Myths and Truths

[Photo of a Wolpertinger]

Myth #1: Current EPO practice is a suitable basis for creating a directive.

Truth: The EPO granted over 30,000 patents on software and business methods which are as broad, trivial and just as damaging as their US counterparts. Fortunately, being illegally granted, they are not yet enforceable (see below).

Justification: The FFII has analysed a random sample from the patents granted by the EPO—see: http://swpat.ffii.org/patents/.

Myth #2: The Council version of the directive will not make anything patentable which is not already patentable.

Truth: Patents on software and business methods, although granted, are not enforceable right now, because they have been granted against Art 52 EPC which clearly says that software programs are not patentable. The Council version of the directive would change that.

Justification: The Council directive was written with the explicit goal to establish current EPO practice. A scientific study has confirmed that the Council directive will introduce US-style patents on software and business methods in Europe—see: http://www.ffii.org/~jmaebe/epecosci0502/SoftwarePatent.pdf

Myth #3: Patents indicate how innovative a company is. Software companies will benefit from this.

Truth: Typical software startups cannot afford patents, thus this “indicator” is attracting potential investors away from the most innovative companies.
In addition to the negative effects on the free market, investors would be more attracted by a less litigious market without the risk and uncertainty that the "mine field" of software patents causes.

Justification: A Patent is not a harmless indicator, but a granted monopoly. And all monopolies are harmful for the free market, as the EU Commission already knows.

Myth #4: Patents are necessary to protect the functionality of software against non-innovative competitors.

Truth: There are two well-established protection mechanisms for software: Copyright and trade secrets, achieved by the separation between the source code (human readable list of instructions) and the actual executable program (binary code that the computer can interpret). When the source code is kept secret and copyright is used to prevent undesired reuse, there is only one economically practical way to copy the functionality: write a new program from scratch.
This costs the competitor as much work as the original innovator and thus should not be impeded by a patent.
Patents are not used for protecting software because they are too expensive, and they need too long to be granted—longer than the typical life-cycle of a software product. In addition they don't even require the software to be written in order to “protect” it, so they create monopolies upon ideas, don't protect software. Is like I was permitted to grant a patent about “an revolutionary engine that uses hydrogen instead of gasoline” without even design one, wait some years until some cars manufacter do realize it and start selling it, and then go after them asking a lot of money because I've patented that idea. This is what happens daily where software patents are permitted.

Justification: The software industry has been one of the most innovative business sectors for over 20 years without patent intervention—which proves that the existing protection mechanisms are suitable.
The existence of cross-licensing agreements between the world's biggest holders of software patents prove that these patents are not used for the protection of products.

Myth #5: Patents help smaller companies to protect themselves against bigger competitors.

Truth: All big companies have big pools of broad software patents. The big company can always find some patents in this pool that the smaller company is infringing on with its own software.
Only when the small company is not in the software business, but concentrates on patents (“litigation company”), it can successfully sue not just a small but even a big software company.

Justification: Both types of lawsuits are already taking place on a regular basis in the USA (see http://wiki.ffii.org/SwpikxraniEn for some samples).

Myth #6: Software companies can profit from patent licensing.

Truth: Writing or buying software patents does not require writing software, so patent licensors for software do not need to be “software companies”. Software patent licenses are paid by software developers and users, so what is certain is that software companies pay patent licenses, not so that they pocket income from patent licenses.
In any case the point is never whether something is profitable for someone, but whether it is in the general interest. Patents are a drain of resources from software innovation to patent offices, attorneys and litigators, resulting in a less innovative and competitive software market. This is a net loss for society, even if you can find a few companies who profit from this general loss.

Justification: CEC consultation before the directive, and recent German survey show software companies are against software patents. No company is against profits.

Myth #7: Software patents are helpful to spread knowledge among software developers.

Truth: Patents are written in a language which is not easily understood by software developers. Given the problem, it is easier for a software developer to create an independent solution than to understand a patent which deals with the same problem. Also, patent descriptions either describe only one of the many implementations they cover or simply describe a functionality without details in how to achieve it.
Furthermore, software patents are typically granted for very basic elements of software development, where there is no need to spread knowledge.

Justification: Ask any software developer.

Myth #8: The Council Directive will prevent the EPO to grant trivial patents.

Truth: The EPO already has granted thousands of trivial patents. With the explicit goal of the Council directive to establish current EPO practice, it is clear that there will be no sudden jump up in quality of the patents granted by the EPO.

Justification: See, for instance: http://webshop.ffii.org, where it is shown how a simple commerce site infringes on at least 20 patents already granted by the EPO.
No precise and proven measures are suggested in the directive to improve novelty or obviousness examination, beyond reinstating the requirements already present in law that have not helped avoid trivial patents.

Myth #9: The directive as suggested by the EU Parliament in first reading will eliminate all patents in modern fields of technology.

Truth: The Parliament directive requires the application of forces of nature—for instance mechanical forces, radio waves, or chemical effects—to make an innovation be a patentable invention. It excludes nothing but pure logic—calculation rules and business methods.

Justification: The “computer-implemented inventions” given as examples against the Parliament directive are in fact computer-aided inventions that do use forces of nature. Of course they are patentable according to the Parliament directive.
No concrete patent numbers are ever given as examples that should be granted but would be prevented by the Parliament directive. The examples given are always about something that could be patented or not depending on whether the innovation was of a material or logical kind.

Myth #10: A Directive can be safely approved since subsequent observation of possible negative effects on SMEs for some years can ensure that it don't harm economy.

Truth: The effects of the new directive will take many years to become apparent, at which time it will be too late to correct this error.

Justification: It takes years just to get a patent granted. Furthermore, those who want to abuse the new patent system may simply wait until the observation period is over.
The legislation cannot surrender the responsibility of choosing the right approach now.

Myth #11: The conflict is between the industry and the programmers of Free Software (Open Source).

Truth: Software patents affects all kinds of software. Whether the source code is open or kept secret does not really matter. The source could just make it easier to detect some kind of infringement.
The conflict is between, on one side, some few big patent holders and "litigation companies" (most of them from the US or Japan) plus the patent industry (patent offices and attorneys), and on the other side the European software industry (SMEs)

Justification: Studies among European SMEs clearly show that they feel threatened by the big patent holders, independently of whether they are focusing on Free Software or not.
Please note that the Free Software movement is not a hobbyists' movement but a segment of the software industry in which European companies have the edge and which is important for the IT infrastructure of all businesses. (See, for instance, http://swpat.ffii.org/papers/bmwi-fhgmpi01/.)

Myth #12: European software patents help to protect the European software industry against competitors in the US, who already have software patents.

Truth: The validity of patents depends on the place of market, not on the origin of the company.
In fact the opposite is true: software patents in the spirit of the Council directive make the European software industry at the mercy of non-European companies.

Justification: Only one quarter of the patents on software and business methods granted by the EPO so far belong to European companies. 42% of the patent holders are located in the USA, 28% in Japan.

Myth #13: Infringement on software patents is done when someone tries to steal some innovative idea from someone else. Honest programmers can do a patent research.

Truth: Since the over 30,000 granted software patents cover the basic elements of programming, infringement on software patents is an unavoidable side-effect of honest software development. It is simply impossible to write any piece of software without—unbeknownst and unwillingly—infringing on hundreds of European software patents.
In any case, carefully examining 30,000 patents, written in vague and not technical language, is almost impossible (it's 821 a day, for a entire year), and you are never sure of the result.

Justification: A random sample of EPO software patents evaluated so far by the FFII (see http://swpat.ffii.org/patents/samples/) displays a collection of basic components for software. In fact there are not even "fields" to classify the patents; each of the 30,000 EPO software patents can affect each type of software. For a patent research, one has to check the software against all 30,000 patents.
The small sample evaluated so far makes it already impossible to implement certain types of software (for instance a web shop—see http://webshop.ffii.org). This will be even more true when taking the full sample.

Myth #14: A directive is needed to harmonise and to clarify the law.

Truth: Every EU directive does harmonise. However in this specific case there is no need to harmonise the law. The current differences are the result of a drift in EPO granting and interpretation of law in the member states, not of different laws which are already harmonised by the EPC.
But indeed the directive could help to clarify undefined terms such as "industry" and "technical". Without a definition of technical contribution this stated objective of harmonisation will fail, because this is where the differences lie.

Justification: Although Art 52 EPC clearly excludes them, the EPO has granted over 30,000 patents on software and business methods.


More about Software Patents – Contact – xhtml 1.0 – Last Update: 25 May 2005