Talk:Seed (level generation)

As of yet, there are a few theorys as to how these things work. I am somewhat new to wikis, so I probably shpuldnt have dumped a theory in the page. Shall we post them here? Lugnut1206 22:45, 22 February 2011 (UTC)

I have double checked that the same seed results in the same map. The spawn is different though, otherwise the seed is pretty much pointless.Lordhazzard 22:46, 22 February 2011 (UTC)

http://en.wikipedia.org/wiki/Random_seed

The seed number is used by the random number generator to make a sequence of "random" numbers to generate a "random" map. So using the same seed number should always generate the same map. I think the first random number generator took the middle four digits of a six-digit number and squared it to get the next number (or something like that).--PurpleKiwi 22:51, 22 February 2011 (UTC)

I've been testing out different seeds from the infamous 103, l33t, Notch, 4, 1. None of these produce the same results as what I was told I would receive. All very standard worlds with a few "landmarks" of sorts but nothing like I was told to expect. Try using a 19 digit seed and see if those produce similar results so there won't be any random digits added since you filled the field. MDR 22:55, 22 February 2011 (UTC)

I thought the same thing, but when i generate the map "1" a few times, i noticed the spawn moves and theres a sand + sign landmark thats a good marker. The map is the same in theory, but i think there are minor changes including trees, sand (if it moves?), mobs etc... Lordhazzard 22:57, 22 February 2011 (UTC)

Also i think to make things a lot easier, it would be good for players to goto x:0 and z:0 of the seed they generate and using that as the basis for the map design/layout/shape/biomes Lordhazzard 23:05, 22 February 2011 (UTC)

Some seeds provide everytime the same map with the same spawn location (like the seed "23"). It's said in the article that the spawn location is generated from the name of the world, but it doesn't work for the seed "23". In fact, if a seed provides a map with a spawn location of x=0.5 y=... z=0.5, then it will be always the same spawn location if you generate a new map with that seed (like "23" or "Glacier") whatever is the name of the map. But if the spawn is located on other coordinates, then a new generated map with the same seed will be exactly the same, except with another spawn location. That's what I experienced in the game. --Zeteube 14:40, 24 February 2011 (UTC)

Lock?
Can someone lock this for now? Most of this is total garbage. All the stuff on specific map features and placements isn't true; these seeds are only general map averages. --Fishrock123 02:10, 23 February 2011 (UTC)
 * I removed all but 5, which seems like a reasonable number of examples, but it looks like people are just going to keep adding more and more.. Manifold 03:08, 23 February 2011 (UTC)
 * Pffft. People. Anyways, I think 0-9 and probably alter the terrain average, and mixing the numbers probably mixes and/or amplifies the terrain alterations. A-Z might aswell. --Fishrock123 03:55, 23 February 2011 (UTC)

== Rename=%25253D

I suggest renaming this page to Seed (technical) (to avoid confusion with Seeds because the current title isn't useful. Is "technical" a good disambiguation for it though? --Gnu32 05:59, 23 February 2011 (UTC)
 * I'd have to say Seeds (technical) due to that's what it is. IKJames 15:36, 23 February 2011 (UTC)
 * How about Seeds (World Generator), to be more specific? [[File:Tjb0607SigIcon.png]] 21:50, 23 February 2011 (UTC)
 * Seems too wordy. (world) might be alright. Also, sans CamelCase when disambiguating. --Gnu32 23:17, 23 February 2011 (UTC)
 * I think it's been fixed now, for the most part. I'ma go ahead and remove the moving notice.--Kizzycocoa 20:31, 26 February 2011 (UTC)
 * Uh, not really.. it's still a long name. More input is needed I think. --Gnu32 10:55, 27 February 2011 (UTC)
 * Consider "Seed (Technical)" and have it redirect to "Seeds (Technical)". --Imdill3 05:22, 26 April 2011 (UTC)

*Facepalm* Edit war over fansite links
Seriously you guys? This is why we can't have a nice wiki. Grow up. --Jonnay 17:54, 20 March 2011 (UTC)

Can somebody tell me the letters or numbers required to get a seed with the water being replaced by lava?

I don't think that's possible with seeds. You could use a map editor though, if you really wanted.--Zkyo 16:03, 28 March 2011 (UTC)

It'd be easier if you went to the Nether. Also if you catch fire how will you stop flaming? That and Ghasts are why I hate the Nether. R ocĸetor talk  07:23, 30 March 2011 (UTC)

Different spawn point every time?
I'm not entirely convinced about this "different spawn point every time" thing. I tried out a seed on the forum that said that you would spawn in a dungeon, and I did, in fact, spawn in a dungeon.

Was this changed at some point? --Abcorntalk 07:39, 3 April 2011 (UTC)

I tested this with a friend- first using notable seeds, such as Glacier and Archimedes, then with random gibberish Result was 100% exact spawn. I imagine most of your probs concerning random spawns are related to tiny things like an extra space on the end of seed or a capital.Mick666 05:03, 8 April 2011 (UTC)

Can we get some clarification on how a alphabetic seed is converted to numeric?
Does anyone know how the String.hashCode function actually works?

The article lists the Calculation as "97 * 31**2 + 98 * 31 + 99 = 96354" (for the seed 'abc')

The Java documentation lists the formula for calculating hash codes as:

s[0]*31^(n-1) + s[1]*31^(n-2) + ... + s[n-1]

(where s[i] is the ith character of the string, n is the length of the string, and ^ indicates exponentiation)

I'm confused, these two examples seems to differ from each other. Help? Superbun 19:45, 12 April 2011 (UTC)
 * I see no contradiction. s = the string, s[i] equals the character in the string, n = the length of the string.  The string position is zero-indexed, so the count starts at zero.  Going by the example above:
 * s = "abc"
 * n = 3
 * s[0] = 'a'. ASCII code of 'a' = 97.
 * 97 * (31 ^ (n - 1))
 * 97 * (31 ^ (2))
 * 97 * 961 = 93217
 * s[1] = 'b'. ASCII code of 'b' = 98.
 * 98 * (31 ^ (n - 2))
 * 98 * (31 ^ 1)
 * 98 * 31 = 3038
 * s[2] = 'c'. ASCII code of 'c' = 99.
 * 99 * (31 ^ (n - 3))
 * 99 * (31 ^ 0)
 * 99 * 1 = 99
 * 93217 + 3038 + 99 = 96354. Thedarkfreak 03:02, 14 April 2011 (UTC)
 * 99 * 1 = 99
 * 93217 + 3038 + 99 = 96354. Thedarkfreak 03:02, 14 April 2011 (UTC)
 * 93217 + 3038 + 99 = 96354. Thedarkfreak 03:02, 14 April 2011 (UTC)

What belongs in here
Let’s talk about why some of you think that interesting seed fit here, and why others disagree. The same with minecraft fansites dedicated to Seeds.

Below a dump so it doesn’t get lost.

Interesting Seeds

 * 404
 * A world with a sinkhole. There is a challenge about surviving down there.
 * watch it here
 * or read about it here


 * worstseedever
 * In fact a nice world with some dungeons nearby. (Video)


 * 3666440496532277820
 * Your spawning point is inside a spider dungeon. Good luck.

Seed Fansites

 * mineseeds.net Seed database.
 * Minecraft Seeds A selection of the best seeds.

These are pretty good examples, but if you can edit this page, why not add them yourself? Also, please add a signature at the end, even if the user doesn't exist, so that we can call you something. -Master7432

Discussion
My Opinion is as follows: Why I understand that the selection of interesting Seeds is subjective, I’d recommend to include Screenshots of some. While this is subjective, too, we have example screenies everywhere, so nobody could be opposed to that solution.

Regarding Fansites; I think they have their place here. Don’t think so? Would you kindly click on the Arrow next to “Minecraft” in the main site bar on the left? Fansite links are even there. We are all a community, everyone who doesn’t agree is free to leave.

Conclusion: If nobody has good reasons against my suggestions, I will carry them out in a few days. Revertings of this edit without a good reason will be reported and the reverter will be punkished for vandalism. ‒Flying sheep 21:10, 12 April 2011 (UTC)
 * I don't agree with putting interesting seeds on the page itself. Links to fansites dedicated to seeds, absolutely, but I don't think individual seeds have their place here. Just my opinion. --Abcorntalk 21:33, 12 April 2011 (UTC)


 * Whenever we allow fansites, we just have people revert warring over them. – ultradude25 ( T at 01:23, 13 April 2011 (UTC)


 * Just put the fansite links under "External Links", and ask people nicely to not revert.--Abcorntalk 01:36, 13 April 2011 (UTC)


 * That's exactly how it was before, but people still did it up until the point where they earned themselves a block. – ultradude25 ( T at 02:24, 13 April 2011 (UTC)


 * Okay, here's a couple more ideas:
 * Put them back and put a comment like, or
 * Put them on a subpage like Random seed/Fansites --Abcorntalk 02:29, 13 April 2011 (UTC)


 * @Abcorn: Of course, a listing of individual seeds is too subjective. It would be educational, however, to post a screenshot of e.g. 3666440496532277820, which proves that you can spawn inside a dungeon, … As previously said; screenshots are there to illustrate a topic and are subjective per definitionem. So we could use them for 1-2 interesting seeds.
 * I think we should keep the links like this (or similar)



Link to my seed reader?
Just wondering if you guys think it'd be ok for me to add a link on this page to my seed reader: https://github.com/thedarkfreak/Minecraft-Save-Seed-Reader/wiki

The reason I suggest it is I find it much easier to use than the MC Seed Scope listed, as the user doesn't have to find and upload their level.dat file to get it. Thedarkfreak 03:02, 14 April 2011 (UTC)


 * I have no problem with you providing a link under External links. It appears what you are offering is different than what is there. However, remember that this is not an advertising service. A single link as an assist to the community is fine, but as noted above, any revert warring will result in a loss of editing privileges. -- Wynthyst [[Image:User Wynthyst sig icon.png ]] talk  09:43, 14 April 2011 (UTC)


 * Understood. If someone objects to the addition, I'll remove it.  I had a link to this up before, but it was removed during the fansite update war.  The reason given was that the original link I provided was a direct link to the JAR file(I didn't have the program on GitHub then).  I thought that was a reasonable explanation, as having the direct link on this page would cause a person's browser to ask to download something without knowing much of what it is.  With this new link, they get linked to a page with a better description of the program before downloading it. Thedarkfreak 00:36, 15 April 2011 (UTC)

"no random seed that will give you (...) entirely desert world."
Uhm, is it actually correct that it's impossible to obtain a entirely desert world? I mean, unless there is some criteria which somehow prevents such worlds being generated, or biomes exceeding certain size, wouldn't it theoretically be possible, although extremely unlikely to get a world in which only desert biome is generated?


 * Worlds are infinite, there's no biome big enough to cover an infinite space. 80% sure they have size limits. – ultradude25 ( T at 02:11, 21 April 2011 (UTC)


 * But wouldn't it be possible to have two (or in this case, infinite amount) biomes of the same kind next to each other, thus appearing like a single big biome? Mastadi 17:10, 21 April 2011 (UTC)


 * Well, given how the moisture scale works for biomes I think it's technically possible, but it would be a freak occurrence that such a seed fits inside the seed generation box. I think the comment on the page, while potentially misleading, is accurate enough to be a worthwhile piece of information to someone who doesn't know much about random seeds.  --Theothersteve7 17:37, 21 April 2011 (UTC)

Relevant info?
While working through some code for random seeds, I noticed something about the datatypes, that suggests that if you're using strings or phrases for the seeds in worlds, then you'll only be able to access a fraction of the available possibilities for seeds. The reason is, if a string is found in the World Seed entry field, then the Java hashCode function is used on the string to obtain a unique hash from it. However, this function returns a 32-bit integer(datatype int). The Random number generator's seed is a 64-bit integer(datatype long), and is stored as such in the save file. The other 32 bits seem to get filled with 1s when the int is upscaled to long. Those bits don't change depending on what string is entered.

However, if a numeric seed is entered, or Minecraft is allowed to generate a seed on its own, all 64 bits can be filled with information. Jeb's blog post on the seed generator says that Minecraft uses the current system time to generate a random seed. The most likely function used is System.currentTimeMillis, which returns a long value, the current system time in milliseconds. This also uses all 64 bits, instead of being restricted to 32.

Using pure numbers, or Minecraft's generation, allows for exponentially more possible worlds. Using a string, you're restricted to 2^32, or about 4 billion possible worlds(4,294,967,296 possibilities, to be exact). Now square that number, and that's how many possible worlds that the full 64 bits allows for. (2^64 possibilities, or about 18 quintillion possible worlds(18,446,744,073,709,551,616 possibilities, to be exact))

So, I was wondering if information considering this should be added to the article, stating that you won't be able to access as many potential worlds using seeds as you would otherwise. Sorry for the infodump :P -- Thedarkfreak 05:16, 26 April 2011 (UTC)
 * Sounds relevant enough to me. Feel free to add it, and in the future remember that it's alright to add this kind of thing right away after you've explained it(if it needs explaining). If it turns out to be unneeded it can always be reverted. - Alphap T ~ C 05:38, 26 April 2011 (UTC)