The other big distinction is that with improv, you don't need to come up with anything extra.but the distinction between pre-prep and improv in terms of output is almost nonexistent if you're doing your pre-prep right.
That sounds like the exact same distinction.
I advocate not detailing too much, so that you eliminate the rails. Everyone "railroads" to some degree if they prep but not improvise.
You're working from a false premise. Proper prep eliminates the rails. A railroad is, in fact, defined by a lack of prep.
None are likely to be unused, unless the PCs skip things or ingore the prep entirely. This is likely a breach of social contract.
This is another false premise. There are plenty of games in which the PCs are expected to skip things they don't find interesting.
(2) You need to design the city of Tharsis. (Where is it? What's it like? What can the PCs do there? Et cetera.)
Only minimally, and only with reference to what the PCs will need for this adventure. "Sinig city on the coast, not far from the derelict ship, white marble city in Greek arcitecture style, with a huge mountain behind it." The rest can be improvised.
In other words, you answer the questions: Where is it? What's it like? But you don't answer the question, "What can the PCs do there?" Even though you claim that all prep should be aimed at the PCs. Would you care to explain the obvious discrepancy in your position?
Sure, the parts that might be a battle site should have a map. Otherwise a short description and improv are sufficient.
No. You need the villains' stats if you want to have a fight. Nothing else is needed. Improvise at the table if the PCs decide to fight the navy.
What you're failing to acknowledge is that the GM isn't making the decision of "where will the PCs fight on this ship". Nor is he making the decision "will the PCs fight the navy?". The PCs are making those decisions (or, at the very least, contributing heavily to them).
You, OTOH, have a vision of the story that you want to tell. You claim that you're "open to other possibilities", but you also advocate the type of limited, narrow-minded prep that you admit causes "everyone" to "railroad to some degree".
The only way you know that (a) the PCs will board there derelict and fight the villain, but (b) not attempt to fight or evade the Tharsis navy is if you either railroad them or have an amazing gift of prescience.
And that's why you're running into problems in this thread: You're refusing to acknowledge that not everyone runs their games the way that you do.
By your own description, you not only design elements of the world, you also pre-determine their purpose. You don't just design a "derelict", you design a "derelict that the PCs are supposed to explore". You don't just design a "bad guy on the ship", you design a "bad guy on the ship that the PCs are supposed to fight". You don't just design a "Tharsis navy", you design a "Tharsis navy that the PCs are supposed to surrender to".
But I don't GM like that. I design a "derelict" and then I let the PCs do whatever they want with it. Explore it? Sink it? Tow it? Repair it? Fight on the deck? Fight below decks? Fight while swinging from the tattered rigging? Whatever happens, happens.
Saying, "You should only design the parts of the ship that might be part of the battle site." is a nonsensical instruction for my style of gaming. Battle site? I don't even know if there's going to be a battle. I certainly don't know where on the ship that battle is going to take place.
You said that you felt "insulted" because I suggested that you just flat-out don't understand how other people play the game. But then you posted a message confirming that you don't, in fact, understand how other people play the game.
World Building - the act of creating background that is independent of plot. The act of creating background FOR IT'S OWN PURPOSE.
Setting Building - the act of creating background that is dependent on plot. The act of creating background TO SERVE THE GAME/STORY.
But if you GM like I do, then the distinction is non-existent. Elements of the game world are like toys in a child's playroom. Your game features structured play time -- you have selected very specific toys that the children can play with and you even have some guidelines on how they're expected to play with them. There's nothing wrong with that. And, in that context, it's fair to say, "You shouldn't buy toys that the kids aren't going to play with."
But my game, OTOH, just features a playroom full of toys and the kids are allowed to play with whatever they want and however they want. In that context saying, "You shouldn't buy toys that the kids aren't going to play with." is meaningless. I buy a toy, I put it in the playroom, and they either play with it or they don't.
Within the confines of the analogy it still makes sense to prioritize my toy buying budget based on the types of toys I know the kids like to play with. And maybe there's some sort of rigorous methodology for organizing the toys, so I shouldn't buy toys that would be placed on high shelves the kids probably can't reach (i.e., designing a village on the far side of the game world that the PCs are unlikely to go to).
But the distinction between "toys they can play with" and "toys they can't play with"? Doesn't exist. They can play with all the toys and they can play with them however they like.
And the reason why D&D rulebooks and magazine articles are predominated by articles supporting my style of play and not your style of play is because that's been the default style of D&D play since 1974.