Posts for Derakon


Experienced Forum User
Joined: 7/2/2007
Posts: 3960
My only gripe is that the moon level music doesn't even get a full cycle before you get to the boss. It's the best theme in the game. :(
Pyrel - an open-source rewrite of the Angband roguelike game in Python.
Experienced Forum User
Joined: 7/2/2007
Posts: 3960
Correct me if I'm wrong, but what we have here is basically two any% TASes, one of which happens to have a much lower effective percentage than the other. As it happens, that means that the other run shows much more of the game. What I would suggest doing is call this run an any% run, and avoid trying to categorize the original run. It can be replaced if someone wants to make a newer "complete" run, by whatever criterion seems appropriate (all sanctuaries, complete story, glitchless, whatever).
Pyrel - an open-source rewrite of the Angband roguelike game in Python.
Experienced Forum User
Joined: 7/2/2007
Posts: 3960
Wow, that was a strange game. Especially, the lead-in to the endgame was all kinds of WTF (and I love how, even though the fight supposedly takes place in the past, Paula's prayers still are heard by people in the present). My only question is, right after Jeff joins, you do the stair glitch, then you die and do the stair glitch again. What was the purpose of the first use of the glitch?
Pyrel - an open-source rewrite of the Angband roguelike game in Python.
Experienced Forum User
Joined: 7/2/2007
Posts: 3960
The main reason to save and open vs. opening them in-browser is to add that extra step of "Okay, did I really intend to view this?" and prevent automated attacks from succeeding. Obviously if you open a malicious file with a vulnerable reader you're going to get hit, but now the site has to convince you that you want to read the file before it can make its attack.
Pyrel - an open-source rewrite of the Angband roguelike game in Python.
Experienced Forum User
Joined: 7/2/2007
Posts: 3960
That was a lot more interesting than I thought it'd be. I figured the low% run would mostly be using tedious glitches to fly all over the place, but I guess since your resources are so constrained you can't afford to do many bomb hovers. I'm definitely more supportive of a low% run now; just have to find someone willing to make it. :)
Pyrel - an open-source rewrite of the Angband roguelike game in Python.
Experienced Forum User
Joined: 7/2/2007
Posts: 3960
I guess it's been a bad three years, then. ;)
Pyrel - an open-source rewrite of the Angband roguelike game in Python.
Experienced Forum User
Joined: 7/2/2007
Posts: 3960
What gave you trouble? I could probably help you get started if the tutorials aren't cutting it for you.
Pyrel - an open-source rewrite of the Angband roguelike game in Python.
Experienced Forum User
Joined: 7/2/2007
Posts: 3960
Hunh...I just thought of something: I wonder if the fact that you only need to do the Shadow and Spirit temples is a nod to LttP, where you only need to finish dungeons 5 and 6 to get the mega bomb needed to get the silver arrows? Almost certainly not, but baseless speculation is fun. :)
Pyrel - an open-source rewrite of the Angband roguelike game in Python.
Experienced Forum User
Joined: 7/2/2007
Posts: 3960
MUGG: damage boosting's only worthwhile if you can manipulate health pickups en route; the healing you get when you pick up the Triforce shard takes longer than the boost gains you. Tompa's point about rooms visited, bushes burned, etc. is a good one and reveals why this issue is so murky. I still (mostly) maintain my stance -- I'm a bit iffy on the blue ring mainly because it's still available for purchase after you've bought it once (right?).
Pyrel - an open-source rewrite of the Angband roguelike game in Python.
Experienced Forum User
Joined: 7/2/2007
Posts: 3960
I think your decisions on what does and doesn't count for 100% are controversial. In general for a 100% game I would err on the side of getting too many things than on the side of skipping things. So for example, I would say that a 100% run would have to get all of the heart containers, both bomb upgrades, the magic shield, the red potion, all three swords, the blue ring, etc. Basically, my rubric is that at the end of the game, every flag relating to a gettable item in the game that is not a rupee, bomb, or key should be set, and the player should be as powerful and well-equipped as is possible.
Pyrel - an open-source rewrite of the Angband roguelike game in Python.
Experienced Forum User
Joined: 7/2/2007
Posts: 3960
The main thing that matters is how much text there is. If there's significant amounts, even if it's stupid, viewers will want to be able to read it.
Pyrel - an open-source rewrite of the Angband roguelike game in Python.
Experienced Forum User
Joined: 7/2/2007
Posts: 3960
He's basically saying that six of those items shouldn't count for percentage since they're just part of a fetch quest. There's absolutely no point in getting the cucco egg unless you plan on getting the biggoron sword, so the only percentage that counts here is whether or not you get the sword. 1%. I'm sure if someone made a low% run of sufficient quality, it'd get accepted.
Pyrel - an open-source rewrite of the Angband roguelike game in Python.
Experienced Forum User
Joined: 7/2/2007
Posts: 3960
Dunno what to tell you except that I've not had any problems tracking whitespace.
Pyrel - an open-source rewrite of the Angband roguelike game in Python.
Experienced Forum User
Joined: 7/2/2007
Posts: 3960
Well, vim at least is smart enough to know that 'return', 'elif', 'break' and the like mean to go out an indentation level, but when I'm ending a block with one of those, it's not too hard to hit delete to go back an indentation level.
Pyrel - an open-source rewrite of the Angband roguelike game in Python.
Experienced Forum User
Joined: 7/2/2007
Posts: 3960
The whitespace is a bit annoying when you start out, but it doesn't take too long to get used to it. Edit: fixed spelling. I r smrt.
Pyrel - an open-source rewrite of the Angband roguelike game in Python.
Experienced Forum User
Joined: 7/2/2007
Posts: 3960
Given that without the sim segments, Actraiser is just a meh platformer, I say do the sims. They're really the most significant feature the game has.
Pyrel - an open-source rewrite of the Angband roguelike game in Python.
Experienced Forum User
Joined: 7/2/2007
Posts: 3960
If you'd like to get into writing games, my advice is to get Python, download Pygame, and start working from there. Pygame gives you easy-to-use access to graphics and sound and is a good way to get your feet wet with game programming. And Python is probably the easiest language I've seen to learn. Not only that, but it's not a "baby" language; it retains all its power without being more complicated than other imperative languages.
Pyrel - an open-source rewrite of the Angband roguelike game in Python.
Experienced Forum User
Joined: 7/2/2007
Posts: 3960
Question -- what are these "cheat roads" you've been talking about?
Pyrel - an open-source rewrite of the Angband roguelike game in Python.
Experienced Forum User
Joined: 7/2/2007
Posts: 3960
I don't think the PSX category has enough entries to give one of them a star., just yet. :)
Pyrel - an open-source rewrite of the Angband roguelike game in Python.
Experienced Forum User
Joined: 7/2/2007
Posts: 3960
Actually, the real issue with that algorithm is figuring out what blocks are accessible from a given block. Simulating jump trajectories is no fun. :) Anyway, I've given up on that approach and have replaced it with something that's much simpler: I walk the walls and look for angles that look problematic. I'm getting decent results so far, though it still requires some tweaking.
Pyrel - an open-source rewrite of the Angband roguelike game in Python.
Experienced Forum User
Joined: 7/2/2007
Posts: 3960
There should be an autoseeder that kicks in if someone wants to download and there's no actual seeds available. It's not very fast but it should get you the movie file eventually. Is that not happening?
Pyrel - an open-source rewrite of the Angband roguelike game in Python.
Experienced Forum User
Joined: 7/2/2007
Posts: 3960
Generally I suspect that commercial games use A* pathing to move units around, and recalculate the pathing if another unit gets in the way (or just have the unit that was obstructed wait for the obstructing unit to get out of the way). RTSes have the advantage of not dealing with remotely realistic jumping -- the jumping units in SC2 are basically short-range flyers, for example. It's really a totally different can of worms dealing with platforming levels. For the moment, I've given up on determining actual accessibility, and am looking into just trying to recognize situations that look like trouble (e.g. long steep walls, overhangs, and the like). That just involves walking along the perimeter and calculating the local slope at intervals, which is much more straightforward. Edit: guess BBCode doesn't like the * in the URL? Weird.
Pyrel - an open-source rewrite of the Angband roguelike game in Python.
Experienced Forum User
Joined: 7/2/2007
Posts: 3960
World Builder I wish Blender had an interface like that.
Pyrel - an open-source rewrite of the Angband roguelike game in Python.
Experienced Forum User
Joined: 7/2/2007
Posts: 3960
Not much point in writing a robust random map generator if I'm only going to use it once. :) The idea is to remake the map every time you play; I'd have stuck "procedurally" into the topic title if it would have fit.
Pyrel - an open-source rewrite of the Angband roguelike game in Python.
Post subject: Algorithms: pathfinding in generated platforming levels
Experienced Forum User
Joined: 7/2/2007
Posts: 3960
This is a problem I've been mulling over for a bit; I'm interested in seeing how other people approach it. I have a program I've written that generates cave systems, like this: I'd like to use it for a Metroid-style game of exploring, finding powerups that let you get past barriers, etc., but to do that I need to be able to programatically figure out what parts of the map are accessible to the player. That map above has some areas that can't be reached (straight up from the middle, for example, or the long vertical cave on the right side) and many that can't be exited once entered (most of the rectangular rooms, for example). My starting assumptions: the player is 3 blocks tall and 1 block wide. He can jump 3 blocks into the air (i.e. if there is a wall that is as tall as he is, he can jump onto it). He can jump over a gap that is five blocks wide (total jump length 6 blocks). He cannot crawl, turn into a morph ball, slide, or anything like that. I started working on an algorithm that starts with a known-accessible block and tries to find out what blocks are reachable from it. The algorithm looks roughly like this:
Pop accessible block A off of queue
Foreach distance d < player jump length:
  Look at space B d to the side of A
  Is that space occupied by a block:
    B is accessible by walking/jumping. Enqueue it
    # Abstracted away: checking for ceiling room, horizontal barriers
  Else:
    Look above B for the closest floor C with enough headroom above it for the player to fit
    Is it lower than the player's max jump height:
      C is accessible by jumping. Enqueue it
      # Abstracted away: previously-jumped-to targets; decaying jump height with distance
    Look below B for the closest floor D:
    Is D far enough below C for the player to fit:
      D is accessible by dropping down. Enqueue it
      # Abstracted away: horizontal barriers
  Repeat
Repeat
All of the "abstracted away" notes are basically things that this "teleport horizontally over and then look up and down" algorithm doesn't handle well. There's also messiness with things like a tunnel that is reachable by jumping, but only if you start out above it (there's an example of that down and a bit left of center in the example map). Basically I think this approach is flawed, but I'm having difficulty thinking of one that would work better. If you think the problem's interesting, I'd love to hear your thoughts on it. Edit: I've come up with an alternative approach that I think is less error-prone. I've written a description with images here. Thoughts on that approach?
Pyrel - an open-source rewrite of the Angband roguelike game in Python.