Quadrilateral Cowboy

Command-line interfaces are less a mystic channel of communication with an operating system and more a compromise. Writing graphical interfaces is difficult and time-consuming, and for the average developer, it’s much easier to implement text-based commands than get mired in the widgets, frames, and event-driven programming that comes with GUI creation. However, a CLI is also enticing for how easy it is to enable cooperation between different applications using a common set of commands: piping, redirection, and a slew of helper programs all support a robust level of control for the user.

Quadrilateral Cowboy uses the command-line interface through a retro microcomputer (or “deck”) as its primary method of control for precisely this reason. The strength of a central interface provides familiarity when testing a new tool, and the “blink” executable opens up tool interaction and sequencing by binding scripts to certain numbers of blinks, which the player can activate at any time with a keybind. Blinking allows script execution while physically separated from the deck, creating the potential to manipulate your environment from afar while you manage other tasks. This, along with the ability to included timed pauses in-between commands, opens up a rich world of puzzle solutions, all enabled by some ingenuity and careful planning.

Those familiar with command-line interfaces will feel right at home with the DOS-inspired look-and-feel of the terminal. [src]

However, Quadrilateral Cowboy never requires you to leverage these capabilities. I would hesitate to call the game’s puzzles “bad;” rather, they’re too timid to ever require complex multitasking, and they’re too rigid to feel open-ended. Between the two primary tools, the weevil (a walking mini-quadruped) and the autocase (an aimable turret), there are no instances of the two interacting together nor are there instances where you utilize multiples of either tool, even though there’s functionality for each that makes it feasible. The weevil itself gets a criminally low amount of time to shine at all: there wasn’t a single instance past the first third of the game where I needed to actually move it, as it’s instead primarily used as a way to remotely control datajack interfaces for obstacles such as lasers and doors. As for the autocase, many of its puzzles solely require aiming the autocase at some interactable, iterating on commands that slightly adjust pitch or yaw. Once it’s in place and functional, the puzzle is solved. Occasional timing constraints will arise that the player needs to account for, but these never extend beyond single alarms or anything that would require intricate timing from a programming perspective.

Part of the fumble here comes from the rest of the toolkit, as these extraneous features rarely enrich the depth of a given mission and tack on additional tutorial time to a game too slim to support them. Out of the nine available missions, the first six are all effectively tutorials, which leaves the last three to extend and combine the previous scenarios. The tutorials themselves are well-designed, but they test so little beyond the basic functionality of each tool that it makes the difficulty on-ramp feel very flat. Beyond this, most of the tools lack the multi-purpose, granular use cases that the weevil and autocase hint at. The launchers (wind-powered launching pads) could have shined if there were relevant moving interactables or reasons to move your other tools (like bouncing the autocase and timing a shot at its apex), but they’re used exclusively for navigating to and from the starting platform outside of its tutorial. The “Enginani” and “Greaser” abilities are entirely for traversal and only exist for brief, unrestrained lock-and-key obstacles. One section in which you’re required to switch between Enginani and Greaser toys with timed cooperation with the duo, but their interaction is relatively minimal in these sections (you only switch twice in either job) and they completely ignore the CLI concept the rest of the game builds on.

The game repurposes id tech 4 into tight, clean characters and models that manage to evoke bits of ’80s cyberpunk without the indulgent neon lighting or hackneyed future tech. [src]

Once you get to the final two missions,¹ the game attempts some more involved puzzles. Unfortunately, both of these puzzles are a rigid series of in-game buttons one needs to press with small tool sections identical to the tutorials. The stock exchange job is virtually one single “aim the autocase at a button and have it fire upon blinking” task padded out with the aforementioned traversal mechanics. By comparison, the final blink tutorial sub-job features two separate buttons that the autocase must be able to fire at remotely, and it has a rotating chandelier obstructing your line-of-fire to add a timing aspect (though you can shut it down a short ways into the mission). No section of the last two missions ever approached that level of complexity. In the final job I skipped a whole section² on accident by walking to the back panel and shooting its glass walls open with the autocase. The weevil isn’t even used in the stock exchange job, and in the final job it’s only used to operate other interactables while it sits motionless!

This game would’ve benefited significantly from cutting everything besides the weevil and the autocase and focusing nine jobs around those. However much the CLI adds in terms of load to an uninitiated player, it contributes nothing as used without a heavier focus on script writing and timing, and the weevil and autocase are the only tools that provide these opportunities for the player. Here’s some examples of puzzle concepts I would have liked to see when building upon the core mechanics:

• There’s a tube (preferably not a straight line) that only the weevil can fit into with a button at the end. The tube has lasers that will trigger the alarm that can be briefly turned off by buttons in a separate room accessible only by gunfire. The player must program the weevil to navigate the tube while correctly timing shots to each laser’s respective “off” button to avoid triggering the alarm.
• There’s a maze of movable walls controlled by a series of buttons in a small chamber only the weevil can fit into. The player must program a sequence allowing the weevil to press each button independently so that the player can progress through the maze.
• There’s a glass-covered panel with a datajack inside that briefly opens a door. When the glass is shattered, an alarm sounds off within X seconds unless disabled by a button on the other side of the door. The player must place a weevil at the entrance of the panel, shoot the glass panel with the autocase, run a weevil script to walk inside and use the datajack, and then finally run through the door to disable the alarm.

Glimpses into the lives of the protagonists show moments of school life tenderness, far removed from the heist settings that most of the game takes place in. [src]

These ideas would get closer to full mechanical coverage: cooperation between tools, sequences of events happening in tight timing windows, and a need to carefully plan out and measure movements within commands. The engine exists here to make these kinds of puzzles happen, but it isn’t elaborated upon enough in order to create more complex situations. What’s left is a shallow puzzle game with a decent variety of mechanics and a eclectic, anachronistic take on cyberpunk stylings.