It has taken me a while to get myself to write this, and I don't feel all that inspired at the moment, but considering how obsessive I can be, if I don't just get this down and wait for a "perfect time", it will never happen.
For years now I have wanted to make a dream game. I would call it a Yume Nikki fangame but it will not be conservative and probably borrow very little from Yume Nikki, even though it is a huge inspiration (if that makes sense). I approached it from a few different angles. I decided it was going to be in 3D, and will contain many of my personal, recurring dream worlds. My ambition for this makes it a kind of "pet project" that everyone should just assume will never see the light of day.
Yume Nikki 3D by Zykov Eddy was important in giving me an art direction. My visions had previously been too good; I mean, the graphics would have taken too long to do anything useful with in a reasonable amount of time. The eduke 3D engine was quite perfect for Zykov Eddy's game, and the fact that it was free software was also a plus. The graphics were pleasing enough to look at, but did not require too much time or ingenuity, and the vagueness even added to it for obvious reasons.
I believe that, because dreams come from a person's mind, they are able to "introspect" them, at least to an extent. For example, I may have a dream that introduces some character my brain just made up on the spot. I may wake up knowing some details, such as the name or background, without it being explicitly stated in the dream, because my brain made up the character and stores all those details somewhere, even if it is in some unconscious part of the brain that reveals itself for a short period of time. I won't pretend to know the psychology that goes into dreams, but I think some of these conclusions are obvious or clear enough to be useful.
Because of this, I have been able to analyze myself during dreams. I can get certain feelings from my mind that I recognize from other things, like if I enter a room, I may notice that something may hesitate to let me in because I might be still thinking of what the room will be like, or I get the feeling in my mind that I get when I have to be creative and make something "random" but with inspiration. Again, I'm not claiming I "cracked the dream code" or something, but I know that some of this research I have done will work well in practice; that is, I would like to implement certain algorithms into my game that simulate what my mind is doing as best it can.
The longer I think about these algorithms, I realize more how LSD Dream Emulator got a lot of things correct. The differences in texture in environments that are still ultimately familiar, or even adding or removing objects from the area seem accurate in how dreams are structured. I have about 100 or so dream worlds that let me know I've been there before, even if the world itself looks dramatically different from how I have previously seen it. It's just something the mind does I guess, and it feels incredible when I enter a dream world I haven't seen for so many years, and may look quite different, and even have two or three dream worlds hacked together (and I can feel the distinction between areas) but I remember it and can even recall dreams I hadn't rehearsed for years. Some false memories in there? Maybe, but I believe this idea to be generally true. I have "checksums" for my memories that involve remembering memories to validate them, and this is a reason I spend time rehearsing many dreams. I have wanted to start a dream diary for years and years, but my obsessiveness has made it difficult to just sit down and start writing old and new ones. I know this is not a good excuse, so I'm actually getting a nice diary this Christmas. Writing this all down has already helped with that incredibly so thank you for reading.
You know how, mostly in free-roam games, there may be hundreds of houses which you cannot enter? The designers simply can't put so much time and money and hard drive space into readily filling up all those places. Neither can my brain in making dream worlds, although I can still enter random doors and have them lead somewhere. Sometimes I appear in a heavily modified version of my own house, maybe with a new character inside. Other times I appear in other dream worlds that take place in buildings, like this one mansion with infinite staircases. Sometimes I just get a generic, small room. Sometimes when I leave, I appear where I left off, I guess my brain kind of saved a pointer or something. Other times I may appear somewhere else, maybe random, maybe where I would normally appear, such as if I leave an area that looks like my house, I may appear in an area outside my house. I would still like it to be mostly linear in the sense that you can go from one place to another without too much randomness, and you can explore all (most) of the dream worlds without dumb luck, where A tends to lead to B and so forth. But there are also cases where this randomness will simulate real dreams much better. This is another example of LSD Dream Emulator getting an idea correctly, just maybe not implementing it the best it could.
So the chances of going from one place to another, and the chances of an area mutating this way of that, with these colors or these other worlds getting added or chopped up or associated near it, where does this come from? Well, I think that we can treat the structure of dreams in a similar manner to music theory, color theory, and some parts of information theory, where we analyze and take things that tend to be true, and then treat it as if it were inherently true. There will be randomness involved to decide things, but decisions wills be skewed based on statistic or intuition. Musicians studying the Baroque period will take Bach's Counterpoint methods as something that is "correct". Music, as all art, is of course subjective, but if something sounds good 90% of the time, it is useful to treat it as if it were true, and then further build from there. If you were to teach a machine to make Chorales, you would make it usually make its chord progression something like
I - IV - viio - iii - vi - ii - V - I
and then give it some chances to deviate in certain ways that may be more statistically likely than others, and of course very small chances to do something a little "odd" just because some people do odd
things on occasion and you want this machine to emulate human composition well.
I think methods like that are called machine learning, and certainly seem perfect for the job. I don't know the intricacies of it, I only studied music theory and information theory and enjoyed drawing similarities between the two. In this game engine, I could give it statistics I may make up based on intuition or write down after having a particularly introspective dream, and use those to help these algorithms make certain decisions in a more realistic and/or more fun and interesting way. This project will not only be a game but research and ideas into the theory of how dreams tend to be created and
I didn't mean to spoil those parts, those should have been bolded, whoops.
What do you guys think? Am I way off base here? You don't need to post anything about it never getting done; it's a personal project that I want to share with people also interested in dreams. At the very least I would like to formalize some algorithms I think would be useful, which is now one of the main focuses of this project. There are many ideas and details I have for the game outside of this but I feel that discussing the way dreams tend to work is more interesting and useful for you guys.
I have a domain and a webserver ready for a kind of shrine I was going to make for Madotsuki, mostyly as a a joke and as an educational experience but maybe I could set it up for you guys if you're interested in this project. I also have an archive of yume nikki and its many fangames in the works that I could post and expand with all of your help, if that also sounds interesting to you. I haven't set up the domain to my webserver yet because I currently have the worst ISP problem ever and it just isn't viable yet but it hopefully will be soon.
I think most of your ideas sound pretty good and generally accurate to what I've experienced in dreams. Especially in regards to how your brain just sort of "knows" things in dreams, even if you'd have no way of knowing it. Knowing where you are, knowing a particular dream characters name & history, etc. That doesn't seem to be something people tend to touch on when discussion dream and dream theory. Not that I've seen at least.
I think it's also important to remember that everybody has their own dream patterns as well. I for example will often times dream of familiar houses and buildings but there will be extra rooms or the layout will be slightly modified. My mother on the other hand, dreams that she's in existing places but the internal layout will be completely different, and often times the rooms will be arranged in grid patterns or even open up to entire forests or oceans. However taking into account everybody elses dream patterns might not be your goal here, since its a personal project you might only want to try and account for your own. Or maybe you're going for more a generic one-size-fits-all dreamscape. In which case I'd advise you to discuss other people's patterns and not work off the assumption that your own dreams are a sort of "default". It's ultimately your project though and you're free to do as you feel is best for it.
As for the websites, I for one would definitely be interested in following a project like this. Maybe wait for more people to speak up though, don't want to pay for hosting if only 1 person is going to be looking at it after all!
>>1910>In which case I'd advise you to discuss other people's patterns and not work off the assumption that your own dreams are a sort of "default".
Very good point, I should make an effort to try and make things more general. I would like to have others' inputs on this, maybe in a more formalized way, so that a program could "learn" off of the way people generally dream. Some kind of statistical model, you know? My mother has told me about the way her dreams work too, and they can be quite different from mine, but equally as fascinating.
As for the website, I would at least like to share my Yume Nikki archive project with people, even if this dream project isn't as popular. Maybe I should post about that in another place, like the Yume Nikki General board? I'm also a bit worried about this thread getting less attention by being on this board, but I figured I ought to put it here rather than the fangames board since this isn't as "concrete" yet.
Thanks for your input, there are so many questions I need to think about with this project
I wish you luck in your project, I do agree that this board isn't the ideal place to discuss it, considering how slow it is. This is really the only board I visit so I wouldn't know which other ones to put it on if you're looking for more traffic, I'm also not sure how strict the mods are on content being directly related to the board it's posted on. I assume since you're posting this on uboachan then you're mostly looking for people with a familiarity to yume nikki to discus this with? Perhaps you should venture out to some of the other yume nikki sites? I'm sure there are more than just uboachan, and I'm sure there's some pretty significant overlap between the userbases of one site to another. It may be easier that way rather than waiting for the few bites on here. Once your project gets off the ground tumblr is actually an excellent site for posting updates and screenshots on.
Again, good luck, hope you can get this all figured out.