How is patent infringement actually proved?
By taking each asserted claim, breaking it into its individual limitations, and locating every one of them in the accused product or process. This is the element-by-element requirement, and it is unforgiving: if one limitation is absent, that claim is not literally infringed, however closely the accused product otherwise resembles the invention.
The evidence that satisfies each limitation varies enormously with the technology — source code review for software claims, circuit analysis and die imaging for semiconductors, testing and materials characterisation for mechanical and chemical inventions. The structure never varies. What separates a strong technical case from a weak one is usually discipline rather than sophistication.
What does a weak claim chart look like?
One where the right-hand column paraphrases the claim language with the accused product’s name substituted in. It looks rigorous, it fills the page, and it establishes nothing, because it asserts the conclusion rather than supporting it. This is the single most common substantive weakness in infringement reports.
A strong chart cites, for each limitation, something the expert did not create: a specific page of a technical specification, a named function in the source code, a measured value with the protocol that produced it, a screenshot of a configuration screen. The test is whether an opposing expert could go to the cited source and check. If they cannot, the limitation has not been proved — it has been asserted in a table.
Does the accused company need to know about the patent?
For direct infringement, generally no. It is effectively a strict liability question, and a company can infringe a patent it has never heard of, sold by a supplier it never investigated. Knowledge is not an element.
Knowledge becomes central elsewhere, and this is where notice letters acquire their significance. Indirect infringement — inducing others to infringe, or contributing to their infringement — carries knowledge and intent requirements. Willfulness, which affects whether damages can be enhanced, also turns on what the company knew and how it responded. The internal email chain following a notice letter is therefore among the most heavily contested evidence in many patent cases.
What is the difference between proving an apparatus claim and a method claim?
An apparatus claim can generally be infringed by a product configured to meet the limitations. A method claim generally requires that the method actually be performed, which means proving somebody did it — not merely that the system could.
That difference redirects the entire evidentiary effort. For method claims the case frequently turns on how customers actually used the product, which lives in server logs, default configurations, user manuals, training materials and support tickets rather than in the technology itself. Cases are lost here by teams who proved the system capable and never established that anyone operated it in the claimed way.
Is the doctrine of equivalents a reliable fallback for a weak literal case?
No, and treating it as a safety net is usually a mistake. The doctrine of equivalents allows infringement to be found where differences from the claim are insubstantial, but it is applied limitation by limitation rather than to the invention as a whole, and it is constrained in two significant ways.
First, prosecution history estoppel: ground given up during prosecution to obtain the patent generally cannot be recaptured by arguing equivalence. Second, the doctrine cannot be used to cover territory the prior art already occupies. Used deliberately for one or two specific limitations with a clear argument, it is a genuine route. Used to rescue a literal case that failed on several limitations, it tends to fail alongside it.