Let the transmutation begin! - Rock, Paper, Zachtronics

[Jurple]

There’s no need for eviscerators

Need? Nope, but it would be fun.

the problem you’re asked to solve here is the same one as Munitions Refill Type 6, only with an extra assembly step that you did in Training Routine 3.

I see that, but my implementation for Munitions Refill Type 6 was … lacking. In this case, with only two sinks instead of three, I ought to do better.

[LTK]

Gaaaaaaaargh. I got halfway through building the bomb before realising I still had to send it through the teleporter. I had such a good design, but it’s completely useless!

[Heliocentric]

Gaaaaaaaargh. I got halfway through building the bomb before realising I still had to send it through the teleporter. I had such a good design, but it’s completely useless!

Did that too, so annoying, I even pre built sausages of my final product to push through the teleporter. And they all got split into blocks.

[LTK]

I did it though. That trick Gus showed me where you use one blocker to pull a row of blocks out from under your entire component buffer has become my bread and butter, it’s just so useful in teleporter missions. You can replace some of the floor of that buffer with welders so that the components that need welding are instantly welded the moment they come out.

To continue the analogy, loose blocks are my peanut butter, and they let me do all sorts of things I thought I couldn’t previously. Usually it’s with a conveyor belt attached to a pusher, which waits to move the product until necessary components have been welded to it. It’s also possible to move a large structure for, say, moving a set of unwelded objects to a welder in sync. Another example:Here’s how I disassembled the anti-javelin satellite in one smooth motion, while keeping the components sorted.

There’s no need for eviscerators

Need? Nope, but it would be fun.

Actually, I see there’s an achievement for completing that puzzle with only two welders, so an eviscerator would be very useful in that solution!

[Jurple]

So, regarding the footprint calculation …

One of the Steam achievements for IF is 'Optimization (Training Routine 1)": solve TR1 with a footprint of 50 or less.

Easy, thinks I. I just need a few conveyor units to steer the 3 kinds of blocks around each other (all on the ground); I don’t even need the neutral blocks, just let the feeder blocks create their own pathway.

[ATTACH=CONFIG]11591[/ATTACH]
And yet: 15 blocks used (all conveyor), footprint of 56. How? Why? Wherefore? Please help me make sense of this!

Ok, I can make sense of it myself: dumped, stationary blocks become part of the footprint. I just happen to disagree with that fact of life.

stamps foot

[LTK]

The footprint is the horizontal area that your solution covers. Any square that either contains a block or where a block passes through is added to the footprint. So avoiding the use of platforms doesn’t help in reducing your footprint because your products are passing there anyway. Even if you pillared a hundred blocks over the path, it would do nothing to your footprint.

That achievement is a good way of getting you to understand footprint and how to minimise it. Making use of vertical space is crucial.

I felt like I did that pretty well in Solar Cell Array, which looks pretty easy until you realise that the solution that first comes to mind is impossible. An alternate solution wasn’t hard to think of though. How I solved it looks amazing really, it all fits together like a jigsaw, almost every available space is filled with pushers, lifters, sensors, welders and conveyors. I devised a system where the welders are lowered into place when the components need to be welded, and then they’re lifted back up so that the completed product can exit the assembly. It wouldn’t look out of place on How It’s Made.

And just when I thought I was getting good at block number efficiency, I see Gus does it with four times fewer blocks. Daaaaaaaaaamn. I need to go back to the drawing board if I’m to have a chance at beating that.

[LTK]

Anyone else spot this little fellow?

https://dl.dropboxusercontent.com/u/20662983/Screenshots/20161208230038_1.jpg

[Jurple]

My first solution to Small Excavator involved a blocker and an eviscerator and I didn’t realise I wasn’t supposed to use them yet until later. I was struck by inspiration and decided to redo my solution and it’s definitely my most elegant one yet. The eviscerator is optional, the excess blocks can just overflow to that side, but it was neater just to destroy them. It’s a shame I haven’t found a better way to deal with the slow pace of the block sculptor.

https://www.youtube.com/watch?v=ckXD9-9asE0

That’s compact!

I’m nearly done with Small Excavator. I have the two sets of sculpted undercarriage and the crossbar all falling in place, welded together JIT and then pushed onwards by a pusher triggered by the next set of blocks falling.

And now I realise I still have a sync issue, as I’m building the crossbar (4 blocks) half again as fast as the undercarriage (6 blocks). Bah.

Still fixable - I can accumulate 6 blocks for the crossbar, throw 2 of them away, and weld the rest. Tomorrow, though. I wonder if there’s solutions for this puzzle that don’t involve throwing blocks away, or using a large buffer/storage?

Also, I wonder if this is the last sync issue I still have with this puzzle. Time will tell!

[LTK]

Gus suggested to weld the undercarriage components together before sending them through the sculptor, which I just tried and it works out perfectly! No more overflow. And using my puts on sunglasses advanced knowledge I shaved a further 16 blocks off the solution. It was mostly conveyor belts.
[SIZE=2]
Also, when I reached the mission conclusion on Atropos Station, I thought “maybe this will be the first Zachtronics game I actually beat!” That hope vanished immediately when I saw what I was expected to build next. I have to put those on that in one piece? How???
[/SIZE]

[Gus_Smedstad]

I assume you’re talking about Navigation Computer? That’s really not as difficult as it looks. Basically, my solution involves aligning all the components next to the circuit board, an AND gate to make sure everything’s ready, and a pusher connected to a welder network that moves the components in place and welds at the same time. No elaborate assembly steps, just a single operation for each board.

Skip Drive (the last mission of the Homeward Fleet), now that one’s hairy. Because the construct is so big overall.

The penultimate mission (Fire Control) is one of my favorites. It’s the only mission where you can alter things in progress - you get a Switch part that can send signals like any other sensor. You’re firing missiles (which you construct) at hostile ships, and firing point defense cannons against incoming missiles and fighters. I liked it not because of the interaction, but because it absolutely demands time efficiency. A slow solution won’t send out your attacks in a timely manner.

[Jurple]

Gus suggested to weld the undercarriage components together before sending them through the sculptor, which I just tried and it works out perfectly! No more overflow.

I was doing that already - I didn’t realize the sculptor moves its blocks by itself, and doesn’t need pushing/pulling by conveyor belts at either end. I don’t see how changing that eliminates overflow.

And firing the game up at $WORK to examine my assumptions sets a bad precedent

[LTK]

You know what, you’re totally right. The solution I recorded above was dividing the blocks in the completely wrong ratio, and moving the welder from after the sculptor to before it fixed that. If the gap between how fast both of your components are produced isn’t too large, you can just buffer one of them until the other one arrives. If that doesn’t work, you have to build in a delay at the input.

I assume you’re talking about Navigation Computer? That’s really not as difficult as it looks. Basically, my solution involves aligning all the components next to the circuit board, an AND gate to make sure everything’s ready, and a pusher connected to a welder network that moves the components in place and welds at the same time. No elaborate assembly steps, just a single operation for each board.

It’s fairly straightforward if you know that you can weld with non-adjacent welders and through components, which I didn’t. Would have been nice to know earlier. I could have solved Control Console in a much less roundabout way.

I was looking at Navigation Computer wondering if I had to drill a pattern into it to expose weld points and then taking a second board and drilling the same pattern out of it to replace in the original board after the components have been welded, but it turns out the board is indestructible. Then I wondered whether the slots were special and would weld to the components automatically, which also isn’t true. Then I tried putting welders at both the front and the back of the board, and lo, that did the trick.

Should have figured that out sooner, really.

[Gus_Smedstad]

Understanding how the welders work is not immediately obvious. Particularly since, like with everything else, the only instructions you receive are a couple of infographics.

A welder creates a “weld point” in front of it. Any two adjacent weld points create a welding arc. The welder’s position and angle are otherwise unimportant - as long as the weld points are adjacent, it will work. So you can have welders coming in from all sorts of odd angles to create welds. You can, for example, have welders in an S-shape, left, weld, weld, right, and it will work.

Thus, “through” doesn’t really matter, nor the relative positions of the welders, just the locations of the weld points.

[LTK]

Got it. Now that I know this is possible, I’m even more annoyed by the lack of bottom-oriented welders and sensors. It means I can weld together the top 6x6 base of the Fusion Reactor with the other components, but to do the same with the bottom 6x6 base I’d need to split it into two to create space for welders. Very irksome.

I completed Navigation Computer with the old rotate, assemble, rotate back method, which is wonderful for keeping the footprint low and makes it much easier to avoid blocking your product in. Timing was a bit of an issue though, by the time all the components had arrived, the next board was already backed up at the input area, and because they’re indestructible I just had to move one of them out of the way. This allowed everything to sync up perfectly at +0 and +1. I also had a constant excess of CPUs. Really weird input ratio there.

[Jurple]

Yes! Finally a solution I’m proud of.

[ATTACH=CONFIG]11633[/ATTACH]

Still not optimal in footprint, and it doesn’t work at higher input rates, but never mind that.

I did use Gus’s trick of having actuators drop into position so they’re not immobilised by adjacent blocks/pipes.

Edit: I really seem to suck at including media. Does this thing move? Nope, can’t seem to include it. How do you guys include large-size images or moving gifs?

Also, I may not need 3 different sensors (one for the conveyor, one for the left weld, one for the right weld & blocker). There might be more tweaking to be done …

[Jurple]

Actually, I see there’s an achievement for completing that puzzle with only two welders, so an eviscerator would be very useful in that solution!

I did just that, and it does wonders for one’s block score, as expected. Now to tackle the footprint …

[LTK]

After a long and painstaking design process, I finally completed one fusion reactor. Assembly of a second one breaks down at the delivery mechanism of the central core, which isn’t too surprising considering both halves of the lower frame are assembled by the same mechanism, sans core for the second half. I guess I have an equally long and painstaking troubleshooting process to look forward to.

Edit: Oh that wasn’t so bad, I just forgot to toggle the mechanism from upper-frame assembly back to lower-frame assembly. Everything works out after that. Praise the gods of synchronisation!

Another edit: "Well, Crew Quarters seems to be just about done, I just need to get the beds in there and that shouldn’t be a problem as long as there isn’t some inane, pointless requirement for the orientation of the headrest…

…are you fucking kidding me."

I’m building a spaceship, damn it, not a hotel!

[Jurple]

“I’m an engineer, Jim, not an interior decorator!”

[Jurple]

Struggling with Relay Satellite. Rotating a one-block-wide pillar is already nontrivial. Well, not the actual rotation, but getting the pillar off the rotator, while none of the blocks involved blocked the rotation, was.

And now I have the central pillar conveyed neatly between the two arms, and can’t get the welding right. I might be able to set up a welding station in the air, fed by lifters, but that is going to be finicky.

Or I could go the route of welding on the two arms separately. That will involve more rotating, but it might be more manageable in the end, as I wouldn’t be rotating pillars but subassemblies that I could then grab with conveyors. Hm.

[LTK]

I ran into that problem as well. I created an expedient solution when I realised that rotators will rotate components even if the rotator itself is in the process of being moved. So I made this:

https://dl.dropboxusercontent.com/u/20662983/gif/2016-12-22-14-45-48.gif

I welded the solar panels on by dropping them into place on a weld arc, a mechanism which is governed by seven pushers and six sensors for just three components. The hard part was assembling the product while still leaving enough room for it to move out of the mechanism. My solution is hardly optimal, but at least it works with a minimum of chaos.

Anyway, yesterday I completed Plasma Engine, which took a long time but all things considered it was easier than it looked. Only two missions left until I actually finish a Zachtronics game!