There are some popular terms out there in the Intarwebz for game rules. One is "rules as written" (RAW). Another is "rules as intended" (RAI). As I understand the distinction between the two terms, one refers to the rules as described by the text, and the other refers to the rules described by the text. Wait, those are the same thing, what gives?
What gives is that there are people out there, ignorant people, who believe that game rules are written in a form of pseudo-code, which can be debugged by means of interpolating colour text, comments made in interviews by the designers, comments made in FAQs, and rules written down in similar texts.
Rules, to my mind, are best described by formal symbolic languages. But as anyone even vaguely familiar with formal logic can tell you, translating informal or even semi-structured natural languages into formal languages is difficult, though not impossible. Jon Barwise and John Etchemendy, in The Liar, go into intriguing detail about how even apparently simple statements in natural languages such as Liar Sentences (ex: "This quotation is false") can be interpreted in myriad ways.
But rules are best described by formal symbolic languages because those formal languages are designed specifically to strip away obscuring features of natural languages to render statements un-ambiguous. Hence Barwise and Etchemendy come to the conclusion that Liar Sentences are ambiguous because it has no definable feature independent of a semantic model that forces it to be paradoxical. And wow do they ever do some fun things with semantics, offering a very readable introduction to both Zermelo-Frankel set theory and anti-foundational axiomatic variations. If I didn't know better I would have said that they were showing off an AFA set theory.
The point being that if you raised the notion of RAW and RAI with a linguist, or a philosopher of language, or a logician, or perhaps even someone familiar with the arcane lore of 'reading', they would think you were illiterate or stupid for suggesting that where the rules indicated some rule A, they actually indicated some rule B because either that was how they were coded, or how they should have been coded. They would think you were illiterate or stupid because the rule is written to convey the author's intentions regarding the rule, 'describing' if you will, to the reader.
Which isn't to say that this curious pathology of literacy is limited to gamers. It's a fairly common phenomenon, and in fact so common that the philosopher Derrida made his career pointing out that the context, and especially the context of reflection and critique, made texts infinitely interpretable. So it's actually one expression of a deeply ingrained cultural need in literate societies to find the gaps in writing, and to fill them with whatever baggage is available. In the West we either have intentions, some sort of vague mythology about goal-oriented behaviour, or we have that bastard offspring of numerology and math: coding.
However, perhaps to Derrida Barwise and Etchemendy, natural languages are full of incomplete, ambiguous, and otherwise loosely interpretable phrases which are grammatically well-formed, etc, but incomplete like the Liar Sentences are incomplete. Yet only conspicuously few such expressions are given the exalted status of Paradox as the Liar Sentence. That is because, unlike the Liar Sentence, most of the expressions and phrases used in natural languages to express things like game rules and why you need to learn to read fit into undefined, yet sufficiently clear semantics.
Games, for example, are often lucky to have as part of their semantic model somatic components, so that when the rules state that object A is moves two squares left towards object B, they refer to actions, objects, and relations. Awareness of how written rules are mapped by readers onto the tasks they wish to accomplish is the essence of technical writing, which is a discipline frequently avoided by people attempting to write game rules, perhaps because they understand that rule books are frequently about more than concise and usable reference material. Also that people using a help system or manual are already angry, and only doing so for instrumental purposes, whereas rule books are used to facilitate pleasure, sell game components, and encourage people to enjoy the game.
That not all books are the same, or have the same goal in mind, is important for literacy. Likewise is the notion that your reading of any material is coloured (if not outright distorted) by one's personal history and experiences, as well as any critical or meta-analytic material written about the source material. If you are going to read properly, you need to do more than recognize how symbols on the page compose words, and words compose sentences, and sentences compose paragraphs, you need to be able to understand how works and phrases express terms, make statements, relate feelings, speak figuratively, encode ideas, and describe the world.
Which isn't to say that authors don't make mistakes, or that they aren't constrained by the number of column inches required for a German translation to retain the page numbering, or simply that they presumed a certain floor-level of emphasis on sportsmanlike game-play would prevail a player's need for reliable game play and reproducible results. It is to say that using terms like RAW and RAI in discussions about rules makes you sound like a hippy sales-bitch trying to describe how quantum mechanics enables the energy in the crystals she's selling to open both your chakras and your wallet.
So, to summarize for the tl;dr crowd.
1. Written rules convey intentions through symbols, grammar, syntactic structure, and reference to a semantic model. The rules as intended are the rules as written. All arguments involving RAW vs RAI are therefore automatically 'false' in value.
2. Determining the meaning of written rules requires literacy skills such as familiarity with grammar, syntax, and semantics.
3. Truth can only be determined tentatively and even then only by harmonizing method, sources, and goals.