There's not just one type of "crafter": there are (at least) two.
Some people like crafting for its own sake; they enjoy the process of making new and interesting products. Others take up crafting as a sales game; their interest is in economic competition. Designers need to recognize that these are different kinds of players with different goals who aren't going to be motivated by just one kind of reward.
The "pure" crafters are best described by Richard Bartle's original fourfold typology of online gamers as Explorers, while the players more interested in crafting as a sales game are much more likely to be Achievers. Explorer crafters enjoy playing with the options and features of the crafting system; their goal is to explore the possibility space. Achiever crafters feel satisfaction when they can dominate the market for the sale of the products they produce; their goal is to outcompete other sales crafters on either quality or volume or both.
What this means is that we get confused when we talk about what "crafting" should be like. For an Achiever, whose interest is in accumulating things (like high-end objects and enormous bags full of money), improving crafting means streamlining the production process (more goods produced per unit of time invested) and exposing more goods to more potential buyers (more income per unit of time invested).
But for an Explorer (like me), improving crafting means that the research and design and experimentation and architectural and creative aspects of building new things should reward ingenuity, not time invested. "Crafting" to an Explorer means playing with the rules of a production system, not actually operating and maintaining such a system.
These are two radically different motivations for play. One is about adding to the economy through the creative generation of novel products; the other focuses on competing for scarce resources (recognition as the "best" and other players' money). Confusing the Explorer and Achiever playstyles -- treating them as identical because they're both "crafters" -- leads to bad gameplay and unhappy gamers.
New crafting recipes are a small benefit to both types of crafter, but ultimately they're more important to the Achiever crafter. For the Explorer crafter, once a recipe has been used, the novelty that made it interesting is gone. The Achiever crafter, on the other hand, can happily crank out many units of the new object if it's perceived as something other players might want to buy.
Thus a developer belief that all "crafters" need to remain satisfied are a few new recipes is incorrect, because not every crafter benefits equally from new recipes.
Explorer-type crafters should not be forced to try to play like Achievers. That's never going to work well. Not only does this disadvantage the kind of crafting they're good at -- crafting as a creatively constructive activity -- there's no way they can compete economically with players who'll do anything to beat them. The result can only be to unnecessarily drive away the Explorer crafters.
There's nothing inherently wrong with sales competition, and for those who like that sort of thing, it's a lot of fun. It's also economically useful in a MMOG because it helps to maintain an efficient market for goods. Crafting-as-sales is good to have in a game.
But why must crafting be only about sales competition? Both creativity and competition are worthwhile motivations, so why not design crafting systems that support and reward both motivations?
The alternative is to explicitly design both a production-focused crafting subgame and a process-focused crafting game. These should complement each other. Process crafting would be the means by which new products and processes are discovered through experimentation, and should be required to purchase specialized tools created by production crafters. Rather than being forced to subsidize their work through sales, they might better be supported by some kind of salary, with perhaps an R&D bonus proportional to how often products made from their discoveries are used by consumers.
Meanwhile, production crafting should be designed as an optimization game that rewards whoever can generate the most units of product at the least cost for the most quality. They need new recipes and fabrication processes from R&D crafters (perhaps acquired through a bidding system for recipes and processes where the identity of the R&D crafter is hidden) in order to offer new products to consumers.
This dual-track but unified crafting system would serve all players. Explorer crafters would be rewarded for their skill in discovering new aspects of the crafting system; Achiever crafters would be rewarded for their skill in making things; and consumers would gain access to desirable products.
Side note: My personal preference is actually to generalize crafting as a resource delivery system. Following this perspective, it might be fun to offer four types of "crafting": R&D, production, distribution, and sales. R&D players could design new products and processes; production players could build optimal products and processes; distribution players would act as traders, carrying products as cargo; and sales players could seek to build fortunes by placing buy and sell orders. It might also be worthwhile to define some skills or classes for hunting/prospecting so that players can travel the world supplying raw resources for production -- this would combine the resource game with physical exploration and some PvE combat.