Author Topic: SnargleQuest, early days  (Read 5904 times)

Snargleplax

  • Rogueliker
  • ***
  • Posts: 50
  • Karma: +0/-0
  • snargleplax
    • View Profile
    • SnargleQuest dev log
SnargleQuest, early days
« on: March 01, 2012, 10:34:35 PM »
I'm developing a new roguelike from scratch in C++, working title SnargleQuest.  My primary focus with this project (my first roguelike, though I've done professional game development in the past) is to advance the state of the art for meaningful structure in procedurally generated content.

Specifically, I'm implementing a system that builds dungeon levels bottom-up from a particular specified structure (expressed as a graph, using the Boost Graph Library).  So, for example, you could have a level that contains a boss and two minibosses, and the connectivity of the rooms will force you to go through the minibosses' areas first in order to reach the boss, rather than allowing any paths that go around them.

Or, similarly, I'm planning to do lock-and-key puzzles of arbitrary complexity.  E.g., the red key unlocks the red door, behind which you find the orange key for the orange door, etc.  Having the room structure as a graph (and with BGL as a rich tool for working with it) means the system can set up really complex puzzle layouts while still guaranteeing that they are solvable.

The other area of PCG I'm hoping to improve on is storyline.  I'm working on a system that abstracts the notion of a basic "mad libs" style quest by allowing whole structural pieces to plug in to each other, rather than just filling out a single flat template.  The branching possibilities implied here should result in a wide gamut of possible content; the main challenge I see is in constraining it all to be sensible.

I've started a dev blog for the project, as a diary for myself and to garner interest and commentary from friends and community.  I'd very much appreciate the attention of anyone who cares to discuss these ideas, or would be interested in playtesting once I get to that point (not yet, but progress lately has been solid).  The blog is at http://snarglequest.blogspot.com/.

In order to take good advantage of the generic systems I'm putting in, I need to come up with a lot of different kinds of content that are enabled by this approach -- I need more examples than lock-and-key puzzles and pinch-point rooms containing monsters you're supposed to fight.  As I flesh out the procedural narrative ideas more, I expect to derive a lot of the environment structure from the plot structure, but I think it'll be good to have a large selection of stuff that occurs in non-plot-driven environments as well.  So, I'm open to suggestions.

getter77

  • Protector of the Temple
  • Global Moderator
  • Rogueliker
  • *****
  • Posts: 4956
  • Karma: +4/-1
    • View Profile
Re: SnargleQuest, early days
« Reply #1 on: March 02, 2012, 01:46:21 AM »
Sounds rather promising, keep at it!  8)

In terms of puzzle-heavy, yet RPG slanted, yet intended to be solvable...hmm....I'd say the top of the heap is DROD RPG: Tendry's Tale---might be worth a gander for some puzzle ideas at the least for meat on the bones.
Brian Emre Jeffears
Aspiring Designer/Programmer/Composer
In Training

Snargleplax

  • Rogueliker
  • ***
  • Posts: 50
  • Karma: +0/-0
  • snargleplax
    • View Profile
    • SnargleQuest dev log
Re: SnargleQuest, early days
« Reply #2 on: June 12, 2012, 08:16:12 PM »
This is coming along; the level layout engine is nearing what I'd call functional.  Here's a nice recent example: