The pseudocode definitely looks like it would work, that's probably all you'd really need to find epic seeds.
Since there are so many seeds, you could also make sure the world includes all of the interesting stuff... reeds, cactus, mushrooms, etc. Shouldn't be hard at all.
The predictions thing is definitely overkill, but I think that's why it could be awesome. I was thinking of sending the entire map data and just letting it figure things out for itself (once we categorize some maps for it to learn from, of course). Not sure if there's a size limit...
The pseudocode definitely looks like it would work, that's probably all you'd really need to find epic seeds.
Since there are so many seeds, you could also make sure the world includes all of the interesting stuff... reeds, cactus, mushrooms, etc. Shouldn't be hard at all.
The predictions thing is definitely overkill, but I think that's why it could be awesome. I was thinking of sending the entire map data and just letting it figure things out for itself (once we categorize some maps for it to learn from, of course). Not sure if there's a size limit...
- Thanks; now to port that to Python for MonkeyRunner and/or MCEditEdit: Too lazy to learn another language; maybe I could try to see if anything else loads Pocket levels.
- % of blocks can also indicate biomes - ice = snow, etc
- Google Prediction, I think, only takes text files with statistics in them, not raw binary.
Edit: could this also be done on the desktop with a older version of the standard Minecraft world generator? Does the Pocket Edition uses the 1.7.3 and older generation code?
- Google Prediction, I think, only takes text files with statistics in them, not raw binary.
Ah, you're right, it takes a CSV, so no binary. I guess it's pretty text-oriented, too. So yeah, by the time you prepared the data for the API, you could pretty much just make your own conclusions from it.
I'm interested actually....can u guys explain a little more? I'm not following all thhe way...
I'm just wondering whether the process of testing seeds can be automated - ideally, you'll feed a text file with seeds you want to try, connect your tablet to your computer, and leave them running. The computer will automatically control the tablet to generate worlds, take screenshots, and fetch the resulting chunks for further analysis if that's possible. Finally, the screenshots and the seeds would be posted automatically.
Try
61068
It is not very at but if u go straight u have a nice flat ocean and if u go the the right of where you spawn u can find sugarcane.... cactus is all behind you and theirs a forest behind where u spawn....
i dont know if I can get a map flatter than the first one I gave u but ill keep trying....
????
Since there are so many seeds, you could also make sure the world includes all of the interesting stuff... reeds, cactus, mushrooms, etc. Shouldn't be hard at all.
The predictions thing is definitely overkill, but I think that's why it could be awesome. I was thinking of sending the entire map data and just letting it figure things out for itself (once we categorize some maps for it to learn from, of course). Not sure if there's a size limit...
- Thanks; now to port that to
Python for MonkeyRunner and/or MCEditEdit: Too lazy to learn another language; maybe I could try to see if anything else loads Pocket levels.- % of blocks can also indicate biomes - ice = snow, etc
- Google Prediction, I think, only takes text files with statistics in them, not raw binary.
Edit: could this also be done on the desktop with a older version of the standard Minecraft world generator? Does the Pocket Edition uses the 1.7.3 and older generation code?
Ah, you're right, it takes a CSV, so no binary. I guess it's pretty text-oriented, too. So yeah, by the time you prepared the data for the API, you could pretty much just make your own conclusions from it.
I'm just wondering whether the process of testing seeds can be automated - ideally, you'll feed a text file with seeds you want to try, connect your tablet to your computer, and leave them running. The computer will automatically control the tablet to generate worlds, take screenshots, and fetch the resulting chunks for further analysis if that's possible. Finally, the screenshots and the seeds would be posted automatically.
61068
It is not very at but if u go straight u have a nice flat ocean and if u go the the right of where you spawn u can find sugarcane.... cactus is all behind you and theirs a forest behind where u spawn....
i dont know if I can get a map flatter than the first one I gave u but ill keep trying....
SEED: i*space*want*space*your*space*head
Try it out.....(also if its not flat enough just clear away dirt and if theirs unwanted lower land fill it with the dirt u shoveled!)
thanks
Don't think he would reply anymore...