On RUP and Agile

This is more of a theory I built around doing this and it is supposed to be a C H I L L jam. I have a bad approach to development (I think of every game I have ever made).

First comes the RUP approach. It might be good for large studios and large products like The Witcher where hundreds of people are specialists in their field (e. g. modelling, coding, scripting). For quick jams like this and a single person, this approach is a creativity killer and might be carried through only by toughest people in the society (like Navy SEALs). It always looks like it is never going to end, especially when the actual development part starts. What's more, implementing any new idea is far from possible. Example: A lot of code is written. There's code for player, objects and mobs are coded but not placed. There's map but it wasn't loaded ever, so I don't know how this all integrates. Puri wants to save the princess but he only knows that Princess extends flixel.FlxObject.

Another one is Bad Agile approach. When developers use Agile, they should focus on developing every field of the application equally in every sprint. Just GUI is not a product evolution and is not an effective sprint. In my case it was polishing HUD elements and writing some form of script parsing for mobs before I did even have a mob and any map relatable to the story. Example: Player can move around an example map, shoot limited bullets, go through teleporting doors. And that's all. Puri can move around his world, doesn't this sound like fun? Are we Greenlit on Steam already? Mine is something between this and RUP.

In my opinion the best approach is the proper Agile. Instead on focusing on entire story, just make a minimal demo. Minimal is a keyword here. Example: Player is on a cloud platform and needs to pass two standing mobs to reach the final door. Player can jump over mobs. If player touched the mob, level restarts. It's one of the moments of Puri's journey to save the princess. Even though it isn't an exciting game, it has all that is needed: a goal, an obstacle, choices of movement. Another evolution could be like this: Player has limited health. Mobs are moving across entire platform. Player can still jump over them but has 3 tries to be hurt. And another can add some bullets, items, etc. Basically: each evolution (Agile sprint) should be still a game with no unused parts. Such single evolution of product is enough to handle for a single person to develop - it doesn't take much time and progress is visible every successful compile. From any chosen point, the game can expand in both directions. You can even start by designing final level first and then step back with the starting point.

Tomorrow, I'm writing an exam from subject about this stuff.