Access and transport

How this class fits together

Access is the first thing a reader needs and the part most guides compress into a single sentence. The documents here separate it into three problems that are genuinely different from one another. Obtaining software you can trust is a supply chain question and is answered once, at install time, by taking the browser from the Tor Project and checking its signature. Understanding what an onion address is turns out to matter because the properties of the address, particularly that it is derived rather than chosen and cannot be shortened or corrected, are what make copying and never typing the correct handling rule. And reachability failures need classification before they need action, because the response to a local circuit problem and the response to a flood against the service are different, while they look identical from the browser.

The order matters. A reader who has NM-001 and NM-002 in place will handle almost every situation in NM-003 without needing to think about it, because they will already hold a verified address set and will not be tempted into the search that follows an outage. Read in the other direction, the documents still work but the reader spends the outage improvising, which is exactly the state this class exists to prevent.

What this class deliberately does not cover is whether the page reached at the end of it is genuine. That is a separate question with a separate answer, and it lives in the verification class, because treating access and authenticity as one topic is how people end up believing that a page which loaded successfully is a page worth trusting.