Google has a great patent search tool for patents across a number of countries (patents.google.com). Google Patents parses some of the information, like the abstract, relevant metadata, various sections of the patent, and most importantly, the claims of the patent.
But the fact remains, it’s still tremendously difficult to understand the specifics behind the science or engineering of the specific patent. While I never expect to be able to understand it all, some sort of shared note-taking service or mapping tool for prior/related art would be helpful.
Crowdsourced notes and explainers, similar to how Medium does highlighting, would be useful for community members to understand difficult passages. The only problem would be vetting to make sure the scientific community keeps notes rigorous and high quality — not an easy problem to solve.
Prior art on Google Patents is currently a list generated from a simple keyword search. Every patent is tagged, so one could feasibly draw a network map of all patents.
Now apply this model to the patent office itself. With a helpful tool to define and cluster patents, the patent office might have an easier time evaluating new patents against prior art/existing claims. That said, I have no knowledge of how the USPTO currently functions — maybe they have an internal tool already.
Just a thought.