LARPing's not my bag, but I am curious about the module building and how it relates to 'traditional' module building.
I don't believe that the main issues I discussed in my previous post (about controlling party size and staging repeaters) matter... really at all to a tabletop module writer. In a tabletop game, 100% of the party usually goes on the module, and they're certainly not going to proceed through the content more than once in anything but outright Groundhog Day plot. The really staggering difference between the adventure writing for LARPs and tabletop games is that LARP modules really don't want to go much past 3-4 encounters, or one running battle with a lot of small encounters. It's just too draining on NPC resources. A tabletop adventure of 3-4 encounters would generally be considered pretty short - enough for one session, or two if your sessions are short. In my experience, tabletop gaming doesn't make heavy use of breaking encounters into multiple waves - something that can be hugely beneficial if the PCs are deeply enamored of alpha-striking or AOEs.
LARP modules also tend to be smaller because the physical locations are smaller, and there are only so many ways you can explore one 40x40 building in a single setup. Of course, some sites are better for module buildings that others - I can't say enough good about how various campaigns have used Indian Springs' barracks buildings, which are large, heated, and have a lot of sturdy walls and posts to anchor tarp and wires. Other sites have pavilions - less in the way of internal anchor points, better for open fighting space. Eclipse and DtD have gotten creative with constructing additional anchoring emplacements, using free-standing PVC pipe structures. Setup takes time and effort in a way that a DM's description doesn't, so having an awesome new idea on the fly is more like l'esprit d'escalier than an opportunity for a new twist. (Sometimes.)
The big lesson that tabletop games can learn from LARPs is that padding an adventure out with fights isn't worthwhile. (I don't know your gaming tastes personally, Matt - this is scattershot commentary on published modules that I've read.) You can have a great module with no fights at all, if you can give the players a clear feeling of both player skill and character skill bringing about victory.
LARP adventures have an advantage of immediacy and tactility. This isn't about how one gaming medium is better than another - they have their strengths and weaknesses, though. Among tabletop games, 4e D&D and Fate are probably the two best for making space and terrain a part of the game the players can really engage and manipulate. LARPing does this intrinsically, though Plot committees are well-served to consider "amplifying" the terrain by adding more features that combatants can manipulate.
Side note: Most of the really obvious uses of modifiable terrain involve moving or upending tables. Safety secondfirst, y'all. Marking areas on the ground, or using clearly pre-defined areas (such as "everything inside the building") is smart - though I'm pretty sure "To the Room" as a targeting tagline is pushing 20 years old now, which makes it one of the oldest tricks in the book that doesn't get all that much use.
The lesson that I would offer to both tabletop and LARP adventure writers is to keep changing it up. There are times when an adventure can involve a string of fights and that's all you need. Mostly, though, a certain amount of variety is everything. As a general rule, avoid using the same kind of encounter twice in a row. The boundaries between one encounter and the next are sometimes hazier in LARPing - for example, the difference between one large trap encounter and two consecutive trap encounters is academic.
Trap encounters: A room crammed with traps is a lot more fun in a LARP than in a tabletop game, because you only bring the trappers (and maybe one dedicated healer) on such an adventure. In a tabletop game, it usually means everyone else catches forty winks while the one rogue in the party rolls a bunch of dice. What I'm saying here is that tabletop games need to get better at making traps interesting at the table - 4e did a bunch of pontificating about this, but didn't fix the problem.
Another of the big differences between LARP and tabletop modules is scope. In tabletop, a published module might be a single dungeon suited to several evenings of play, or several maps (none of which quite live up to the name of dungeon) with encounters and a closed narrative arc. That is, the questions raised at the beginning of the module are usually resolved by the end; if not, you're probably looking at an Adventure Path. When LARP runners talk about "a module" or "an adventure" (some communities get very humpty about the former term), they're talking about the span of time that the characters are out of the main area of gameplay, as well as travel time and time spent setting the module hook. There is game content relevant to a module's narrative arc before the module begins and after it ends - modules advance a plot, but it's relatively rare that they are the whole plot in itself.
There are whole kinds of adventures that one very nearly cannot stage in the kind of boffer LARPing that is found in the US. (Make a statement about "you can't do X in a LARP," and the Nordic types will move heaven and earth to start a new game that does that, just to be contrary.) The travelogue-style adventure doesn't work well in a game where the players are demonstrably in the same location for 36 hours, and are probably in about the same location for the next event. Sure, you can rearrange the event site to make it seem like a different place - every game does this, and should. But one cannot simply LARP a fourteen-month hard march to Mordor on a state park in middle Georgia. (If you are a New Zealand LARPer reading this, you have my full permission and blessing to snicker derisively. Also, you suck.)
There are now a few direct examples of classic tabletop modules translated to another medium. The Temple of Elemental Evil has been available as a tabletop module since 1985, and a computer game since 2003. Likewise, DDO has just announced that they'll be releasing The Haunted Halls of Eveningstar - which has the extraordinary honor of being the only published module I've ever run, in nearly 21 years of D&D. My point here is that there can't help but be a lot of changes from that come with a shift in medium. In terms of the verbs available to players, video games are unavoidably restrictive, while tabletop is limiting in idea (depending on rule set and group dynamic, players may not think to pursue means that aren't spelled out on their character sheets) and LARP is limiting in feasibility (LARPs generally don't let players solve problems by breaking windows, burning down buildings, or doing any of the things that violate safety).
This is a really digressive way of saying that the things you have to plan for as possible player solutions vary widely by medium. The glory of tabletop is that there's no visual to worry about rendering, so the players can try anything that the group's social contract and the GM's patience allow. LARPs don't allow all of that, but there's an incredible wealth of emergent detail that flows out of the physical staging. On the gripping hand, imagine a tabletop game in which there could only be about half a dozen maps. You can add walls and other terrain features, at some cost in time and labor, but there are also walls and other terrain features that cannot ever be removed. Oh, and it takes a lot of work to make creatures be anything other than medium-sized bipeds or quadrupeds (with an exception made for spiders - I think the spider costumes CI/Ro3 has been using since the dawn of time are surprisingly good).
I may be reiterating the same idea a lot here. I do want to point out that the traditions and styles I'm describing for CI/Ro3 games are far from the only things out there, though in the US they are predominant - CI/Ro3 has more in common with NERO and its lineage than IFGS. (Matt - I'm well aware that this is probably obscure jargon to you - NERO and IFGS represent radically different approaches to the basic game loop.) I do have some experience with the latter, though - it's how I was first introduced to LARPing. IFGS does a number of things that make it a lot more like D&D than NERO is: modules are published and are often run more than once, in different parts of the country; the story focuses on the adventuring party rather than the playerbase as a whole; there's very little roleplay that doesn't occur in the context of a module. (Also, this is not criticism of their style. They were the ones who first got me hooked on this hobby, so they must be doing something right.) Without extensive interstitial content, though, it changes how you think about setting, character, and story. It is to counter this exact point that the 3.x-era Adventure Paths published in DungeonMagazine had extensive setting detail. IFGS, and a D&D campaign made entirely of modules, have a lot more in common with a TV procedural than a serial drama. CI/Ro3, and I think the whole NERO lineage of LARPs, still have a few predictable elements in their event's format, but no more than serial dramas conform to Freytag's Pyramid.
Wow, this answer is all over the damn place. I hope I've said a few useful words about module-writing in the two different media.