One of the exciting things about dungeoneering is opening a door into the unknown. The potential for death and treasure is clear, and maybe there will be some curious or exploitable feature. Maybe on the other side is a portal to hell, a shaft descending in darkness to the 6th level of the dungeon, or the hall of the mountain king. The gradual construction of a player map, which can be exploited to circumvent lairs and reach desired destinations with minimal risk, is a satisfying activity. In OSR systems where there is not much system-mastery to be had, world-mastery can fill the same role, rewarding players for their attention to detail and their skill. Additionally, when players keep their own map, they necessarily end up a little uncertain about it, and this adds to tension.
Given that this works well in the dungeon, why don't we do it in the wilderness? Notably, a number of very successful wilderness adventures and settings actually have given players a blank map and said "have fun filling this in". The original Wilderlands of High Fantasy did this way back in the 80s. The player map was a central feature of the widely-imitated Western Marches campaign, and Pathfinder's popular Kingmaker adventure path did the same. In these cases, the hexcrawl is actually a crawl, with decisions along the way forced by lack of information. By contrast, in my attempts at wilderness campaigning with a high-detail ground-truth hexmap, typically the party looks at the map, calculates an optimal route, deals with any random encounters along the way, and arrives at their destination. Rations are laid in with perfect accuracy, so logistics are almost never an issue; hunting and foraging never occur, and resting only results from forced march. The travel becomes interactive, nonlinear, only when the party fails a navigation roll (an exceedingly unlikely occurrence when you have two characters with Navigation, although re-reading the rules it looks like multiple characters with Navigation don't stack, but Navigation and Explorer do stack so it works out about the same), and the satisfaction of exploration, the joy of discovery, never occurs.
Alexis very accurately expresses this major source of discontent with my last campaign in his book:
A mistake I made many times as a young DM was in recognizing, unfortunately, a common pattern. If I made the party aware of a place that contained something valuable, they would, most times, unquestioningly march out and get it. My campaigns rushed from target to target, and as I assigned each target, the next sessions would become lamentably predictable. I began to feel trapped in my own campaigns. After a time this became quite dull... Parties are more interested in and involved with targets they cannot see... If the party is told where to go, they will go there. But if the party is told only that the destination is 'nearby', they will go everywhere.There is also an anachronism argument to be made, that accurate mapping and surveying techniques were not developed until the early modern period to assist the developing European states in clearly delineating their territory into legible administrative districts. I think the realism concern is less important than the excuse to drop stylized, dubiously-accurate maps with crazy sea monsters on my players, with hopefully hilarious results.
On the flip side, running wilderness navigation is hard. You do not have the structure of the dungeon, and your extensive experience in indoor spaces, to help you. My earlier interest in mountain peaks and smoke visibility were part of an attempt to build usable language and conventions for wilderness exploration play (in the way that "You open the door, and see a 30x30 room with a door in the opposite wall, containing five gnolls" is clear and actionable. "You are in gently-rolling forested hills, on a narrow trail; from a treetop, you can see a mountain range to your west and multiple columns of smoke to the south. The sun is low in the sky to the west, nearly touching the peaks" is a little less useful; it does not present an obvious menu, other than "make camp"), but I recognize now that without concealment, without prolonging wilderness travel and making it interactive, there can be no fun in exploration and such conventions and language are useless.
On further reflection, I guess what I'm really aiming to do is to turn the wilderness into the megamost dungeon. The megadungeon structure is sort of dumb and deeply unrealistic, but it works. A lot of the things that make good megadungeons work really well (multiple paths, varying danger levels, factions, restocking, path reuse, too-big-to-clear) actually make more sense or are simply unavoidable in the wilderness, but for the most part we've (well, I've at least) failed to put it all together properly. I've failed to see the dungeon for the trees.