Code:
# Factors to modify for technological development.
technology = {
	endgoal = { ... }			#list of endgoals to achieve.
	preference = { ... }		# list of techs to give bonus too.
	ignore = { ... } 			# Techs we will never research

	#preference to weight importance of research in these field.s
	armor = 2.0
	infantry = 3.0
	industry = 2.5
	aircraft = 0.6
	naval = 0.1
	land_doctrines = 4.0
	air_doctrines = 0.9
	naval_doctrines = 0.1
	secret_weapons = 0.4
}
This is an excerpt from ai_file_doc.txt. The preference to weight importance of research in these fields does not work as it should as techs in progress are ignored. The best example is if a nation has no landdoctrine, researches one, gets a new techslot and will try everything to use it for landdoctrine aswell, usually resulting in hospital techs researched by CHI. This special case is solved by some changes but the basic problem remains. Depending on the number of slots available the ai will try research many tech from one field at once, often more than proper tech team are available.