Without this playing with fantasy no creative work has ever yet come to birth.Been working on my RPG-setting fantasy stories. If I may be so bold, I think it may actually be safe to say that the distinction between RPG-type and other fantasy is, essentially, nonexistent; RPG-setting fiction, when it's worse, is worse because it's a tie-in, not because of some characteristic of the setting—and Dragonlance is better than roughly half the non-RPG fantasy on the market.
Look at Conan and Lankhmar—there is a reason those two things were Gygax's major inspiration. The gent behind "Empire of the Petal Throne" is the only person who ever undertook something on par with Middle Earth, and he published a game based on it with TSR. Moorcock may talk like a White Wolf LARPer, but his books are pure Munchkin from end to end. Even George R. R. Martin can be compared to a certain RPG.
Yeah, I said it. Say it again if I had to. Seriously, how has the existence of FATAL not put an end to "dark fantasy" once and for all? The phrase "Socratic irony" might be bandied about.
Anyway. So, I'm working up my writing systems—helps that you can make truetype fonts with Inkscape—and, dude, I'm a better linguist than Tolkien. Okay, not really (if I have occasionally seen further than him it is by standing on his shoulders), but now that I have your attention, I will say that I'll put the realism of my method up against Tengwar and Cirth any day of the week.
Tengwar and Cirth basically work like hangeul—due to, like hangeul, having been created by a linguist. They add features—loops, arms, lines, etc.—based on sounds being related to one another. E.g., in Cirth (the "rune"-looking one, as opposed to Tengwar, which is more like Arabic or Devanagari), "P" looks like the letter P, while "B" looks like R. "M" looks like our B, and "F" and "V" are backwards versions of "P" and "B" (possibly they're ɸ and β rather than F and V, since those are the fricative versions of P and B, while F and V are labio-dental rather than purely labial).
My fantasy alphabets, on the other hand, like all known real writing, are derived from pictographs. You can't tell what their pictographic origin is, but then, I defy you to see the Egyptian hieroglyph-origins of any Roman letters except M (it comes from mw, "water", and depicted a little wavy line).
I first pored over tables of the development of several of the world's scripts (Phoenician prominently), found letters whose look I liked, then found words in my Elvish language for the things they derived from. Then I assigned the letters to sounds based on which sounds the relevant words started or ended with. I had to fudge a few times, e.g. for the one deriving from a hawk (it kinda looks like a 3), I used "hunt", symbolized by a bird of prey.
Of course, an elvish generation in my stories is c. 75 years—the civilization humans call "the Ancients", that fell a thousand years previous, is only as remote from the elves in the story as the French and Indian War (or the Seven Years War of which it was the New World theater) is to people nowadays...except nobody who saw those wars is still alive, and still-living elves saw the Ancients fall. But the point is, the elvish alphabet would have to develop rather slower than, e.g., the 700 years from Proto-Sinaitic to the earliest forms of Canaanite writing, or the 1500 years from Egyptian hieroglyphs to Proto-Sinaitic (all told, 2200 years). I think 12,000 years isn't too long of a time for that development to occur, given it'd be 8250 just based on the length of their generations. Of course, most of that development happened before they ever met a human, but then again, most development of our alphabet happened before any Egyptian or Canaanite ever met a Greek.
Another thing I do different from Tolkien is that the dwarves have their own writing, rather than using an elvish one (Cirth, the rune-looking Tolkien script, was the writing used by Sindar and Moriquendi before the Noldor brought the Feänorian Tengwar back from Valinor—the Khazâd like it because it's easy to carve into things). Of course it'll be more angular and rune-like than the cursive elvish script (remember, shoulders of giants), but I'll use the same method to create it—the earliest phases of the Canaanite script were quite angular, and South Arabian (ancestor of modern Ethiopian writing) has actually been compared to runes, much like Hungarian rovás (literally "notching, scoring") has.
PS. If you look at that chart of Hungarian "runes", you'll notice a lot of the consonants use different letters depending on the preceding vowel, e.g. there's a different letter for "ek" and "ak". That's because Hungarian has "vowel harmony"—the feature is also seen in the Old Turkish "runes" they derived rovás from, since the Turkic languages also have vowel harmony (we're still not sure if Hungarian and Turkish are—albeit distantly—related, or if vowel harmony is just what's known as an "areal feature", where unrelated languages—e.g., Tibetan and Sanskrit, or Finnish and Swedish—start doing things similarly because their speakers interact).