the key issue is that people cannot mass produce these things. (I've heard good arguments against patents in some of those fields too, but I'll leave patent policy in those areas to experts in those areas. I'll stick to software.)

Will the software ideas being developed today still be developed in a software patentless future? Will innovation in general be increased or decreased, and will the quality or utility of innovation go up or down?

Some things that need to be highlighted more are that patents are a market distortion, the opposite of a free market, and every patent is bureaucracy.


back to table of headings

References

  1. TRIPS, Article 27, on what is patentable.
    http://www.wto.org/english/docs_e/legal_e/27-trips_04c_e.htm#5
  2. TRIPS, Article 10, on how software is classified.
    http://www.wto.org/english/docs_e/legal_e/27-trips_04_e.htm#1
  3. Business Software Alliance (BSA) study which says that "CII" is another term for software patents.
    http://wiki.ffii.org/Bsa050609En
  4. ZDNet news: Row erupts over software patent directive, see second paragraph
    http://news.zdnet.co.uk/software/developer/0,39020387,2104807,00.htm
  5. UEAPME Press Release 2005-04-27
    http://wiki.ffii.org/Ueapme050427En
  6. UEAPME Press Release 2005-06-21
    http://ueapme.com/docs/press_releases/pr_2005/050621_Computer_Patent.pdf
  7. US Federal Trade Commission: Report on innovation, October 2003
    http://www.ftc.gov/os/2003/10/innovationrpt.pdf
  8. Patent Reform is Not Enough, an essay by Richard Stallman
    http://www.gnu.org/philosophy/patent-reform-is-not-enough.html
  9. Saving Europe from Software Patents, an essay by Richard Stallman
    http://www.gnu.org/philosophy/savingeurope.html
  10. The European Patent Convention, Article 52
    http://www.european-patent-office.org/legal/epc/e/ar52.html#A52
  11. The 21 amendments, supported by those who oppose software patents (sometimes called "the 21 cross-party amendments" or "the 21 compromise amendments")
    http://swpat.ffii.org/papers/europarl0309/amends05/komprom0506.en.pdf

If any of these links does not work, please email me.


(Go: homepage of Ciarán O'Riordan)

Valid XHTML 1.1! Powered by GNU Made with Emacs FSF

© Copyright 2009 Ciarán O'Riordan. Verbatim copying and distribution of this entire article is permitted in any medium, provided this notice is preserved. Distribution of modified versions of all or part of this article are permitted, provided that such works carry three things: (1) this copyright notice, (2) prominent notices stating the that it has been changed, and (3) information for how to obtain the original (such as a URL).