javadoc-stylelisted

# Javadoc style House rules for javadoc in `src/main`. A comment that breaks one of them is either wrong, or a reason to change the rule here deliberately. Code, javadoc and message text are written in English. ## What a comment is for Javadoc says what the signature cannot: **the contract, the meaning of `null`, the invariant behind a value** and - rarely - **the reason for a decision**. Restating the name of a method is worse than writing nothing. The norm is two or three sentences; a second paragraph has to be earned. A reason earns its place only where the code without it reads as a **mistake or as unfinished work**. That is the whole test. "Publishing happens after the transaction commits, because until then the operation does not exist" passes it - the ordering looks arbitrary until you know. "A record rather than a class because that is all it is" does not. ## What never goes in The middle ones are what creep in while the code is being written; the last is what creeps in while it is being tested. **1. The name, spelled out.** `getTtl()` returning "the ttl" documents nothing. **2. A defence of an absence.** Why the type has no further parameter, why a method was not added, why an alternative nobody proposed was not chosen - none of it belongs in the doc: ```java // no * The declared type of the result is not among the components on purpose. It travels on Operation but * stops here: it belongs to the call site, which every entry that reads a result back alre