The Early Problems (Loc Part 3)

So, we were excited to begin working and gave ourselves a rough release window of the end of August 2011. Originally we intended to do 99 levels for Loc, the normal 1 – 6 sided puzzles, but some additional mechanics we dreamed up that changed the rules in new and interesting ways.

I would go to work, spend all day fixing other people’s problems, then return home crank on levels for Loc, fall asleep and repeat. It was a loooooong summer. During that time Mike headed up the coding on the huge horrible task of dealing with the Loc Solve algorithm while Matt attempted to get Tile Movement down.

We didn’t realize how long just those two functions of the game would take to get setup. Solve was a continual nightmare. It would have been so easy, so unbelievably easy if each puzzle only had one solution. But the elegance of the design was that each player could potentially solve the puzzle in their own way. Which is incredibility rewarding, both from a design stand point and from a player’s. Every path movement needed to be accounted for every face with every conceivable tile.

This is what he had to deal with…

The solve algorithm didn’t come together as a 100% working piece until late March. At the end of everything Mike had worked on Solve, on and off, for nearly eight months, rewriting it three times. The completed thing is somewhere in the realm of 13,000 lines.

Matt’s challenge, though seemingly simple was also dangerously complex. Early on we wanted to click and drag tiles, just like you could in any tile sliding, traffic style game. The thing was, every single one of those games that had only done so on a flat two-dimensional surface. A few of them made it even simpler by only having a single space available.

But again, this freedom of having a different number of tiles of different faces was important for the design. One face could have 6 tiles on it, and another could only have 1! It made for interesting, brain tingling challenges as well as keeping every level, even within the same section, fresh and exciting. Matt, got a rough working copy finished early on that we could at least test with. However, tile sliding was really finicky. The camera needed to be rotated dead onto a face in order for the mouse click to register and for the tile to be moved.

During Winter break he ended up polishing it a bit, which allowed for a greater range of the camera angle to move tiles, but it wasn’t again until March when he somehow created a miracle and tiles no longer needed to stop before changing directions.