I'm going to work through the elements of the SNAFU Cookbook publicly, because I don't think I have all the answers, and I'd be interested to see other people's input on the project. Once the series of posts has reached it's conclusion, I'll revise my ideas in light of those suggestions, and compile the final cookbook from those revisions. If there's enough interest, I might even throw a few commissions out to people willing to write short pieces (200-500 words) that can be added to the book so that there are more voices than just my own within it.
First a look at the ingredients. Note that these ingredients aren't discrete and separate things, they overlap, they exist in relation to one another. There are probably other ingredients to consider as well, and I'm certainly open to hearing ideas about additional concepts that contribute to a game.
An oeuvre is a body of work, it typically has themes running through it or elements where one piece in the wider collection connects to other pieces creating a continuous whole that transcends the sum of the individual pieces.
Orchestrally, we might look to the works of a single composer, where similar musical motifs are embedded into different compositions, or where the composer explores particular themes from different perspectives, or even a propensity for using a specific type of instrument.
Cinematically, the term oeuvre is most commonly associated with directors like Alfred Hitchcock, Kevin Smith, or Quentin Tarantino. Hitchcock's oeuvre is linked by thematic elements, methods of lighting, camera angles. Smith's oeuvre is linked by locations (New Jersey playing a prominent role), characters (notably Jay and Silent Bob in his early works), and again by exploration of themes. Tarantino's oeuvre involves hyper violence, pushing the boundaries of language, evocative soundtracks, and a stable of regular character actors. The individual works within a director's oeuvre will be different, but by virtue of a singular artistic vision there will be something that links them. You generally know what you're going to get when you go to a movie theatre to watch a film by one of the three directors mentioned. It can be enough to get a certain crowd to buy tickets, and can similarly drive off other segments of the population who have shown a dislike for films from that film-maker.
I'm certainly thinking of the term oeuvre cinematically here. The players are the regular stable of actors; where some may be character actors who exhibit the same range of traits regardless of who they portray, others may be method actors who really immerse themselves in a role. Themes are common elements and types of scene encountered by the characters; there might be regular violence in the game and the game might say something about this (or might actively remain silent...both are just as important as artistic choices), perhaps the game plays with the concept of relationships, or transhumanism, or struggles of race or class. Another element of the oeuvre might be the simple way responsibilities in the narrative are assigned; does everything come down to a GM's adjudication of the dice, do different players on the table take responsibility for different elements of the world, is there a GM at all?
Every session is like a new movie, or a new episode of a TV series. The tighter the oeuvre, the more similarities there will be between one session and the next, the more elements will be common across the collective works. The looser the oeuvre, the more free-ranging the individual sessions are, there might be vague linking concepts like a recurring NPC, a monstrous entity or hideous situation that appears in a few sessions, a focus on groups and politics rather than individuals... a few things might define the oeuvre, but this doesn't mean that every session requores mandatory appearance of every element.
Often an oeuvre is not predefined. Like an artistic movement, it's only in retrospect that the common themes and elements start to become apparent. That can make it difficult as a tool to guide future elements of gameplay when a campaign is first begun, but after a few sessions the ideas that work can be retained, while the ideas that don't can be discarded as a part of the play environment.
A traditional fantasy game is divided into races and classes. Races are things like "humans", "elves", "dwarves", "halflings", and "gnomes" on the side of good, while "goblins", "orcs", and "trolls" are on the side of evil. Classes are things like "warrior", "wizard", "thief", "cleric", and "everyday peasant who isn't important". But why is this the case?
A science fiction game might simiarly divide it's population into races and classes (but might instead call them occupations). The races of such a setting would still tend to include "humans", but might add a range of "aliens" with unique powers and weaknesses, "mutants" who have deviated from their parent race in some way, or "AIs" created for some purpose but now transcending it. The classes/occupations might be more refined in such a setting, with "marines", "mercenaries", "bounty hunters", "pilots", "smugglers", "psychics", "scientists", "engineers", "technicians", and others.
Wider demographic categories tend to be fewer in number but allow for a variety of options within them, while narrower categories tend to be more plentiful but very specific. Sometimes there is an option between two defined demographics, "half-elves" are a common racial example of this in a fantasy setting, while a "marine/pilot" might be a science fiction example of a multi-classed person who can fight reasonably well, and pilot long haul ships reasonably well, but when it comes to flying a starfighter they are truly accomplished.
Consider the types of stories you want to tell, consider the types of characters who would take on major roles in those stories, make them options for player characters. Then consider those who would play supporting parts (or potential antagonists), make them into a list of retainers, allies, and contacts; possibly allow them as player characters but make the players aware of what they are getting in for. In a story about samurai, the main characters would probably want a swordsmith around, but unless the swordsmith has other abilities, it could be a very deadly or very boring story for them. Finally, make a list of things you definitely don't want in the story. The elements which aren't present in an environment can be just as defining as the elements which are. A fantasy setting might be expected to have elves in it, but if your story doesn't revolve around elements where they are a natural fit, why have them? If they might add something unusual to the setting, include them but change them from the stereotype in some way. Science fiction can still be science fiction without aliens present... stories can also work without humans.
When defining the races, classes, and other groupings in the setting, consider what you want to be commonly available to everyone, what might be generally available, what might be restricted, and what might be exclusive to one particular group. If you expect lots of physical violence, then make combat abilities available in different ways to different groups, if ypu want a grittier world where violence will get you quickly injured or killed, then maybe restrict who gains access to these abilities. The same applies to magic, technological know-how, political intrigue, or anything else the story might be intended to revolve around (or have the potential to revolve around).
Different groups (regardless of race, class, or other method to define them) should have different agendas and motivations, different tools and techniques are their disposal to achieve those ends, and different obstacles that are preventing them from achieving them. It's often at the intersection of these that stories come alive. Consider who needs what, why they need it, who could help them and what might be necessary to get that help. The characters will be at the centre of all this. Who are the "bad guys" unable to be portrayed by the characters, and what is their motivation? Remember also that the demographic divisions don't need to be defined by race and class/occupation, they could just as easily be divisions between noble houses or guilds, geographical distinctions, socio-economic, religious, anything that you think might make a good story.