“Would “my” software infringe on the ‘002 patent? My answer? It probably does not infringe. But I probably should pay nevertheless.”
The purpose of Project Troglodyte is to hunt for bad patents and to show what went wrong. The project is now being continued at http://www.project-troglodyte.org/
I will now try to see guess whether “my” software would infringe the ‘002 patent. And I do mean “guess”; there is no way of finding the “truth”, as has been discussed earlier. The “truth” can only be discovered in court, when the patent owner sues someone.
To make one point clear: a patent cannot prevent me from making a Java program for myself to do this, at least in Europe (in principle it could in the US, but in practice this cannot be enforced). In practice it is difficult to prevent an open source program from being done (but only because it is difficult to know whom to sue). What it can do is prevent anyone from making a commercial application.
(Open source fanboys will say that in that case there is no problem. But the problem is that, overall, open-source programs are just not famed for their usability. Usability is all about the boring details, and money makes it much more motivating to use weeks and weeks polishing up those details. In my opinion (and I may be wrong), if this software cannot be made as commercial software, then its use is going to be severely restricted. I am therefore assuming that “I” am a small company which sells this software.)
What the ‘002 patent appears to cover is: using a combination of measurements and subjective data inputs, making a hypothesis based on those inputs, and then presenting various outcomes to the user. A more detailed analysis is in Appendix 1.
So: if I were to receive a letter from the patent owner, asking for a “reasonable” licensing fee or risk being sued, what should I do? This is not hypothetical: see the EFF’s Lodsys FAQ. Although this is limited to the Unites States, a foreign company can in principle be sued if they sell the software in the US (although I am not quite sure how that would play out in practice).
Would “my” software infringe on the ‘002 patent?
My answer? It probably does not infringe. But I probably should pay nevertheless.
Since IPR is “non-conventional warfare”, it makes no real difference what the “facts” are. It’s my perception that matters. If I were a small enough company, just hiring a patent lawyer to look at the letter might already strain my budget.
And this is not something that a patent lawyer can just glimpse at. There are so far 16 patents and more applications that need to be crawled through, about 1000 pages of text so far. Even the cheapest and friendliest patent attorneys will charge upwards of 150 EUR per hour. This is easily at least a week’s work, so I would expect to pay the attorney 6,000 EUR just to sit down and look at this.
(Note that this kind of insanity does not currently happen in Europe, where we have no software patents. We are lucky, for now. But if patent laws get harmonized, we will get there too).
It’s up to the patent holder to decide whether or not to sue me, so I could certainly try to call the bluff and refuse to pay. The holder cannot get any money from me unless he takes me to court. And there is an interesting paper by Allison et al which suggests that the holder might be more likely to lose than to win.
But there is a mismatch: the holder in this case is a huge company whose core “product” is litigation, and which has enough of a buffer that is can easily risk losing a court case every now and then. I cannot lose — most likely I cannot even afford the court case in the first place. So, if the license fee is priced just right, so that even talking to a lawyer would be more expensive, I would probably pay it.
Anything wrong with this? Legally, no.
How did the patent owner manage to pull something like this off? (Subject of next posting)
APPENDIX 1: CLAIMS, CLAIMS, CLAIMS
If this were a professional prior art analysis, I (or actually a patent attorney) would have to write down the steps into the form of claims, and then try to see whether the claims overlap. I simply do not have the time to do that, so I will cut some major corners here. What I am doing below is not legitimate patent analysis; real analysis would take perhaps ten times more effort (and require a qualified patent attorney).
Looking at the text and claims, it is clear that the biggest original focus of this idea has been on health monitoring of a person. (This is particularly clear when looking at the earliest patents in the family). This opens up some potential vulnerabilities for the patent owner, which I will discuss in Part 4.
“My” software has the following four blocks:
A) Collect and monitor objective pollution data at the location, and transmit that information to a server
B) Collect subjective estimates of the pollution level from people near the location, and transmit that information to a server.
C) Use the information to make an estimate of the pollution status
D) Inform the user about the pollution status.
Claim 1: A computationally-implemented system in the form of a machine, article of manufacture, or composition of matter, comprising: means for acquiring a first data indicating at least one reported event as originally reported by a user and a second data indicating at least a second reported event as originally reported by one or more sensing devices; and means for developing a hypothesis based, at least in part, on the first data and the second data.
Basically, this may be what my program does in steps A,B, and C. Of course it all depends on how the court chooses to interpret terms like “acquiring”, “reported”, “developing”, “hypothesis”. I follow a Twitter feed, and monitor sensor data, both of which I know to have their weaknesses and biases. Therefore I have to make a hypothesis about what the actual pollution situation is by combining the two.
Claims 2-10. Refer to possible ways in which the data is collected. This can be status reports, blogs, or other means of collecting data.
Claims 11-23. Refer to information that is collected about the user himself (such as blood oxygen, accelerometers, and so on). These would not be infringed.
Claim 24: Location of the user is determined by GPS or other method. This might not be a problem.
Claim 25-26: Air temperature and pollution are measured at the location of the user (determined from GPS of claim 24). There is no infringement if the sensors are somewhere else!
Claim 27. The computationally-implemented system of claim 1, wherein said means for acquiring a first data indicating at least one reported event as originally reported by a user and a second data indicating at least a second reported event as originally reported by one or more sensing devices comprises: means for acquiring the second data including environmental characteristic sensor data indicating at least one environmental characteristic.
Claim 28. …. comprises: means for acquiring a third data indicating a third reported event as originally reported by a third party and a fourth data indicating a fourth reported event as originally reported by another one or more sensing devices.
These are the crucial claims. If it is “true” that the detection (especially the third and fourth data) can be in another location, then it is quite possible my SW infringes. If, on the other hand, they still refer to to location of the user, I would probably be safe.
Claims 29-35: Hypotheses are made, and can be refined and compared to the facts. If we have gotten data from A and B, anything that we do in C could be argued to be about forming a hypothesis. “Is the air clean enough for me?”
Claim 36: A hypothesis is made that relates to the user.
Claim 37. An action is made (?).
Claim 38. An advisory relating to the hypothesis is made. This could be D.
Claims 39-43. The hypothesis, or facts are that related to hypothesis, are presented to the user. This is D.
Claim 44. A device is prompted to make an action.
Basically, as far as I can see, much boils down to what claims 27-28 “mean”. Claims 25 and 26 refer specifically to sensing the pollution at the location of the user, and do not pose a risk. However, if claims 27-28 mean that the system can also use pollution measured elsewhere, then I am at risk of infringing.
APPENDIX 2: FULL LIST OF PATENTS
The four patents with asterisks may (or may not) be relevant to pollution monitoring.
It is almost certain that there are additional patent applications in the queue, but I simply do not have time to look through them