I was going to mention earlier that if general purpose computers with attached peripherals end up being ruled as legit machines depending on the patent, then would it be acceptable to put the thing together on your own if you own the various components already using them for legal uses? Note that the software itself would not be patentable. [See example here http://boycottnovell.com/2008/10/30/bilski-decision/#comment-32129 ]

One argument is that it would be illegal, ie, the combination would be a patent violation, but then that would seem to mock the idea of a patent since a patent would be granting a monopoly to something whose components would be fairly well understood/unprotected “inventions”.

One ruling recently said you couldn’t put together two obvious items in combination and get something patentable, but what about 3 or 4? If 1 and 1 obvious leads to obvious, then 1 and 1 and 1 is obvious by induction. Ie, 1 and 1 was shown to be obvious, so now that same 1 and 1 together with another obvious 1 would also be obvious, at least it would once the first two were put together.

This argument might point to how courts might ultimately rule. A combination of 3 obvious items might be non-obvious if no one could find an intermediate obvious use for 2 of these or if such use would have the addition of the 3rd item be non-obvious.

The value here is that perhaps existing patents can be shown to be “obvious” if we can break down the invention components into a series of steps, all such steps shown to be obvious constructions from the 2 component parts. This would apply to all patents.

Patent laws are horrible. They kill growth and cleverness in people for the sake of giving the first person to put that combo together (and file for a patent) a monopoly for 17+ years. Imagine if every person coming up with a mathematical theorem patented it so that no one could leverage that theorem in their own future proofs for 17 years. That would kill mathematics and science and much more as we know it. But why not allow it for mathematical algorithms yet then allow “inventors” to gain that huge anit-social power grant when they make some aspect of the invention physical/machine? Why kill technological advancements? Frequently, it’s actually the abstract ideas/algorithms the ones that require the greatest craftiness. So we don’t allow patenting of the truly difficult for the sake of social advancement in math etc, yet we then allow the patenting of the frequently more obvious physical device inventions [though maybe this won't be allowed in the future if we can show a clear "proof" to the courts as indicated in the earlier paragraphs above].

Presumably a just reason for granting patent monopolies would be that further advancement along those lines would not be likely in the short term (17 years) and we would want to help subsidize the investments that led to such a discovery/invention. I think this would make some sense for some of the inventions that have been patented over the years. But then this fails horribly for sw patents for the most part (if not in all cases) because the sw industry has shown that basic inventions and inventions supported by past inventions happen frequently — lead to better products for end users frequently. Meanwhile, FOSS has shown these inventions have real value to users/society and to businesses (eg, Red Hat) at *nominal costs* to those contributing to the development.

A basis for granting any patents should be that it could not be used to restrict products if those products could be shown to be derivable and mass produced at a low investment by those (eg, inventors) taking part. For example, if they came up with the invention pretty much on their own (and can show clear progression of evolution of ideas) and at an affordable cost and could then get it into the hands of users also affordably.

I think future arguments before the courts (and before Congress, if laws need to be adjusted) should focus on the unaffordability condition and on the lack of a clear recipe understandable by many practitioners as two prerequisites for granting a patent and for validating a granted patent. Perhaps, based on the recipe and cost functions, if a threshold was met and the patent was granted/upheld, a suitable monopoly period and potential royalty conditions would be determined by the PTO/courts ..or perhaps these limits and allowance would only be determined by the courts (not the PTO), once a challenge was filed, in order to help relive the PTO of such burden for each patent granted.