Some initial thoughts:
1. Context Dependence. The problem with crafting magic items in a D&D-type game is its value and effectiveness varies wildly depending on the individual table:
- does the DM customize magic item drops for characters (wither through a wishlist or a subjective, Billy-uses-a-scythe-so-here's-a-+1-scythe approach)". DM-compicty reduces the value of any crafting.
- are magic items purchaseable? Magic emporia reduce the value of any crafting.
- are magic items plentiful? Not just the Christmas-tree thing, but whether a character needs to work to achieve a magic item -- is it part of the actual campaign? Crafting becomes an alternative to that, but it risks pushing it into the background.
Given those variables, any crafting system that works in one context is going to fail in another. We can see that with the Artificer, for example, which trivializes getting certain high-end items (it presumes that any player would have access to them by another means in any case).
2. With checks involved, there is the question of permanence. If a player doesn't like what they receive, can they get a do-over? If yes, then the value of crafting is again reduced. If I roll badly once, and don't get what I want, what are the consequences?
3. Obstacles in real life play differently in-game. For us, the limit that something takes significant time is essentially handwaved in game ("okay, you all work on your items, and meet again in this spot two years later") and requirements for specific resources are individual, and not party based: I have never played a game where multiple characters needed specific ingredients, and they took turns pursuing each others' quests. If it is possible to reduce the difficulty (or achieving time/resources), they no longer become obstacles.
4. I've been thinking about the 2024 D&D crafting rules recently, and there are some hacks that let players have what they want with relative ease. There are fixes to them available, but unless they're errata'd, it's unlikely to have a wider impact.
That said:
I can see a place for crafting a personal item, an ARTIFACT that gains abilities when you advance to a new tier, whose effects are not certain in advance. "I want a magic trident"; "... ring", "...staff" "...robot dog" or whatever, and players can draw from a series of small tables, varying selection and randomness (e.g. 6 tables with six possibilities on each, and players either roll for which table they use, but then get to choose, or choose a table and then roll for the ability gained). Once you make your character's artifact, it exists in the world, but it's yours, bound to you, and the player doesn't get a do-over.
These items should be multi-use, with some minor properties and flavour mechanics to go along with the major abilities.
Variant: artifacts are made with a certain number of build-points, and those points advance by tier. or level. Maybe 1 build point per level when first crafted, and at 6, 11 and 16, the item (or items) bump up -- so there's a bit of an incentive to delay (getting your artifact at level 8 means you have a stronger one until the party's at 11). Major abilities cost 4-6, minor ones cost 1-3, perhaps. And the elements of randomness could still be at play, as could disincentives for an extra point or two.
This approach, with one major item per character, has its parallels in fiction, and the challenge then is for it not to feel too much like the D&D cartoon from the 80s. But it solves the problems outlined above:
1. it's only ever one item (or one pool of build points) per character.
2. it grows and develops with the character, shaped by what the player wants but not limited to a wish-list
3. it allows some player agency as to when it comes into the game (also shaped by in-game events), but does not create illusory obstacles of time or money
4. It circumvents hyperoptimized meta-strategies, such as true-strike scrolls, which (for me at least) work against verisimilitude within a fantasy world.
I'm under no illusions here; this won't be the solution everyone wants. But I do feel that unless a solution addresses these concerns to some degree, especially the tension between player desires and costs vs character desires and costs, any crafting system is going to feel icky.
