I was writing in an earlier post about converting Pathfinder modules to our older systems; not because I expect anyone to be doing that at this point but more because of the baseline assumptions about what constitutes old-style and should thus be converted.
For me, the parts of any module that don't inherently need conversion are the backstory about what the place is and why it's now dangerous, and the tactical set-up of the map/monsters.
On the other hand, the monsters themselves may need adjustment, as I discussed in my previous post. There are a couple of other factors involved, though, in adapting a modern module into the old-style rules.
First off is the level of characters for which the adventure is designed. There is a particular feel to an adventure -- killing a powerful, evil wizard works best as a higher level adventure because of its feel. Giving a brilliantly thought-out way for a lower level party to kill an evil wizard, even if it's a way that doesn't have a problem like a powerful NPC ally, STILL doesn't feel right if the adventure's concept doesn't feel low level.
I had this problem surface, by the way, with the "Tower of Mouths" adventure I wrote for Knockspell magazine. More than one of the readers told me that going into a totally abandoned wizard's tower in the middle of a city didn't feel like a really high level adventure (at least, not the way I'd written it). So for that adventure I ended up reducing the threat levels to toggle down the adventure -- not because of any problem with the adventure itself, but to make the situation match up with the character levels.
In Pathfinder and 3e, the game is designed to go to higher levels, which means that the whole bell curve where game-threat is matched to game-feel doesn't map directly from PF to OD&D or vice versa. It's close, but you have to watch for it, especially at the higher levels.
Thus, I tend to approach conversions as kind of a holistic process. I start with the monsters. I screw around with the methods I mentioned in the earlier approach: avoiding a monster group that is entirely leveled-up, toning down monsters that are too leveled-up relative to the "standard type," and making sure that several of the encounters involve monsters that offer different types of threats at one time.
The result, since I've been working more from feel than from numbers, is that a lot of these encounters will vary pretty widely in terms of how difficult they are. That actually kills two birds with one stone, because it removes the "game balance" which is generally advocated in more modern games. I want that gone anyway. However, I still have to make sure that the right ones are the difficult ones. In general (generally), a module should go from easier to harder as the adventurers move through it, assuming that they follow the most likely pathway. This definitely isn't a firm rule, but it's a rule I'd have to have some reason to break if I wanted to break it.
Which leads to the fact that, again, since I've been working from the feel rather than the numbers, the resulting set of encounters might or might not fit the character levels for which the adventure was designed in the modern system. I change -- if it's necessary -- the character level to fit the difficulty, not the difficulty to fit the character level.
Finally, there's another issue, which is that modern game systems (a) don't necessarily resolve traps based on player skill, using a die roll instead, and (b) traps and similar difficulties are often made more difficult to match the character level in a modern adventure.
Problem (a) isn't something I've really had to deal with, since the Necromancer Games authors are already on board with the idea that a trap or trick needs to be fully described. It's a poor author, even in a modern game, who just describes a trick or trap as a die roll. Thus far, I haven't run into it. Part (b) requires some changes. In general, even a high level old-style adventure contains only a couple of "more than usually difficult" traps, locks, or whatever. The whole point of being a thief is that you have a higher and higher chance to open locks. If the locks routinely get more difficult, your chance of opening them never changes. That sucks. So I remove lots of references to "difficult" inanimate challenges, and make them standard. I keep in mind, though, that it's cool to make some of these into unusual challenges - variation is the spice of an adventure, and the thieves should encounter some difficult stuff (and some easier stuff, but I tend to forget that). But the difficult stuff should be the exception rather than the rule.
Nevertheless, in several cases, I have to beef up, not necessarily the description of a trap's trigger and appearance, but I do generally have to add in the description of the clue-to-disarm and the disarming mechanism. Those tend to be glossed over in modern modules; again, in the Necromancer modules this has tended to be less of a problem than it would be if I were converting some of the other 3e material that I've seen. The Necromancer writing style tends to provide a lot of this description even though it's technically not necessary within the structure of the modern rules.
And then, there's often some description of skill checks in the modern modules. These I treat with a whole variety of changes, based usually on the tactical importance of the skill check. Some, I totally ignore as completely unimportant to the module. If there's a book in some foreign language that might or might not be readable based on a "knowledge check?" Who cares. Elves and magic users can read it. Or it's totally unreadable without a spell. Whatever. On the other hand, if climbing a wall should tactically involve a risk -- I might use either a saving throw or even (gasp) an ability check if I think character level ought to be irrelevant to the task. In the process of converting inanimate threats, I tend to focus (as I did with monsters) first on the tactical importance of the challenge (whether to use a check at all) and then on the description that's given (what sort of check to use, which is usually 1d6 or a saving throw), and then I tailor the game-threat to match the description (how hard is the check - is there a penalty or a bonus, or is it preferably just normal).
I guess that's it. But the interesting thing is what the process says about what old-style gaming actually is. It's probably not a complete list, but here are the things I work from:
a) Often monster encounters have numerous monsters compared to modern games
b) The differences between the types of threat offered in a mixed-foe encounter is more distinct; the threats are quite different in terms of tactics, and create bigger tactical decisions. There's not the subtlety of five-foot steps, attacks of opportunity, reach of weapons, etc., but the broader brushstrokes of the tactics -- the big tactical decisions -- are clearer.
c) Monsters tend to have less variation within a species in the older games. This fits again into (b) above. They move in squads (or couples, or individuals) that have considerably different tactics and capabilities from group to group than in a modern game where the monsters have more subtle gradations in their tactics and skills -- and more of the monster have these gradations instead of having squads of identical monsters.
d) Traps and tricks have relatively rich descriptions, so the DM can figure out what happens when the players need information and then start tinkering. That information is more necessary. This also isn't a problem in all modern game writing, even though it's there more for flavor than in the old-style approach.
e) Game balance less important - duh, we all know that one.
f) Character level needs to match up with the epic-ness or lack thereof of the module's theme. That's true of modern modules too, but the character levels don't map onto each other.
Whew - I don't think I'll need a part three on this one. I believe I covered it all. :)