u
p
d
a
t
e |
UPD: photo gallery, etc.
(01 Jun 2004 at 19:56) |
We swung by the Pittsburgh zoo this weekend, and I snagged a few photos before my camera's battery ran out.
Mike and I are printing up photos for "budget chic" decoration of our apartment. This is sweet.
I beat Metroid: Zero Mission in two days; 3:36 of game time. It was pretty easy, but I recommend it if you like that platform collect-the-items style. Getting 100% of the items is significantly more difficult; there are loads of infuriating dash-recharge-dash-etc. puzzles to solve.
The passport office totally screwed up my passport order; they forgot to take my birth certificate and somehow attached someone else's to my application. They seem to think it will all be fixed in time for my trip.
Today I had a minor type theoretic emergency as I discovered that my current design for the new version of our mobile lambda calculus won't work. I am pretty sure I understand how to fix it, but I hate the feeling of beginning the day with the belief that I'm almost done, and ending knowing that I am less done than I thought. As I've lamented before, I like to live monotonically.
Also, over the holiday weekend I got some good hacking done on escape. Now I'm adding ridiculous things like a pattern-based search and replace feature and automated level generation (assistance) based on retrograde analysis, which means I must be getting pretty close to a "1.0" release. |
 |
|
Yeah, I did retrograde analysis level generation for my game (the UFO puzzle, a.k.a. Lunar Lockout). I think it was based partially on your suggestion but it was something that I also thought of, but the idea isn't new, anyway. Sometimes RG is good but sometimes just randomly generating stuff and trying to solve it works nicely also. However, the more complicated the game, the worse the latter will be, I would think, so don't try it on anything but a limited subset of Escape.
I never released my game because I never came up with a good difficulty measure. I think the best one is probably enlightened random walk (random walk never repeating a position and never doing anything which would instantly kill you at least), but you have to try many random walks to come up with a good measure. |
| (RG here is short for Retrograde Generation, I guess.) |
| (Also my game appeared to have some bugs due to having been coded in C. And by "repeating a position" I mean "repeating an entire board state including character's position".) |
I'm sure retrograde generation is about as old as it gets. There are some 'heuristics' to what I'm doing, which amount to doing it only for a subset of escape. Otherwise the levels get pretty boring.
Why do you need a computer-generated difficulty measure? |
Damn. It segfaults whenever i try to connect to the internet. With option 3 or 4 or R on the main menu. I don't know why. (this is with the http://escape.spacebar.org/escape-alpha.tar.gz version) :'(
Cvs is giving me errors, so i haven't been able to try to compile it myself. :-(
Bug# 1? When editing a level, the border starts as blue blocks. But then if you clear it, the border is redrawn with whatever your currently selected piece is. >:O
Dashing was a pain in the ass, by the way. Then again, it was sort of nostalgic to have a video game demand dexterity. Like some of those waterpackless levels in Super Mario Sunshine. I'm glad part of that game was hard. And I'm pissed that there's no way to get back to them after you pass them the first time. =-O****
(_!_) |
| Also, one of my favorite gameboy games was Dexterity. It was fun. |
Bug #1 is not a bug, it's deliberate. But segfaulting when it tries to connect definitely is! I don't know anybody else trying to use the statically linked linux executables (other than me), so maybe they are not as portable as I thought. (Why is it so hard to distribute binaries on linux?!)
dash dash revolution |
| lovely photos :-) who's mike, by the way? |
| Mike is my brother -- see mike.soboring.org. |
| The most obvious reason why you'd want a difficulty measure is to throw out all trivial levels. After all, almost no level will have a unique solution, and I'd bet that almost always, the retrograde generated solution is not the fastest one. Also, it would be neat to see if you can get the computer's notion of difficulty to correspond with yours. That way you can come up with a "Tombot Difficulty Analysis" for every level, which would roughly correspond to your opinion of the level's difficulty, so those who are beginners can know where to start and those who are seasoned experts know which ones to skip. |
Oh, I see. I thought you were talking about using the difficulty measure for something else. Well, in my case you just have to look at the results and see if they are pleasing!
For escape we have a rating system based on user feedback. |
| "Type Theoretic Emergency" would be a good name for a band or radio show. If I ever get myself air-cleared I may steal that. |
| I agree. Do it! |
| By the way, that was actually me (ekorber), not apw2. apw2 considers the mere existence of type theory an emergency. |
| Heh heh, ok. ;) |
|
|
|