I contend that one of the major causes of misunderstanding in discussions of tabletop roleplaying games is differential prioritization of where rules should live. At base, games are bundles of practices that can be stored and communicated in various ways. For example, baseball in the United States started as a game played by amateurs using informal rules that lacked textual basis. A social club wrote the first baseball text in 1845: the Knickerbocker Rules. Tabletop roleplaying game rules can also reside ultimately in culture or in texts.
The community that developed around the Forge in the early 2000s is one influential example of prioritizing texts. The games developed by designers from the Forge and the subsequent Story Games forum tend to have clear, explicit procedures and written creative agendas. Critique often focused on potential disconnect between what a game text promises, either explicitly or implicitly, and what the game delivers. For example, Vampire: The Masquerade promised “struggle for humanity” but “always ended up as emo superheroes” (ref). This disconnect tends to be framed as a flaw or a sign of broken rules. In this tradition, to understand games one returns to the original text, tries to follow the procedures carefully, and then evaluates the outcome. For example, here is Luke Crane on Moldvay D&D and Ron Edwards, quite recently, on 4E D&D. A different community that exhibits strong textual fundamentalism in a different way is the portion of Pathfinder players concerned heavily with tactical game balance and character build optimization.
In contrast, some tabletop roleplaying game traditions deemphasize singular texts and are more likely to prioritize norms and expectations communicated informally. The community that developed around the old school renaissance seems closer to this approach on balance. People seem perfectly comfortable with a more distributed storage of practices, drawing from resources such as the old school primer, assorted blog posts, and recalled experiences. In this approach, no single text contains the full game and various aspects may lack textual basis completely. This is not just house rules, where players modify a canonical text to suit local group preferences, though house ruling is a part of how rules evolve culturally. Instead, the full game is more like a cultural tradition rather than a solid, defined, bounded artifact.
In terms of game design, both approaches have pros and cons. A strong textual basis serves as a shared landmark. People or groups that differ in expectations but share a specific, relatively explicit text, such as Apocalypse World, Pathfinder, or 5E D&D, might be able to communicate more effectively compared to people that lack such shared singular text, all else equal. Culture, however, is extraordinarily effective in social coordination, requiring little explicit deliberation to function. For example, norms are more influential in directly coordinating social behavior than laws, and when laws do come into play, few people other than judges and lawyers are familiar with legal details.
The major misunderstanding, as far as I can tell, is the idea that the choice is between system and individuals making things up rather than between prioritizing text and prioritizing culture. For example, in a classic Forge document, Ron Edwards writes, to characterize an objection to the importance of well-designed systems: “It doesn’t really matter what system is used. A game is only as good as the people who play it, and any system can work given the right GM and players.” And, from a DIY D&D perspective: “sometimes people just suck and redesigning the game won’t fix that” and “Reading the book is not and never should be essential.” Now, of course there is also an effect both of well-crafted procedure in text and individual creativity, but players (and cultures of players in aggregate) can differ on preference for textual versus cultural embodiment of rules.
Personally, I see the benefit of both approaches. I spend a lot of effort trying to effectively proceduralize rules that I develop, particularly for many of the less fluent aspects of traditional fantasy games, such as encumbrance, resource management, and bookkeeping, with the Hazard System being a more involved application of this kind of thinking. However, insisting that individual texts be highly pedagogical and entirely self-contained both creates strangely unmoored documents, seemingly unaware of their likely audience, and ignores the massive, brilliant social fabric to which we all belong.
That said, this post is more about facilitating communication than advocating for a particular approach to game design. People with a strong focused design bent might want to consider that games can be stored culturally, similarly to the way traditions develop and propagate more broadly. And, people who feel like they already know how to play and just want the text to get out of their way may find that trying to play some innovative or experimental games may suggest new techniques or even be enjoyable as completely self-contained games. At the very least, recognizing that people may differ on this preference might help prevent misunderstanding.
Credit to a post by Dan M. that helped me crystalize this idea and provided some useful language.
See also: game design as common law, though that post concerns a method to craft rules in the context of a particular group, rather than broader cultures of gaming.