I run into this fairly recent (sent 2017) patent that seems to just describe an incremental parser which is used to 'enrich' textual program code with syntactic and semantic annotations. I don't really understand what the patent tries to protect and what is supposed to be the patent worthy invention here as for example tree-sitter existed in 2017 already. I'd be interested to hear any thoughts on this. https://patents.google.com/patent/US10684829B2/en

Nominally, everything after "what is claimed is." How protection works in practice is beyond me.

This is from intentional software right? Intentional software have been real innovators in this field since 2002, so it’s possible that their work precedes tree-sitter. Of course that doesn’t change the fact that software patents make no sense.

Yes, or at least from people who used to run Intentional Software. Intentional was acquired by Microsoft around the time this patent was filed. I'm not necessarily so much against software patents, but my understanding is that what you're trying to patent can't exist in the wild at the time of sending the patent application. So I'm puzzled how this got accepted. Maybe @Jason Morris can clear things up for us a bit?

No such luck. Studied IP in law school, but software patents never made any sense to me. I suspect a lot more patents get approved than are enforceable, but I have no real clue.

But even "enforcement" has gotten pretty fuzzy because I suspect many software patents aren't intended to ever see the light of a court room... Rather they're just a hedge against accidentally infringing someone else's patent, so you have something to horse trade with them if you need to.

Think of patents as one of the many legal mechanisms larger firms use to communicate with one another.