I'm trying to figure out why silverbullet regularly reports (using Stats: Show) substantively higher word counts than the output of wc -w $file. Today's journal page for example reports 1659 words while wc reports 1560. I've checked numerous different pages and there's always a higher number recorded by silverbullet than wc, although there is variance in the difference.
Reading the wc(1) man page states that it uses a "nonempty sequence of non-whitespace surrounded by whitespace" to determine a word, which is how I imagine silverbullet does it too? Or does silverbullet use another implementation?
I never put much thought into making this number super accurate, because… I didn’t care much if it’s a bit more or less (I don’t charge anybody by the word). If somebody’s interested in making this better I’d happily take a PR.
Oh cool, thank you for answering! I just assumed they'd be the same method of calculating words. Completely understand why an accurate count isn't a priority. Sadly contributing a PR is probably beyond my skill set at the moment.