Oh boy, opening up with Krugman. I can see the Neo-Classicists who try to paint all Keynesian thought of every worthwhile economist in the past half century as a delusion spread by the "evil" Krugman getting ready to swarm already.
To be fair, it's difficult to for neo-classicists to complains about Ricardo, given that he's pretty 'classic'. They may as well complain about Mr Smith.
OK, first, it's probably worth bringing up a basic link to Comparative Advantage itself, because Krugman isn't talking about what Comparative Advantage itself, but the reaction to it.
Thank you for this link, which I should've given. Almost all of your post is along lines with which I would agree, so I'll just focus on a couple of particular points.
For comparative advantage to work, you need goods to cost more for a given culture to produce them. Until we have the internal mechanics of having to pay dwarves for their work again, this can only make sense in terms of how much labor it takes to make a good. Elves must somehow produce a good with less labor than a dwarf does... but there is nothing in the game that does this. Every good takes the same amount of labor to produce - one harvested raw material, and one step in a workshop. In fact, the act of using the trade depot is more work than any other single industry. The "cost" (in the only currently meaningful definition of the term - effort and time on the part of the player) of trading itself is greater than the "cost" of just making it yourself.
There's at least two options here: simply lowering skill acquisition rates and hence the length of time the workshop step takes, or adding in additional player input through multiple steps for a particular type of labour. The latter would add to micromanagement, which isn't necessarily bad - this is DF, after all - but may prove annoying. The former would slow down the early game considerably, but could push the player into focusing on a stronger division of labour, which can only help to personalise the dwarves. It's important to recognise that the biggest cost to the player - at least in the early game - is that of opportunity, and imposing additional opportunity costs makes decisions more difficult, and hence more interesting.
The problem is that Fortress Mode never lacks for productivity. There are no goods to distribute out to your dwarves, there is no consumerism, there is only having enough to survive, and then just sitting on your fort and trying to see how long you last. Maybe building a big monument or fighting something more difficult, but basically, all you need is to not starve and have clothing and weapons. Everything else is just a waste of time and effort and FPS to produce or import anything.
This is literally true - but to be fair, everything in a sandbox game is a 'waste of time and effort'. Dwarf Fortress is effectively a toy for grown-ups, and it would be unfair to understate the value of play when it comes to big monuments or similar. Of course, this is value to the player, rather than within the game itself.
In order to add the latter, we could consider the social structure that DF is attempting to emulate and consider how that could cash out within the context of comparative advantage. Nobles currently undertake an interesting form of rent-seeking which places heavy demands on the player and relatively smaller demands upon the fortress's economy; a couple of extra ballista parts won't hurt a fortress's supplies overmuch, but may be a faff for the player. There is very little reason for any dwarf to service nobles' demands at all, beyond the slight risk of catching a stray punch during a tantrum.
If we could ramp up their rent-seeking behavior in some way, we could increase demand for goods and hence the scope for the economy, in advance of any consumerist-driven change to the game. There are various possibilities, such as nobles now demanding a hoard (i.e. a room full of chests in which they store a percentage of the fortress's output, which can't be used by anyone else), which is pretty dwarfy. They could select retainers - dwarves whose only responsibility it is to look after the noble's needs, fetching and carrying and so on - thus diminishing the workforce.
Of course, you'd need a cost to
not supporting nobles in this way to make it work, and the obvious one is to assume that nobles get their rent-seeking privileges from their traditional feudal function: organising the military. If only nobles can appoint militia commanders ("Urist McPriceyPants has knighted Urist HoleyPants"), then the player is forced to meet their whims unless they resort to exclusively using traps. This would go part of the way towards raising demand; the richer the fortress, the more nobles you can attract.
Basically, the work that has to be done to make any of this make any sense is more monumental than most people realize.
This is true. But getting it right would provide a free-of-charge economic simulator to everyone, which is worthwhile even outwith a fun game.