This doesn't really speak much to software patents (I'm assuming those are of particular interest to HN). Does anyone know what the status of software patents in Canada is currently? If I were to judge solely based on the articles I read, the problems with patent trolls seems to be an exclusively American problem.
In June 1993, the Canadian Patent Office replaced its August 1, 1978 guidelines and published them for the "information and guidance" of practitioners. They were:
"1. Computer programs per se are not patentable.
2. Processes which are unapplied mathematical calculations, even if expressed in words rather than in mathematical symbols, are not patentable.
3. A process and/or computer program which merely produces information for mental interpretation by a human being is not patentable, nor does the process or program confer novelty upon the apparatus which uses it.
4. Claims drawn up in terms of means plus function which merely produces intellectual data are not patentable.
5. New and useful processes incorporating a programmed computer, are directed to patentable subject matter if the computer related matter has been integrated with another practical system that falls within an area which is traditionally patentable.
6. The presence of a programmed general purpose computer or a program for such computer does not lend patentability to, nor subtract patentability from, an apparatus or process."
The guidelines are actually pretty clear and the Canadian Patent Office mostly follows those. You will generally not be able to get a patent for a piece of software per se, although it is not very hard to attach it to a physical process to make it patentable subject matter. Any experienced patent prosecutor/attorney will be able to draft something that is acceptable to the patent office.
The Amazon 1-Click patent, for example, was held to be valid because the software and algorithms were attached to a physical system (of collecting and storing customer information along with an automated process of inventory gathering to shipping).
Most patent lawyers here in Canada are of the opinion that software _can_ be patentable, and that the trend is towards patentability. The main difference between US and Canada is not their respective Patent Acts (which are fairly similar), but the attitude of both the Patent Office and the courts in dealing with these things. It seems that the USPTO very easily accepts software patents, while Canadian patent examiners seem to be a little more strict on the requirements set out regarding software. Indeed, the Amazon 1-Click case arose in Canada because the patent office _rejected_ Amazon's patent, rather than the other way around as it seems more common in US.
Patent trolls are largely an american problem because America is one of the only jurisdictions that doesn't award legal costs to the successful party. Additionally, most of the 'deep pockets' are found in america, and since English is an international lingua-franca, its easier for foreign litigants to come and participate in litigation here, than it is to litigate elsewhere.
In general, Canadian patents treat software as any other invention; if it falls within the definition of an invention, and you can show utility, novelty and non-obviousness, you can acquire protection. The requirement for attachment to a physical process is trivial to circumvent for an experienced patent agent.
Obviousness is normally considered one of the big issues in software patents. Sanofi is the leading case on determining obviousness. Its got a succinct 4 point test that people like to gravitate towards centered on how to apply the 'obvious to try' standard, but the case also states that obvious to try isn't the only acceptable standard and that the determination should be industry specific.
Its against many large players' interests to bring litigation in this type of climate because of fear that courts will find that the software industry has a higher bar than most industries where satisfying obviousness is concerned.
N.B. one of the most common defenses to being charged with infringement is to show that the patent is invalid, so enforcement proceedings are risky.
I think you have to be careful to separate out aspects of patentibility or patent lifecycle:
There is,
1. Patent application classification.
2. Patent application testing.
3. Patent infringement.
4. Patent legal validity.
5. Patent technical validity.
You are discussing classification, dated 1997, which can easily be misrepresented (or at least represented incompletely), as it is here.
In most jurisidictions around the world, there are very high thresholds for the first, second and third.
Both the US and Canada are unusual in having much lower thresholds for the first and second, especially in the case of business process which can be easily redefined to fall under open classifications like "art" in Canada. This is how Amazon were able to obtain their 1-Click patent, which is really a software patent (that also easily fails technical obviousness tests in most parts of the world) wrapped up in a business process and classified under "art".
The US is also unusual in having a low threshold for the third, infringement, due to their court system that very effectively supports patent trolls (for any type of patent, not just software).
Regardless of the sustainability of patenting software, the problem that is becoming clear is that waiting until the very final stages of the lifecycle to defend a patent system has very high irrecoverable structural costs and an impact far wider than the restricted economic activity to which the patent relates.
This accurately describes the current situation, in broad terms. Amazon went through an extended litigation process to extend Canadian protection to their US 1-Click patent.
Reading: http://www.jurisdiction.com/spcanada.htm says that in 1993, the Canadian Patent office updated its terms regarding software patents to:
In June 1993, the Canadian Patent Office replaced its August 1, 1978 guidelines and published them for the "information and guidance" of practitioners. They were:
"1. Computer programs per se are not patentable.
2. Processes which are unapplied mathematical calculations, even if expressed in words rather than in mathematical symbols, are not patentable.
3. A process and/or computer program which merely produces information for mental interpretation by a human being is not patentable, nor does the process or program confer novelty upon the apparatus which uses it.
4. Claims drawn up in terms of means plus function which merely produces intellectual data are not patentable.
5. New and useful processes incorporating a programmed computer, are directed to patentable subject matter if the computer related matter has been integrated with another practical system that falls within an area which is traditionally patentable.
6. The presence of a programmed general purpose computer or a program for such computer does not lend patentability to, nor subtract patentability from, an apparatus or process."