DWeird nails a lot of things on the head that I've been thinking about. I largely agree with this suggestion.
In my opinion (keep in mind I'm no historian or anthropologist) DWeird's suggestions would fit well with a model of technological progression akin to that of human civilizations as far as I understand it. I've been looking a tad more into the histories of various technologies, particularly those dealing with warfare-related technologies. From my forays into the subject matter, I have come to the conclusion that the development and implementation of technologies is not quite as much a matter of some serendipitous "aha!" moment that happens at a lab, forge, bakery, apothecary, or what-have-you (though those are still very important!!!) but more of a matter of economic development. In other words, the existence of the economic foundation required to implement a technology is what I believe to be the biggest determinant in the appearance of any technology throughout human history, and should be the case in dwarven/elven/gobbo/etc. histories.
A great example of this is steel, something I see that people are very keen on dwarves having access to. Steely irons have been known about and made since iron was able to be worked, which is a daaaaaaaaaaaaaaaaaaaaaaaaarn long time (since around 2000BC or thereabouts...). However,
consistently making steel is something that took a while to develop. It wasn't a lack of knowledge of the existence of steel; blacksmiths since time immemorial knew that introducing a carbon source to a good, pure iron led to a higher quality iron (i.e. steel). The largest limitation to making steel was having the infrastructure to do it: control of mines and ore deposits, presence of bloomeries, good roads, government, etc. That's not to say that the advent of such devices as the Catalan forge and other devices capable of smelting steel efficiently weren't important, but these are small potatoes compared to the afore-mentioned limitations. Think about this as well; before they fell, the Romans could mass-produce steely irons of decent-enough quality for their huge armies on a scale not again attained until the late middle-ages centuries after their fall. People didn't forget the processes involved, at least I don't think so given some metallurgical and historical treatises I've read. What happened is that the infrastructure of the Romans disintegrated and with it the means to implement the technologies already discovered. Tools broke down faster than they could be replaced. This goes for other technologies besides steel, such as sanitation, construction, medicine, etc. I hope I'm making sense.
And if anyone thinks this example is bull-crap, feel free to call me out on it. I won't get butt-hurt or start some dumb flame-war.
I realize that there are a lot of things I've mentioned and will mention that can be torn apart because I am not taking the time to cite sources and I am writing this on the fly without as much careful thought as I would put in a publication. This is all just food for thought. Anyways...
So, basically, I feel DWeird's idea fits my examples nicely. Most ideas are already there or are quickly thought of in beginning of dwarven civilization in a generated DF civ. What are also present are the basic tools needed to eventually make better tools/economies/infrastructure, and, with those, even better tools/economies/infrastructure. So, basically, as civilizational economies and infrastructures increase in capacity, certain technologies become feasible to implement on a civilizational scale. Along the way, certain ideas come to some possessed/fey-mood/whatever dwarfs and they develop something like a novel concrete mixture or the afore-mentioned Catalan forge or whatnot.
So, yeah, no active tech research in the sense of selecting something from some sort of list. Techs largely arise through economic development... something DF can do awesomely and already does on a tiny scale within the fortress (i.e. building a wood burner and smelter before being able to get a metal-works in order).
I guess I can conclude from these musings that the biggest thing that needs to happen in DF in order for suggestions like mine and DWeird's to come into fruition is the implementation of a true supply-and-demand model in DF.