9 hours ago, frob said:

In small games, I've seen merchants with a capacity over time. I've seen merchants that are available only for limited hours, and while present they can only accept a limited count of items. This could be expressed in many ways, like "I'm sorry, my cart is full, I can't take anything more." In other games (particularly Nethack and family) merchants have a specific amount of money; If you buy things they have more money, but once you sell them enough things they stop giving you cash and give you store credit instead.

Well... if we make an assumption here and say that the cart is full of wood.... The Merchant would need to purchase that wood from the grinder.... When the Merchant has enough wood, the grinder would then sell to Housebuilder or Breadmaker or any other multitude of characters.

Store credit is a big gamble, because if the Merchant finds that the conditions have changed and suddenly is getting short on funds, chasing up creditors, will be a chore unto themselves... Its a concept that shouldnt be dismissed and a great point  

9 hours ago, frob said:

They scale badly to enormous online games.  It isn't hard to track a small number of merchants for a single player.  It is more difficult (but not impossible) to scale that to thousands of merchants and millions of users. The dataset is small (probably smaller than each user's profile image) but the infrastructure is not typically a priority, and the story breaks down: a merchant with thousands of customers per hour doesn't have cash flow issues.

I agree, however, who buys the iron and wood and leather???
What if another Merchant opens up down the road, selling near identical items cheaper...? (by cheaper, it would mean less profit margin as the items would cost approx the same to craft and the Merchant would simply add his/her profit on top)
Being just a Merchant can be a lot more involved