I'm trying to decipher a competitor's patent (US only), which follows the usual system-method-CRM format for computerized patents (in this case, a standard computer with standard peripherals running innovative industrial control software). Although I've been granted some patents, this is my first foray into the wild world of claim construction. Juggling the different canons of construction is quite bewildering.
Claim 1 is a system / apparatus claim, and it's very functional, including two uses of the phrase "operative to" followed by several vague steps that correspond to clear structures and algorithms in the specification. There is certainly no way for a PHOSITA to practice Claim 1 without those algorithms. Moreover, there is at least one term in Claim 1 that is totally novel, with no dictionary meaning, no ordinary and customary meaning, and no meaning to a PHOSITA... and no clear intent to define it in the specification. So I understand from Indacon v. Facebook that this term will be strictly defined by the corresponding structure and any equivalents.
This all works out in our favor, as our system isn't really very close to theirs. It's the method claim I'm worried about, because it recites the same vague steps, but if those steps are unanchored to the specification, it would be absurdly broad. But is that really the case? It uses the same terms as Claim 1, so given the strong presumption that terms have the same meaning across all claims, doesn't that mean that those terms should be equally narrowly construed in the method claim as well? That would be great news for us.
FWIW, nothing in the prosecution history seems applicable to construction.
I would link the actual patent but I don't want the patentee to find this question on Google... at least, not yet.
Edits: Add US jurisdiction. Fix my mistake calling this a software patent (I've been looking at the software part for weeks); it does recite a bunch of standard hardware connected in a standard way. Note prosecution history.
Edit: The patent was granted within the last 5 years.
Edit: At the suggestion of Eric S, here's an obfuscated version of the claims.
1. A system for detecting bad chickens by listening to clucking noises, comprising a data-processing device connected to at least one microphone and operative to: (a) record the clucking noises from the at least one microphone; (b) detect clucking phrases within the clucking noises; (c) identify biometric fingerprints from the clucking phrases; (d) decode the hidden information of biometric fingerprints using at least in part the identified biometric fingerprints; and (e) locate a danger signal within the hidden information of biometric fingerprints. 6. A method for detecting bad chickens by listening to clucking noises, comprising: (a) recording the clucking noises at a data-processing device connected to at least one microphone; (b) detecting clucking phrases within the clucking noises; (c) identifying biometric fingerprints from the clucking phrases; (d) decoding the hidden information of biometric fingerprints based at least in part on the identified biometric fingerprints; and (e) locating a danger signal within the hidden information of biometric fingerprints.
Yes, the phrase "hidden information of biometric fingerprints" is just as awkward in the real patent (but not so nefarious, as the real patent is about industrial processes). That phrase is also the novel term that's defined nowhere in the spec except as the output of an algorithm for calculating it. So it would seem that in Claim 1, the term should be completely tethered to the disclosed algorithm. But would it be so tethered in Claim 6? I would think so, because the term is contextually meaningless except as defined by the disclosed algorithm. But if it's not tethered to that algorithm, then the term could be construed very broadly. It would also make the whole method very broad because steps (a)-(c) are industry standards and wouldn't be amenable to a design-around.
I do understand that the patentee wouldn't want to construe the term too broadly because that would certainly read on prior art and invalidate their patent entirely. But the invalidation defense requires a whole trial and we don't really have the funds for it. I'd rather design widely around step (d) so that the patentee can't even clear the pleading stage, or at least make it obvious early in discovery that there's no case.