That's interesting. I can't help you with the 1e levelling, but I can tell you that 3.0 (not sure about 3.5) aimed to have the PCs level after every ~14th encounter. That was based on their research on how fast the players and the DMs wanted to go.
The 4e 8 encounters per level thing was also alleged to have been based on such research. But they may have been talking about a slightly different market. My impression is that WotC at that time was leaning heavily into organized play, like Living Forgotten Realms.
It would not surprise me if people who were playing with strangers half the time, in a very structured campaign with no time to spend on anything that wasn't directly contemplated by the modules, needed fast advancement to stay engaged. Because, whatever people may say about LFR, you will have a hard time convincing me that those modules were engaging in their own right.
I actually prefer slower advancement, mostly because there are so many interesting things that can be done in a given level of play, that I don't know why you would rush your way through them. But I definitely have players who would get cranky if I slowed it down. I'm hoping with my new group (if I ever get that going), I can keep it slower from the start.
I'm also hoping that my training rules will provide a steady drip of player facing goodies that keeps them engaged even if they level more slowly. I don't know if I've talked about this before. When a PC in my game gets enough XPs for a new level, he gets the basic stuff - purely mathematical things like bumps to attack bonus and more hit points - after a decent night's sleep. But if he wants a new feat or power, he is going to have to train for it. I don't make it hard to find people you can train with, at least for common classes, but it does mean that the purely additive parts of leveling and the features part of leveling can happen at different times.
I'm very curious if you've got an example queued up for a module that actually does overland travel well (that is to say, an example module where travel is not the entire adventure, like it is in something like Isle of Dread). I'm at a loss to think of one; even old school modules like Assassin's Knot were usually all abbreviated like "It is 2 days travel from Restenford to Garotten. When the party arrives...", and maybe at most you could get a random encounter table (usually a painfully generic one with entries like "2d4 goblins" and "1d3 giant spiders").
B2 has a small but workable map. N1 has a very small hex map, with different random encounters in each region. ToEE's map is a bit bigger, but again has regional variations in challenges; whereas I think T1 suffers from having no regional map. U2 and U3 both have overland components. I think S4 (Tsojcanth) and WG4 (Tharizdun) theoretically share an overland map, although the presentation is different in each module.
However, none of these deal with long distances. So no, I can't point to a module that exemplifies what I am discussing above. Sometimes you just need to come up with a new procedure. I think I would just eyeball the distance, terrain, weather, likely inhabitants, and the precautions taken by the players, and come up with an ad hoc assessment of frequency and danger level of random encounters, as well as how they would be affected by local colour. I would probably use the 1e DMG as a baseline. And it might not actually require significant changes in procedure, so long as you are aware of the problem when building your encounters/encounter tables.
For long distances, I think I would reduce the number of encounter checks for a given distance, based on the unit of time we are talking about. If travel is measured in days, it would be checks per day. But if travel is measured in weeks, it would be checks per week, and probably more significant encounters, because of the amount of time to rest between encounters makes every encounter the only one that day (breaking from the attrition assumptions of D&D), and because I would be handwaving more frequent but lesser encounters as being something that occurs but is not worth gaming.
As for making a system for publication, that would take more playtesting than I have time/inclination for, but then nobody is paying me to design or play D&D modules.