When Combining Things That Already Exist Is Obvious

Almost every software invention is an arrangement of known components. That makes this the question most engineers actually have, even when they ask it in other words.

In 2007 the Supreme Court reaffirmed the Graham framework in KSR while telling the Federal Circuit it had been applying a supporting test too rigidly. The part that matters here is the warning that followed: combining familiar elements to yield predictable results is likely to be obvious.

That is why the first rationale on the examiner's list is combining known elements, and why obvious to try sits on the list as well, framed as choosing from a finite number of identified predictable solutions.

There is a limit built into the statute that pulls the other way. Section 103 requires the claimed invention as a whole to have been obvious. That phrase forecloses an analysis which establishes each element separately and stops there. The comparison the statute directs is to the differences between the claim and the prior art, not to whether any document resembles the invention.

What This Does Not Decide

• It does not say whether combining what you combined was obvious.
• KSR is a Supreme Court decision; MPEP 2141 is the Office recording it. Neither is a rule that a combination is unpatentable.
• It does not address eligibility. Whether a claim is directed to an abstract idea is a different question under a different section.

Educational, not legal advice. Whether any of this reaches a particular invention is a determination for a registered patent practitioner working from the actual prior art.