Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Root Infinity]

So I decided to start a megaproject. I've partially planned it out mentally - I'm going to build a castle controlled by a Turing machine. The Turing machine is what I'm currently focusing on; hopefully the rest will follow. This will mostly be done as Quickfort designs, to save my sanity.

Mods/utilities:

Spoiler (click to show/hide)

• [SPEED:0] dwarves
• [INVADERS: NO]
• DFProspector
• DFAutodump
• DFCleanmap
• Stonesense
• Quickfort

And that's it - no DFliquids for me!

General schematic assumptions:

Spoiler (click to show/hide)

• Any letters are gears connected to the respective input
• Any letters that are lower-case mean that the gear has been pre-toggled
• Pressure plates trigger on 7-7 water
• Rows of dashes on side views show where floors are

^ Pressure plate
# Wall
. Floor

Code: [Select]

  ^
  %
<% %> Pumps (pumping away from center)
  %
  v= | Axles
* Gear assembly (either not connected to anything or currently engaged)
+ Gear assembly (currently disengaged)
~ Water
… Grate (Side view)

Todo list:

• Embark Site
• Specs
  Spoiler (click to show/hide)

  Need:

  • Lots of wood (machinery)
  • Stream or brook (water; anything larger kills FPS)
  • 20+ z levels aboveground (castle)
  • ~20 z levels belowground before caverns (4x4 for tape + 2 for logic + 2 for castle traps/power)
  • No aquifer

  Want:

  • Volcano
  • Flat
  • Traders

  In other words a flat forest volcano, with sand & no aquifers, and a stream or brook. And preferably it should be all in a 3x4 area (fits the tape nicely), for FPS' sake, with the volcano in the corner so that it doesn't impact the tape. This will be !!FUN!!...
  
  Actually on second thought a flat desert volcano would also work, as long as it has a good source of flow, as DWRs kill my FPS...
• Genning/Finding
  Spoiler (click to show/hide)

  I found a location! A tropical woodland (4x3) - not as much wood as I hoped but it has sand, no aquifers, a river, and a volcano in the corner.
  DFProspector output (selected):
  Layer materials:
                   SAND_TAN : 378
  >>> TOTAL = 734859
  Vein materials:
               TETRAHEDRITE : 27620
                NATIVE_GOLD : 21172
                     GALENA : 9322
                   HEMATITE : 7181
              NATIVE_COPPER : 4326
              NATIVE_SILVER : 4179
             RAW_ADAMANTINE : 3618
            NATIVE_PLATINUM : 455
            NATIVE_ALUMINUM : 334
                HORN_SILVER : 307
  >>> TOTAL = 422695
  Shrubs:
  >>> TOTAL = 1632
  Wood in trees:
  >>> TOTAL = 1645
  
• Turing machine
• Tape
  • Specs
  • Test
  • Build
• Incrementer/Decrementer
• Specs
  Spoiler (click to show/hide)

  Must be scalable and fast. Two inputs (U&D); n outputs. U increases outputs by one; D decreases outputs by one. Also should be fairly easily resettable
• Design
• Test
• Build
• Controlling logic
• Specs
  Spoiler (click to show/hide)

  4-tuples (read; write; inc/dec; change state); 16 states*2 memory states; one permanently set to HLT. An array of levers will program it.
• Design
• Test
• Build
• Castle
• Design
  Spoiler (click to show/hide)

  Cast out of obsidian
• Build
• Cast

[Root Infinity]

Annnnnd... nothing. Not a single reply? odd...

[Organum]

Wut is a turing ma-cheen i just don' know

[Sphalerite]

Good luck with this.  I've considered building one of these myself.  I don't think an actual turing machine has ever been built in DF, although there was of course the working (simple) computer someone built, and my Game of Life engine.

[expwnent]

The game of life is actually Turing-complete, so technically you did build something as powerful as a Turing machine. Ignoring space restrictions.

[rangarkash]

You realize the concept of a Turing machine is highly dependent on infinite storage space right? 

[Montague]

So like, you can umm... calculate things? or... stuff... and junk?

I dunno what you'd do with this... how would it work, even?

[AutomataKittay]

Looking over at your requirement, though I know you already found a spot, it seems like an one cavern world with large amount of layers above and below it, as well as high Z set, would be perfect for your purpose. Though you can remove the [AQUIFER] tag to make it easier and seek a sandy spot for glass, I don't think you would need much glass. Unless you're doing a lot of magma works, or just like the looks. With a few layers just to dedicated a tree farm in, and an obsidian-casting layer or two, and hopefully a lot of water in cavern ( water's adjustable in advanced perimeter also) sounds like it'd be perfect spot to make an automated fortress.

I thinks tower-cap would make for nice contrast with the obsidian surrounding! Though black-cap probably could show up even on one-cavern worlds and blend in, color-wise, at least in ASCII fort.

I don't really have much to say otherwise, other than good luck! A true Turing machine would need a few ( or a dozen plus ) layers of storage just to run the tapes, I'd imagine. Would be nice to get some snapshots of the innards, and maybe a wiki entry once it's done and working 

[tomblifter]

Do you have the design ready? I'd love to see it.

[uncool]

The thing I'm curious about is how you're going to implement the tape. Since there are no fully moving parts, are you just going to have a global thing that shifts water from one part to the next?
=Uncool-

[Cheese]

Interesting. I hope this works out.

[LastDwarf]

I've spent a good bit of time working out the most efficient tape using the motion of water. Here is the template for a two-cell tape:

t%>s
^  %
%  v
s<%t

The 's' indicates a hole for storing the fluid, and the 't' indicates a hole used for transferring it. To make the tape longer, repeat the pattern diagonally along the s holes.

I've got two different ways of controlling the water motion.

One is to selectively power the screw pumps. Start by powering either the north pumps, or the south pumps. Then turn them off. Then activate the east/west pumps.

The second is to have the pumps powered all the time, and have the t-holes be doors rather than holes, and have the east/west pumps a z-level higher with a hatch at their output. To move, close the hatch blocking the output of the east/west pumps, and open a set of doors. Then close all the doors. Then open the hatch.

I think the first is easier to build, but the second would be faster.

[Root Infinity]

I've spent a good bit of time working out the most efficient tape using the motion of water. Here is the template for a two-cell tape:

t%>s
^  %
%  v
s<%t

The 's' indicates a hole for storing the fluid, and the 't' indicates a hole used for transferring it. To make the tape longer, repeat the pattern diagonally along the s holes.

I've got two different ways of controlling the water motion.

One is to selectively power the screw pumps. Start by powering either the north pumps, or the south pumps. Then turn them off. Then activate the east/west pumps.

The second is to have the pumps powered all the time, and have the t-holes be doors rather than holes, and have the east/west pumps a z-level higher with a hatch at their output. To move, close the hatch blocking the output of the east/west pumps, and open a set of doors. Then close all the doors. Then open the hatch.

I think the first is easier to build, but the second would be faster.

I tried that, but it seems to be space-inefficient. Your idea using doors is much better though (It drops the z-levels required from 3 to 2) making it a possibility again... Thanks! It seems much more dwarfy than a bank of memory and a incrementer/decrementer (Which I have designed, by the way...)

Looking over at your requirement, though I know you already found a spot, it seems like an one cavern world with large amount of layers above and below it, as well as high Z set, would be perfect for your purpose. Though you can remove the [AQUIFER] tag to make it easier and seek a sandy spot for glass, I don't think you would need much glass. Unless you're doing a lot of magma works, or just like the looks. With a few layers just to dedicated a tree farm in, and an obsidian-casting layer or two, and hopefully a lot of water in cavern ( water's adjustable in advanced perimeter also) sounds like it'd be perfect spot to make an automated fortress.

I thinks tower-cap would make for nice contrast with the obsidian surrounding! Though black-cap probably could show up even on one-cavern worlds and blend in, color-wise, at least in ASCII fort.

I don't really have much to say otherwise, other than good luck! A true Turing machine would need a few ( or a dozen plus ) layers of storage just to run the tapes, I'd imagine. Would be nice to get some snapshots of the innards, and maybe a wiki entry once it's done and working 

• 3-cavern world (Wanted as much wood as possible)
• Z levels (top to bottom) 20 1 20 1 20 20 1 (Lots of aboveground; cavern close to surface; lots of building space below cavern 1; almost no space before cavern 3;lots of space below caverns for both minerals and cotton candy
• I am using sand for pumps
• Almost full cavern (will be pumped out/diverted, as the stream is just for power)

The thing I'm curious about is how you're going to implement the tape. Since there are no fully moving parts, are you just going to have a global thing that shifts water from one part to the next?
=Uncool-

Started with thinking about a bank of memory and an incrementer/decrementer, but I think I will have pumps moving water. Seems more dwarfy somehow...

[expwnent]

You realize the concept of a Turing machine is highly dependent on infinite storage space right? 

The game of life is actually Turing-complete, so technically you did build something as powerful as a Turing machine. Ignoring space restrictions.

[Root Infinity]

You realize the concept of a Turing machine is highly dependent on infinite storage space right? 

The game of life is actually Turing-complete, so technically you did build something as powerful as a Turing machine. Ignoring space restrictions.

Yea... any "Turing maching" IRL isn't actually a TM, but a finite approximation of one.