Jump to content




[COMPETITION]


139 replies to this topic

#21 matejdro

  • Members
  • 324 posts

Posted 29 September 2012 - 04:36 PM

Very interesting idea. Just few questions:

1. By "No internet access will be provided." you mean no HTTP API access right?
2. AFAIK There are no chunkloaders that work with 1.3 (CC 1.42) yet.
3. It would be good if we would get an example map or at least a picture of the starting area.
4. In case of an unintended interruption (for example server crashing), will turtle start from beginning?

Otherwise I love the idea. Even without the competition thing, it is still awesome challenge to do. I will try this for sure.

#22 pruby

  • Members
  • 48 posts
  • LocationWellington, New Zealand

Posted 29 September 2012 - 11:34 PM

View Postmatejdro, on 29 September 2012 - 04:36 PM, said:

1. By "No internet access will be provided." you mean no HTTP API access right?

Yes, I mean the HTTP API will be disabled.

Quote

2. AFAIK There are no chunkloaders that work with 1.3 (CC 1.42) yet.

Still looking in to chunkloader options. If necessary will port one and test it.

Quote

3. It would be good if we would get an example map or at least a picture of the starting area.

I've set up a sample environment and am working on some pictures - will let you know. I'll see if it's easy to share the map also.

Quote

4. In case of an unintended interruption (for example server crashing), will turtle start from beginning?

If Minecraft crashes, the run will be re-started.

#23 pruby

  • Members
  • 48 posts
  • LocationWellington, New Zealand

Posted 29 September 2012 - 11:56 PM

View Postimmibis, on 29 September 2012 - 01:41 PM, said:

There is guaranteed to be another tree trunk within 8 blocks of the starting tree. Is the a maximum height difference between this tree and the starting tree?

I won't throw you a curve ball here - will be pretty close. The top wood block of the starting tree will be level with or above the bottom of the nearest tree, and vice versa. Not guaranteeing that searching at any one height will find it though.

#24 immibis

    Lua God

  • Members
  • 1,033 posts
  • LocationWellington, New Zealand

Posted 30 September 2012 - 12:28 AM

View Postjag_e_nummer_ett, on 29 September 2012 - 02:23 PM, said:

View Postimmibis, on 29 September 2012 - 01:41 PM, said:

There is guaranteed to be another tree trunk within 8 blocks of the starting tree. Is the a maximum height difference between this tree and the starting tree?
Dosen't matter, you can just do turtle.detectDown() to find out if you are at the ground level. And if you got a log left in the turtle's inventory you can turtle.compare() with the block in front of it.
It could be on a cliff, with a cave under it.

#25 pruby

  • Members
  • 48 posts
  • LocationWellington, New Zealand

Posted 30 September 2012 - 12:54 AM

OK, have photos and sample world. Since I don't have a chunkloader mod yet, an iron block has been placed where it would go. This will not be the world the actual judging takes place in - is a sample only. I recommend testing your program on a wide variety of similar sites to ensure it's general.

Site 1 - tree, turtle and chunkloader:
Posted Image

Notice that sand and water are within 16 spaces. This is the edge of a forest (seems water & sand most often found there).

Closer view of turtle:
Posted Image

Loader will be where this iron block is:

Posted Image

Site 2 is similar:
Posted Image

Posted Image

EDIT: Site 3 tree was too short. Not valid placement. Replacing site.

#26 pruby

  • Members
  • 48 posts
  • LocationWellington, New Zealand

Posted 30 September 2012 - 02:31 AM

New Site 3 set up:

Posted Image

World can be downloaded at:

http://www.goddard.n...%20Challenge.7z (4.7MB)

Unzip with 7zip in the saves directory. Environment includes:

* Minecraft 1.3.2
* Forge 4.1.1
* Computercraft 1.42
* Rei's minimap (for navigation)

The .points file in the 7z archive needs to be moved under mods/rei_minimap/ to get the sites as waypoints.

#27 Tahg

  • New Members
  • 7 posts

Posted 30 September 2012 - 10:47 AM

To avoid cheating on the challenge would it be permissible to provide a hash of our zip file before the submission deadline, and the actual zip within the day afterwards?

#28 matejdro

  • Members
  • 324 posts

Posted 30 September 2012 - 10:54 AM

Why not just send the zip at the deadline day?

#29 soccerboy5411

  • Members
  • 14 posts

Posted 30 September 2012 - 04:12 PM

Hello, supper excited to attempt this; however, I have a few questions:

1) Is the starting position going to have a set y value or will it be variable?

2) How do you plan on ensuring that all the needed resources will be found within the 16 x 16 chunk? Doing the math, in a perfect world that area will only provide us with 4 or 5 diamonds, which prevents us from making 2 crafty-mining turtles..much less any other turtles to complete bonus objectives. If expansion is the answer then it should be an inherit objective not a bonus one. :)/>

#30 jag

  • Members
  • 533 posts
  • LocationStockholm, Sweden

Posted 30 September 2012 - 05:40 PM

View Postsoccerboy5411, on 30 September 2012 - 04:12 PM, said:

Spoiler
Well make more chunk loaders then!

#31 Orwell

    Self-Destructive

  • Members
  • 1,091 posts

Posted 30 September 2012 - 06:03 PM

View Postsoccerboy5411, on 30 September 2012 - 04:12 PM, said:

2) How do you plan on ensuring that all the needed resources will be found within the 16 x 16 chunk? Doing the math, in a perfect world that area will only provide us with 4 or 5 diamonds, which prevents us from making 2 crafty-mining turtles..much less any other turtles to complete bonus objectives. If expansion is the answer then it should be an inherit objective not a bonus one. :)/>

Don't you have a 32x32 area to work with? So that would be 16 to 20 diamonds...

#32 matejdro

  • Members
  • 324 posts

Posted 30 September 2012 - 06:11 PM

You don't know where boundaries of chunks are and you are not guaranteed to spawn in the middle of the chunk (32 in every direction). If you spawn on the border of the chunk, it might be 16 in one direction and 48 in another direction.

So to be safe, you will have to assume that maximum range is 16x16.

#33 Orwell

    Self-Destructive

  • Members
  • 1,091 posts

Posted 30 September 2012 - 06:21 PM

View Postmatejdro, on 30 September 2012 - 06:11 PM, said:

You don't know where boundaries of chunks are and you are not guaranteed to spawn in the middle of the chunk (32 in every direction).
Wrong, it would be 24 in every direction (1 and a half chunk).

View Postmatejdro, on 30 September 2012 - 06:11 PM, said:

If you spawn on the border of the chunk, it might be 16 in one direction and 48 in another direction.
So this would be 16 in one direction, 32 in the other.

View Postmatejdro, on 30 September 2012 - 06:11 PM, said:

So to be safe, you will have to assume that maximum range is 16x16.
Well, in each direction you have a minimum of 16 blocks and a maximum of 32 blocks that are available. So to be safe you take the minimum everywhere. This makes 16 blocks in every of the four directions. So 16 blocks north + 16 blocks south = 32 blocks in that axis. And 16 blocks west + 16 blocks east = 32 blocks in the other axis. That makes a safe area of 32x32 blocks.

#34 matejdro

  • Members
  • 324 posts

Posted 30 September 2012 - 06:31 PM

You are correct. My bad.

#35 GopherAtl

  • Members
  • 888 posts

Posted 30 September 2012 - 08:15 PM

he is not correct, your turtle has no way of knowing where in the chunk it is, so it can't know which direction it can go 16 in and which it can go 32 in. Since more chunks will load while a player is present checking on it, then unload when they leave, it can't count on getting false back when trying to move out of the loaded chunks, either.

re: this, I'd suggest a revision to the rules: The judge player logging in to check on their torches, should log out only immediately after witnessing the turtle replace the broken torch. This ought to ensure the turtle is not caught with it's pants down, moving into a chunk that unloads when the judge leaves. Then we could identify the edges of the full 3x3 chunk area and exploit it all effectively.

#36 Orwell

    Self-Destructive

  • Members
  • 1,091 posts

Posted 30 September 2012 - 11:07 PM

View PostGopherAtl, on 30 September 2012 - 08:15 PM, said:

he is not correct, your turtle has no way of knowing where in the chunk it is, so it can't know which direction it can go 16 in and which it can go 32 in. Since more chunks will load while a player is present checking on it, then unload when they leave, it can't count on getting false back when trying to move out of the loaded chunks, either.

re: this, I'd suggest a revision to the rules: The judge player logging in to check on their torches, should log out only immediately after witnessing the turtle replace the broken torch. This ought to ensure the turtle is not caught with it's pants down, moving into a chunk that unloads when the judge leaves. Then we could identify the edges of the full 3x3 chunk area and exploit it all effectively.
I am still correct. You are right, we can't know which direction we can go 32 blocks in. So we have to assume it's 16 blocks in each direction. But still, then we have a 32x32 area. Because it's 16 blocks north, east, south and west, right?

And I still haven't seen any confirmation in this thread that turtle.forward()/turtle.back() indeed returns false when it tries to move outside a chunk. I'd have to check this to be sure.

#37 GopherAtl

  • Members
  • 888 posts

Posted 01 October 2012 - 02:03 AM

Oh, sorry, I misread your post. I saw the "1.5 chunks in each direction" re: center, then "16 in one, 32 in the other" for the edge, and apparently skimmed over the rest without entirely reading it. :)/> sorry bout that. Yes, 32x32 area, 16 blocks in each direction. This is only guaranteed to find the diamonds in the center chunk, though, meaning if you're unlucky there might be no diamonds at all, if there's an intersection between your one vein of diamonds in the center chunk and a canyon or dungeon or whatever. Unless the generation rules have changed in 1.3.2? I've not really looked into it since the upgrade from 1.2.5...

#38 pruby

  • Members
  • 48 posts
  • LocationWellington, New Zealand

Posted 01 October 2012 - 05:20 AM

View PostGopherAtl, on 30 September 2012 - 08:15 PM, said:

re: this, I'd suggest a revision to the rules: The judge player logging in to check on their torches, should log out only immediately after witnessing the turtle replace the broken torch. This ought to ensure the turtle is not caught with it's pants down, moving into a chunk that unloads when the judge leaves. Then we could identify the edges of the full 3x3 chunk area and exploit it all effectively.

Hi GopherAtl,

I can't guarantee that any chunk will _not_ be loaded and later unloaded except the 3x3 block with the chunkloader in the centre one. Please don't rely on any loaded chunk staying loaded unless in that area - I'd rather not see people's submissions failing due to chunk mechanics. You have a safe area of 33x33 (16 blocks to each side plus the centre line). If resources are missing from there, your program may have to accept a degree of risk by moving outside, but this should be managed carefully.

Sorry all about the chunk loader delay. Looks like at least RailCraft has them working in 1.3.2 again - will confirm.

Cheers,

Tim

#39 pruby

  • Members
  • 48 posts
  • LocationWellington, New Zealand

Posted 01 October 2012 - 05:28 AM

View PostTahg, on 30 September 2012 - 10:47 AM, said:

To avoid cheating on the challenge would it be permissible to provide a hash of our zip file before the submission deadline, and the actual zip within the day afterwards?

I'd rather avoid this, but can manage it. SHA1 only please (no MD5s), and the final file should be attached within 24 hours of the deadline.

#40 pruby

  • Members
  • 48 posts
  • LocationWellington, New Zealand

Posted 01 October 2012 - 05:35 AM

View Postsoccerboy5411, on 30 September 2012 - 04:12 PM, said:

Hello, supper excited to attempt this; however, I have a few questions:

1) Is the starting position going to have a set y value or will it be variable?

No guarantees here except that it will be at the base of a naturally spawned oak tree. I'm not going to plant one on a single block of dirt in layer 1 :)/>

Quote

2) How do you plan on ensuring that all the needed resources will be found within the 16 x 16 chunk? Doing the math, in a perfect world that area will only provide us with 4 or 5 diamonds, which prevents us from making 2 crafty-mining turtles..much less any other turtles to complete bonus objectives. If expansion is the answer then it should be an inherit objective not a bonus one. :(/>

There will be three sites and the world will be the same for everyone. If one is missing resources within the guaranteed safety area, it'll be the same for everyone. If, extraordinarily, multiple sites are missing resources then I'll re-start the judging for everyone.

I should clarify with submissions being run at three sites:
* If any site completes the base goal your submission is eligible for the prize.
* Extra goals can only be met at a site which has met the base goal.
* Credit will be given for the second site to meet a goal - shows consistency.
* If two sites succeed, no additional credit will be given for a third doing so - feel free to have one dud.





1 user(s) are reading this topic

0 members, 1 guests, 0 anonymous users