Adrift! Behind the Scenes

I’m posting this How-I-Did-It and Spoiler guide on my blog instead of on the main forums to prevent unnecessary spoilers for the general public.  If you found it by following one of my breadcrumbs left on the interwebs, congrats!

How I Made Adrift!

First, the tools and what they do.  All of these are third-party applications which modify game data.  I don’t use game behavior mods, primarily because it’s hard enough to get people to put the downloaded save game in the right spot, let alone install a mod.

Invedit:  A simple inventory editor.  Allows you to place or remove items in your inventory, including things that cannot be natively generated inside Minecraft (like a grass block or a spawner).

Loledit:  I don’t know how long this one will work as the map portion is no longer being developed, but I use it for two things:  changing mob spawner characteristics and setting the game time.  It also can be used as an inventory editor.

MCEdit:  The key interactive game-map tool.  The controls take some time to learn, but it gives you powerful search-and-replace editing of blocks on a map.  Primarily used for changing the map, moving the spawn point, and moving the player location.

Cartograph:  Not a building tool per se, but a nice mapper that may help create interesting images for promotion.

Step One: World Generation

You need a raw map to work with when you’re making a challenge map.  If you want a minimalist map, you can make use of pre-made maps like the flatgrass map.  I wanted naturally-generated caves and landscapes instead of having to make them from scratch.  I started a new game in Minecraft, then did a short circuit around my spawn point to force new chunks to be simulated.

Next, I opened MCedit and placed a stone box about the size of the “play zone” that I wanted in the center of the map, up high enough to not intersect with the terrain in any way.  I did this to give myself a visual reference for the important second step in terrain generation.

Back in Minecraft, I proceeded to do a large circuit around my world with the maximum viewing distance setting.  I kept an eye on the reference box I’d placed in the world, attempting to keep it just on the edge of my viewable distance.  The result is a sufficiently large world that shouldn’t show discontinuities in the terrain, so long as your players stay inside the play zone.

Now that I had established my map, it was time to do…

Step Two: Large Scale Terrain Modifications

All of this is done with MCEdit.  Some of these will take a long time to resolve and re-light so plan on several hours.  Unless I indicate otherwise, each of these edits was applied to all the horizontal dimensions of the map.  To make an ocean map like mine, I made the following changes:

1. Deleted the reference box I placed to help get the right amount of terrain generation
2. Deleted all solid blocks located more than five blocks above sea-level.  I used this elevation instead of sea-level to give a bit more variety in elevation.  I still ended up with vast flat zones.
3. Searched and removed all logs, leaves, pumpkins, sugarcane, etc, that might have still existed from sea level and up.
4. Searched and replaced all grass and dirt blocks with stone from elevations -20 blocks to +5 blocks above sea-level.  I didn’t want dirt on the ocean-bed or in the shallow 1/3 of the underground environment.
5. Wanting a bit more clay on the map, I replaced a portion of the top solid layer stone with clay.  Because it wound be underwater, it wasn’t going be obvious that it was mostly a big square.
6. Last I added water depth.  Approximately 20 vertical blocks of water represents a one-way trip for a Minecraft avatar, preventing people from working on the ocean bottom without danger.  I replaced all air from sealevel to +20 blocks with static water.  In shallowest parts, one can just barely touch bottom and resurface without suffocating.
7. To make life easier, I placed a small square of solid blocks on the surface and moved the spawn point to it, since the next phase required in-game work.

Step Three:  Moderate-size Map Modifications

The next major phase was the subterranean renovations.  I have four distinct underground areas:

• The Lava Tube:  A cave filled with lava that stretches from the ocean bed to near the bottom of the map.  First I manually seeded the entire length of the cave with about 2 stacks of diamonds.  Next, I used MCEdit to replace air with lava.  This has to be done in small sections, in order to prevent overlap with other caves in the selection zone.  Last I tinkered with the top of the cave, which was originally open to the water so that it was mostly roofed-over.
• The Monster Cave:  A long top-to-bottom cave that contains approximately 8 spawners.  Placed the mob spawners manually in game, digging out spots in a few places so that I could have a second spawner located above a visible spawner that would drop additional enemies on the unwary player.  Near the end of the project, used Loleditor to change the placed spawners to their final monster assignments.  Near the bottom I dug the remaining distance down to adminium and installed a nether portal.  Also plugged a few cases of lava bleed-through from the lava tube.  Plugged the top of the cave in with clay.
• The Mushroom Cave: A medium-to-small cave that starts near the ocean floor.  Clayed-in the entrance and covered every inside surface with mushrooms.  Easy.
• Benson’s Office:  A very small sandstone structure sticking out of one of the deepest parts of the ocean.  The only underground area cut entirely out of solid rock, I used McEdit to create a working space, then did the rest in-game.

Throughout all of this, I continued to use McEdit to move the player position around.  This prevented me from leaving tell-tale signs of excavation to access the sealed caves.  Used Invedit for all of the placeable items.

Step Five:  The Ship

I’ve written about this elsewhere.  Several players have commented on the nice design of the ship, but I think it merely looks adequate.  I used Google image search to look at cut-away views of galleons.  I observed that the keel is almost exactly shaped like a straight line with a quarter circle on the end.

Just like any good shipbuilder, I made my boat on dry land and then transfered it to water with MCEdit.  I started a second save and flattened a bit of beach before setting down a keel.  After laying the keel, I then put in a single-block width section for the widest berth of the boat, trying to get a proper sense of proportion before committing to the rest of the shape.

From there it was an exercise in eyeballing the curvature from fore to mid-section.  Decks went in next, once I’d decided on the appropriate water level.  Fore-and Aft-castles came next.  Then I had the genius idea of using dispensers to make “cannons” on the cannon deck.  Last I added the masts, opting for three main sail masts and one lateen sale (the aft-most mast).  The foremost mast should probably have been build a block or two further forward.

Made sure I had a way to re-board the boat, so dropped ladders down the outside in the most natural-looking position I could find.  From there it was accessories:  chests and goodies within, stairs, ladders, that single precious square of grass, etc.

At the end I had to go back through and check all my signs to make sure my adventurer’s log links were correct.  Moved the spawn point, cleared the inventory, and then used Loledit to set the time to start of the day.  Finally, I had a testable map.

Step Six:  Play Testing

Probably the most essential part of the process is when you set aside time to test it yourself.  This allows you to avoid the most embarrassing mistakes and identify flawed concepts.

In my case, I found that it’s hard to avoid the monster cave if you mine directly below the ship.  I decided that was okay, the extra challenge was worthwhile.

I also found my sign for the top of the lava tube had caught fire and disappeared.  An easy fix, but one that could have been game-breaking.

Once I convinced myself that there wasn’t anything on the map that would make me ragequit in frustration, it was ready to turn it loose.