Donjon Design: More than Just a Cellar Full of Monsters

So after a six month hiatus, I'm getting ready to dive back into running Fourth Edition D&D, albeit on a limited basis.  I'm taking over one of Citizen Ben's three D&D games for a brief spell and I'm also planning a massive weekend long mega-dungeon in celebration of the 30th anniversary of my escape from the womb...   As as I get together my notes and actually start the design process, I've started keeping a list of goals for how I want to design these two dungeons. **Note, for the purpose of this post, consider Dungeon to be synonymous with Adventure.**

(brainstorming for Iskender's Tomb)

Here, then, are the tenets I've decided so far are worth shooting for:

• Design for Interesting Encounters, Instead of Merely Challenging Ones
• Hard ≠ Fun, Challenge should be Dynamic
• This is a big GM trap that I know I've fallen for many a time, the notion that every encounter needs to be a CHALLENGE, that every battle should be Life and Death and WORLD SHAKING... Okay, okay, that might be a little hyperbolic, but I do think there is a tendency among GMs, even experienced GMs, to try and make combats Challenging instead of Interesting.  Why?  I think part of it is that we fall into the false assumption that "difficult = interesting" and another part is that it's easy.  It's easy to build encounters where the players and PCs will find the going tough, it's a little harder, sometimes it can take a little bit of outside the box thinking to make an interesting encounter that goes beyond merely difficult.
• If every combat is a struggle just to stay alive, it's very possible that they all start to blend together.  It's hard to keep that time you almost (or did) die in a fiery volley of arrows from well organized and tactically disciplined goblins from that time you were almost bludgeoned to death by the clubs of well organized and tactically disciplined ogres from the time you were impaled on the spears of well organized and tactically disciplined hobgoblins.  I think those three combats would be much more memorable if the goblin archers broke once threatened with melee combat, fleeing to safety and then attempting hit and run sniping and guerrilla tactics, if the ogres waded into the PCs without any sense of tactics, swinging wildly, attacking the last thing that hit them without any pretense of doing the tactically effective thing and the hobgoblins, well the hobgoblins should be the tactically disciplined foe, where maybe the PCs need to take a different tact and try to maneuver the hobgoblin  phalanx into a place where they can use the rigid and disciplined order of the hobgoblins against them.
• Make no mistake, I'm not saying at all that every encounter should be a breeze.  There needs to be the gritty fight to the death, but if every encounter is knock-down, drag out, barely make it out alive (or maybe not make it out alive) battle, the game starts to feel, at least to me, like a meat grinder.  Maybe the PCs haven't fallen into the grinder yet, but that falls starts to just feel inevitable.
• That's why, with both the dungeons I'm planning, I want to make sure to include some interesting, but not traditionally difficult, encounters.  I've got several ideas of how to do this, my favorite right now is a grand battle against nearly countless lesser foes.  After all, I think you'd be hard pressed to find a D&D player who wouldn't enjoy having a chance to take their PC and wade into a scrum of orcs or skeletons and have the satisfaction of hitting with nearly every attack and killing with nearly every blow.  Of course, the little cuts and bruises they'd take would add up, keeping the fight from being a complete breeze without any consequence and of course no DM would be able to resist having that moment in the midst of the melee where a PC disappeared under a wave of mooks, only to have the chance to burst out ala the Hulk or to have the other PCs have to wade in as well to rescue their too bold comrade from a death by a thousand tiny goblin knives.
• Give Every Room or Encounter a Hook
• Ever watch a movie or TV show and start getting bored as the same scene plays over and over again, without really advancing the plot and with only a pretense to differentiate it from the others?. "Oh, here we are, another spaceship battle against the Grays, oh, but I guess this one is above a moon instead of above a gas giant.  Wah-wah..."  Poor dungeon design often falls into this.  How many rooms can be filled with goblins before the players are bored?  "Oooh, another room full of kobolds, but this one, I guess, has several raised platforms."  Let's be honest, nobody remembers that one room that had the raised platforms, or those ruins that were full of lizardfolk, if there's nothing to distinguish one set of platforms or ruins from another.  
• Another common problem is a sense that an encounter just delays the PCs from their ultimate goal without adding anything else.  Is there anything more boring than a hallway?  Adding a half dozen fire beetles helps a little, but really, a month from now, are the players going to remember that fire beetle hallway?  Are they going to remember those kobolds in the forest clearing?  Probably not, sirs and madams, probably not.  
• What those encounters need is a hook, a logline, a one sentence summary that the players may never know but that they could probably come up with on their own if they were describing the dungeon to a friend who hadn't played.  Like: ruins with lizardmen who ambush the PCs by pushing teetering columns over... "And then we came to some crumbling ruins full of lizardmen, and the lizardmen started pushing over columns on us, but soon we turned the tables on them and pushed a column onto their chief."  
• Just like nobody wants to watch a TV show where one scene can barely be distinguished from the next, the players will soon be bored if each successive room features yet another group of zombies and maybe this one has a pit trap and the last one had a brazier.  It's not enough to fill a room or ruins with full of terrain, traps and monsters, you really need to figure out how those interact with each other and the PCs.  It seems like a simple thing, an obvious thing, a thing that could easily be skipped, but taking the time to decide that the pit trap is well hidden and if a PC falls in, several zombies will fall in with them or placing the brazier on a pedestal in the center of the room with zombies behind it so that an enterprising PC could push it over to set said zombies on fire is better.  Even better is to have the zombies have enough hit points so that for a few glorious rounds the PCs get to fight flaming zombies, because let's face it, "Flaming Zombie Room" is way cooler than "Room with a Brazier We Had To Walk Around to Fight Zombies"
• Now, some of you are probably going, wait, what about Dungeon Ecology, shouldn't things make sense, you can't just fill a dungeon with random cool things because you feel like it.  Well, no, you probably can't get away with doing that too much unless you have a group of players who only want to fight things and don't care about any semblance of plot or story, but even if you decide that you'll make a set of mines infested by kobolds there's plenty of ways you could take that and turn it from 7 encounters in variously sized rooms full of kobolds and maybe kobold allies into something fun with seven memorable rooms.  Maybe I will make that dungeon someday.  Well, probably I will make that dungeon someday, but not today.
• The Hook Needs to Draw them In (Or at least be present at the Start) or No Cheating! The Hook can't be What Is Waiting for After the Encounter
• The hook for the encounter can't be that once they beat generic goblin group in generic ruins that they get to explore the ruins.  The hook can't be that once they beat the ogres they learn about where the kidnapped noble is.  Much more interesting is a fight with goblins who do all they can to keep the PCs outside of the ruins but who refuse to enter any part of the ruins themselves.  The savvy player will recognize that something is going on with those ruins.  Likewise, just having the PCs find a clue to the whereabouts of the kidnapped noble after fighting a group Ogres is much lamer than having one of the nobles attendants swinging in a cage above a fire as the fight goes on, leaving the players to decide if perhaps they might need to save that NPC in the midst of the fight so that they can learn what he might know (also, so they can continue to pretend to be good guys).
• Design for the Players and PCs, not the DM, Monsters and Obstacles
• But It's So Tempting...
• It's so tempting, as a DM, to just design the dungeon that you want to play, but after designing many a dungeon that I want to play, I think it's more effective and enjoyable for all involved to design dungeons that I want the players and PCs to play.  It might seem like a small distinction, but I think it's important.  It's easy to write up a dungeon that is full of things that are interesting to me, to put monsters and traps and terrain together in such a way that it amuses me, but if you don't design with the players and their PCs in mind you might as well be masturbating onto a sheet of graph paper.  Design dungeons that will be fun to play when the players and their PCs explore them instead of dungeons that just tickle your own particular fantasies.  If you're thinking to yourself, "I'll really enjoy having the players trapped between fight three red dragons at once in a room full of pit traps" maybe you need to think about what the players might enjoy about that encounter.  I'm not saying they might not enjoy it, maybe all the PCs hate red dragons and have good athletics/acrobatics/thievery that will keep the many pit traps from becoming overwhelming frustrations, maybe that could be an enjoyable encounter, but if the PCs are all built around being scholarly undead hunters and their skills lie more in the arcana/history/religion realm, maybe a battle against 3 red dragons in a room full of pit traps will just feel like the DM being a jackass and enjoying himself at their expense.
• Build in Moments for PCs to Shine
• This is one of the easiest ways to increase player involvement and enjoyment and frankly I'm surprised at how overlooked it is as a DM tool.  Does one of the PCs have a crazy high resistance to Lightning?  Why not have a lightning trap that they can walk through with impunity to go toe to toe with an enemy that the rest of the group might have to attack across a chasm or a horde of minions.  Does one character have a ridiculously high Religion check bonus? Present an encounter where quickly determining the correct demonic idol to push over helps disrupt the summoning ritual the baddies are trying to complete. Have a character that specializes in ranged attacks, have a battle with a chasm in the middle of the area that allows them to pick of the goblins crossing the rope bridge as their compatriots cross (apparently I like using chasms as examples).
• A good player will always watch for opportunities that randomly come up for their PC to shine and a great player will actively work for opportunities for their PC to shine, but really, a good DM should be providing opportunities for each PC to take advantage of things that character excels at or otherwise is uniquely suited to handle, because, really, wouldn't it suck to finish playing in a campaign, look at your character sheet and say, "Huh, I had 5 pts of Resistance to Radiant damage and didn't leave any trace when moving through mountainous environments... Too bad we never fought anything that dealt radiant damage and we spent the whole campaign in marshy wetlands."  
• Don't Build Challenges without Considering PC Sheets and Player Options
• Sure that chasm the PCs need to jump across sounds like a great idea in the abstract, but don't forget to look at their sheets to make sure that its not an insurmountable obstacle.  If you decide they need a 20+ to make the jump, maybe you need to consider how the wizard with a +3 to his check and the cleric with a +4 to her check are going to make it across.   If none of the PCs have a Will defense higher than 22, then maybe that Mind Flayer with a +16 to his Save Ends Domination attack is just asking for unhappy players as PC after PC is brainwashed with little hope of resistance.
• I've found that whenever I come up with a difficult to overcome obstacle, it's best to think up at least two ways, and ideally four or more ways, that the players could have their PCs surmount it.  If I can't think of more than a single way for the PCs to get around a chasm, then I know I need to add some options.  Maybe there's a dead tree on the other side that the first PC across and push over for the others to use as a bridge.  
• Monsters Aren't the DM's PCs, Don't have Them Behave Like PCs.
• One of the things you give up on when you get behind the DM's screen, even if it's only a figurative screen, if playing your own characters.  Sure you can run excellent NPCs, but it's a terrible pitfall if you start to think of them as "your" characters as that makes it hard to make sure they're servicing the story, and ultimately, the enjoyment of the players.  Likewise with monsters in combat, they shouldn't be "your" characters.  You're not on the Orcs side battling against the Players.  You're running the orcs the same way you're running the story, not to beat the players but to collaborate with them on a fun story.   I've made the mistake before of having the monsters be every bit as tactically devious as I can be when I'm playing a PC and it was a disservice to the players I subjected to that combat.  Yes, the monsters should have tactics, but if you're designing/running encounters where you make sure the monsters make all the tactically smart moves I think you're forgetting that above all you're a referee.  Just as even the toughest video game bosses have a tactical weakness or a trick that a player can use to beat them, your monsters should have a weakness or tactical blind-spot that the PCs can take advantage of.
• Let the monsters behavior be distinctive and comprehensible to the PCs/Players.  Goblins should be cowardly, Ogres should be dumb, Gelatinous Cubes should be mindless, at least mostly, as then it's always fun to have goblin berserkers show up, or throw an Ogre with a really clever plan that the PCs need to figure out or a Gelatinous Cube that's being controlled psychically by some intelligent enemy, but if all the monsters behave in combat just the way they would if you were running them as PCs that you wanted to "win" then you're robbing yourself of those opportunities.

I'm sure that I'll come up with more design tenets/goals for these two dungeons, but those are probably plenty for me keep in mind as I start my planning...