[this is also probably a "part one" of something I'll come back to later]
I'm trying to work up some rules for Sorcerors for the Skin of Naranpelo, and it occurred to me that part of it is that I'm still too much in the space where magical theory slows me down as a designer.
See, a "spell" to me is a manufactured item of the scientific method being applied to magic. "If I match this ingredient and this chant and this hand movement, every time I am going to have this result." This is classic AD&D territory. For game purposes, this is going to be how Alchemy works in Naranpelo.
It doesn't feel like "magic" to me.
(This is also why I tend to play Clerics as having no real "spells" but the ability to 1) pray for divine (or divine messenger) intervention/invocation, 2) manipulate energies relating to the sphere of influence of their deity, and 3) handle basic "spiritual" details, from blessings/curses to consecrations/desecrations. Being a Cleric in one of my games generally allows you the ability to "put your faith" in the deity and let the GM decide how that's going to come out based on your rolls/method of determination.)
When I run my amended AD&D games, a basic magic-user has a grimoire that contains all the basic spells for 1st through 3rd level. You still need to be able to Read/Write Magic for scrolls, to figure out traps and glyphs, and really, the journey to 3rd level spells generally is a matter of figuring out, "I know it says THIS, but it just isn't making sense. That completely disregards the laws I know from the first level spells," and that sort of thing.
(The advantage of being a high-level wizard in my campaigns is that you can actually use "conversational magic." [grinning])
In one of my more famous Shadowrun campaigns (actually, all of them are famous amongst a certain group, although infamy really is a better term...) we had an assassin/bodyguard type who had learned a certain set of movements that usually indicated a very high-power, very nasty spell was about to be cast.
This was a character detail. He was cyber'd enough that casting magic really wasn't in the rules, nor would he be able to cast this particular spell, and, well, I don't recall anything in the rulebook that said there were any common somatic ingredients to spellcasting.
The reason I allowed it is because I love the drama of the moment when you see the spellcaster start to draw energies. It's the same argument against demonstrating telepathy on television - the actors alway have to point at their heads or narrow their eyes in concentration or something. Whether the wizard waves his arms, or dances her steps, or goes to his happy place that smells like cookies, that's a dramatic moment and it ought to register in some fashion... if only as a vulnerability.
(For me, magic loses something if it can be activated at the push of a button, which is another argument against "spells" in a design. Automation starts to lose its appeal as "magic.")
I love situations in games where I can develop spells based on the ingredients and events "at-hand," with the usual provisions applying. [1] I want it to qualify by the rules (even unwritten) in play. I want it to be based on possibilities that make sense by the feel of the game.
When Princess Krrsed came upon the Real Newspaper, a spell to change herself into a paper airplane did not seem entirely unreasonable given the other over-the-top events of the scenario.
So that's where my sense of the home rule begins. Communication with the player and GM: "What do you want to have happen if you succeed? What happens if you fail?" For a spell, "Is this replicable?" "Do you have a name for it?" "What makes it special?"
[1] "Usual provisions" include a GM with a generous sense of humour, a fiendish sense of opportunity, and a friendly ability to handwave for the sake of story, I suppose.