I can understand designers thinking of crafting as being about servicing the game economy. In other words, designers design a crafting system as an economic support system -- its purpose is to insure that other players can have the stuff they need/want to play the game.
Well, yes, that's a useful function of a crafting system... but what does that offer the people who like to make things?
My problem with crafting systems in nearly all current and discussed MMORPGs is that they're not focused on satisfying the people who craft because they actually enjoy crafting. "Crafters" aren't all automatons who thrill to the opportunity to grind out gazillions of copies of some uber weapon, nor are they all hypercompetitive economic PvPers who see crafting as a sales game -- they are creative people who enjoy imagining new things and bringing them into the (game) world. It's the enjoyment of making new things in and of itself that appeals to the crafter personality, not the utility of the end product.
Accordingly, my one great interest in any MMORPG's crafting system is that it be more about process than product.
Don't worry about whether crafters will crank out enough units of Product X to satisfy demand. Make the process of crafting so much fun in and of itself that lots of people want to do it and (assuming you also implement a free-market player economy) the results will take care of themselves.
So how can this be accomplished? I believe the three keys to enabling a fun crafting process are:
1. Great variation in resource and configuration inputs is possible.In a little more detail:
2. The attributes of outputs mirror the attributes of their inputs. (a nod there to the Doctrine of Signatures)
3. Players can choose to trade quality for quantity in outputs.
1. By variation of inputs I mean two things. First, there should be a vast number of resources/components needed to make things, and all those resources and components should have multiple attributes of varying degrees. And second, there should be a appropriate number of ways to connect each of those resources and components to each other.
Variation of inputs allows players to experiment with inputs to see what works best for a given purpose. It supports treating crafting as an exploratory process.
2. Allowing outputs to derive their attributes from their inputs gives structure to crafting exploration. If through crafting some test items I can learn that using a certain kind of wood or a resource with a high level of copper turns my finished product green, that's information I can use later if I ever want to make something that's green.
Similarly, letting crafters experiment with how resources and components are connected to each other (in more complex crafting procedures) should also have some plausible effect on the final product. Perhaps connecting a flywheel to an engine through a gear will make the complete device more efficient, though perhaps also resulting in a higher maintenance cost due to more parts....
3. Letting players decide whether to optimize an individual crafting process for quality or quantity is a way to support both creative crafters as well as those current players who regard crafting as a competitive economic game. If you want to try to corner the market on widgets, OK, but optimizing your crafting process for quantity will mean that you can't make the "best" widget possible. Likewise, if you want to try to make the most perfect widget possible, you're only going to be able to make a very small number of them, and the cost in failed experiments will probably be pretty high.
A crafting system designed to incorporate these principles for focusing on process would, I think, be one that satisfies everyone's needs. The artistic crafter can create; the exploratory crafter can tinker; the sales crafter can meet market demands; and the developers can all buy Ferraris with the revenue from the subscriptions from happy gamers.