Tracery for LPC

[tsathoqqua]

Check Kryssa's work on the Tracery library for LPC.
https://github.com/GwennKoi/tracery_lpc/blob/main/tracery.c

What could we use this for? Could it be integrated into something cool? M_CONVERSATION? M_ACTIONS? Introduction system?

[azerwyn]

Three things pop into my head when I went to Kate's Tracery repository.

1. Semi-procedural NPC chatter in the spirit of "madlib" style constuction.
2. Semi-procedural quest/mission construction where the objectives end up being crafted into mutable quest/mission briefing(s) or stages.
3. It could be a precursor or dependency for phased descriptions and interactions. It could open the door to a transformative experience. As one progresses, the mud itself changes with a player's progression through their story.

The tool becomes the means of delivering repeatable outputs that might help reduce story/dialog grind. "Similar but not the same."

I suppose one could also use it to seed a body of story elements, rumors, red herrings, hints, etc into zones, domains, or mudwide where a player is hooked to hear what s/he is supposed to hear when they need to hear it; otherwise, they see/hear nothing (or something entirely different/unrelated). The same NPC could take on different roles for different players and their "unique" playthrough.

Assume for a moment it was done (coded into the mudlib). To successfully build with it might require storyboard-like planning similar to what might go into a novel or a AAA RPG. It's a big bite but I think it is (r)evolutionary and I think it complements the behavior tree stuff.

[tsathoqqua]

I think we're aiming for a module in /std/modules/m_tracery or m_babble or something. I regretted suggesting putting it into M_CONVERSATION, that's a fat enough module already.

I like all your suggestions @azerwyn! We need to make the interface a bit lighter than what Gwenn Koi did with hers. She made this example grammar:

mapping grammar = ([
  "rebel": ({"$vic shouts: Stay out #reb_insult# or #reb_threat#!",
    "$vic says: We don't want #reb_insult.a# like you here.", 
    "$vic threatens: #reb_threat.capitalize#, you #reb_insult#.",
    "$vic says: All of you #reb_insult.s# get off of our #curse#ing planet.",
    "$vic says: Our guild halls are closed to #reb_insult.a# like you.",
  }),
  "neutral": ({ 
    "$vic steps in front of you like a big #curse#ing jerk.",
    "$vic says: Are you sure you really want in there, #neut_insult#?",
    "$vic says: Why don't you pick a side in this fight? If you pick the "+
      "right one I might let you through.",
    "$vic says: $name was it? I heard you were #neut_insult.a#.",
    "$vic steps in front of you just to be a jerk."
  }),
  "gen_insult": ({ "#curse#head", "sonnuva gundark", }),
  "neut_insult": ({ "#curse#ing coward", "#curse#ing sympathizer", 
    "reb-hugger",  "fencepole sitter", "opportunistic #curse#hole",
    "war profiteer",
  }),
  "reb_insult": ({ "Rebel Scum", "#curse#ing dissident",  
    "#curse#ing rabblerouser", "#gen_insult#",  }),
  "reb_threat": ({ "I will kill you", "I'll toss you in the protein recylcer", 
    "I'll use you as target practice", }),
  "curse": ({"fikk", "dosh", "dank", "kark", "kriff", "krit", "pfassk", })
]);

But I think we could perhaps make that even simpler to work with so you would have to add this for each and every NPC, but perhaps have a library or a stack of different modules it could include inject_grammer("guard") ... I dunno. Something! :-)

Don't know if we could abstract the grammar more and make the sentences even more different?