Thank you, I think you're probably right overall. Do you think I should edit the title now? Ironically, the parts you (reasonably) most object to are really a misleading result of trying to edit in response to comments: in fact though some documents are highly miscellaneous, most are flat A4. Also, it does seem the problem I actually have is knowing about features (since that's what prevents me starting), and not about "how to quickly scan a bunch of old documents": most people do know how, in outline (get a scanner), so I suspect answers to that would be unhelpful because too generic.

I think in the FTP case, that's not opinion-based, but rather an insufficiently constrained question (though of course the questioner might not realise that). If different people make different guesses re the question intent, then aside from that guess itself, their answers are likely still objective.

Understanding the design of tools often helps use them. I think it's a mistake to require every question to flag up their use. Instead, we should proceed as if they were asked in the spirit of good SE questions. If they were, all is good. If not, the questioner might not get what they wanted, but others likely will, because the answers will still be good answers to the generous interpretation of the question. It's only when there are multiple good (i.e. generous) interpretations that there is a problem.

I think as answerers / moderators we should at least consider the possibility that the questioner has reasons for asking the question that can't reasonably be communicated in the SE format. Not every question we don't understand is caused by a failing of the questioner's understanding. I can for example recall times when I knew the answer to a question asked by somebody smarter than me who was pursuing a problem that I do not understand. Sometimes I even thought the question was misguided, when in fact it was not. These issues cannot always be resolved, so we shouldn't try too hard.