Let's all play Spacechem

Time is up - as mentioned earlier, the final writeup will be a couple days late. I've checked out all the solutions and am putting together a video.

whoops, i didn't have time to work on this one as i've had family in town for the past couple weeks. oh well!

the first idea i had for a puzzle was a little too complex i think, so maybe i'll get it back out later. here's one that should be fun for people who haven't gotten too far into the game and hopefully not super obvious and boring for people who are way better than me? i guess we'll see if i get two identical solutions or something

#5: Nucleosynthesis

Only moments after the Big Bang, the first atoms were formed, making the basis for all of the stars in the universe. Now, you, too, can experience the magic of creation and the wonders of the cosmos, in two easy steps!



You can import with this code:
In case anyone is unfamiliar, the strange object laying partly in the upper output is a fusion device. You activate it by passing a waldo over the "FUSE" command, similar to bonding. When atoms are on both sides of the object, this command combines the two atoms, creating a single new atom on the right side. The resulting atom is the atom whose atomic number is equal to the sum of the two components. In this case, the product of Helium is created by fusing two of the Hydrogen atoms together. You don't have to make a waldo drop the "receiving" atom before fusing, although I don't think this fact is relevant to any solutions of this specific puzzle.

Solution Criteria: Least cycles, with least symbols as a tiebreaker.

Solution Deadline: Solutions should be sent to me via PM by 11:59 PM EDT on Wednesday, July 15, 2015.

Enjoy!

(Also, the more physics/chemistry/astronomy inclined may notice that this puzzle produces a 2:1 proportion of H:He when the true ratio is roughly 3:1 instead, so here's a version which calls for 10 He and 30 H if anyone finds the original version easy and is interested in playing with it; I haven't been able to figure out a satisfying solution since the need for a fifth H per output cycle breaks most of the obvious symmetries...)

Challenge 4 Results

Thanks to everyone who submitted a solution (or might still take a shot at this after the official deadline)! I think we got some interesting variety.



Results:


Challenge #5 is already up!

Nice! I didn't get a chance to optimize after you dropped your hint (nor very much before, really), so I'm pleasantly surprised by my rank this time.

edit: spineshark's puzzle seemed too simple -- and then I noticed that there are 20 H outputs. If you could switch the position of the two outputs, this would be much easier

As it stands, after correcting for this, I got an even worse time, but one that's compatible with the 3:1 version of the puzzle.

Hmm, of course there's a solution that edges mine out slightly by cycles and is more elegant symbol-wise at all. For reference, here's my first submission which did not exploit the multi-bondable Chlorine, before I updated it in response to DFalcon's hint.

Wow. I feel really silly for not taking advantage of putting the bonders where inputs appear. Oh well.

At least I got my line to nowhere in. I love that.

to clarify, i think any practical solution to the 20 version will also complete the 30 version; i can think of a way to make it crash for the first time after 20 outputs but before 30, but it's pretty difficult and extremely slow if it even works. what i was trying to get at is that, if you *do* just reuse the same solution, then you end up making 5 extra heliums, which seems like an obvious inefficiency. but i'm not actually sure that it is, either
i of course playtested this the other way around (and also with two hydrogen inputs, which was way, way too simple) and thought this one was a bit trickier/more interesting for people

Well, yes, any practical solution to a 1:1 will also work with any other ratio. But what I meant in this case is that I'm pretty sure that the time I got with my second attempt would be the same even if you changed it to 30 H.

Before noticing the outputs, I had red delivering He and blue delivering H. My second attempt had red deliver 1 He:1 H while leaving blue mostly unchanged, meaning that each complete loop of the program delivers 1 He and 3 H. It was actually only about 10 cycles slower, and it's got some inefficiency, so I could probably tweak it for better results. I think I have another approach I want to try, though.

After puzzlemastering I do appreciate the ability to duplicate in custom-puzzle mode even more; besides keeping all the submissions separate, I tried out a bunch of variants.

Challenge #5 makes me want flip-flops that use ratios other than 1:1. By the time I work up the ratio I really want, I've spent too many cycles just moving around in the flip-flops for it to be worthwhile.

While we wait for spineshark's results to go up, here's the next challenge.

#6: Atomic Arithmetic



With oxygen (atomic no. 8) as an input, create lithium (atomic no. 3) and boron (atomic no. 5).

This puzzle requires you to use both fusion and fission. We saw fusion in spineshark's challenge. Fission is the opposite: give it an atom, and it will divide it in half. One oxygen atom becomes two beryllium atoms, one beryllium becomes two heliums, one helium becomes two hydrogens. Odd-numbered atoms divide as closely in half as possible, so lithium becomes helium and hydrogen, in that order. (You probably won't need to split an odd-numbered atom in this puzzle, but far be it from me to tell you how to compose your solution.)

While this puzzle gives you an atomic input that's equal to the sum of the outputs, remember that it's not required for you to perform all the operations on a single input. It might be more efficient to work with two oxygens as you try to synthesize your output atoms. Then again, it might not. I haven't drilled down into this puzzle too deeply, so take this with a grain of salt.

As usual, this is a lowest cycles competition.

i moved the deadline to wednesday night, so there's a bit longer. i've already gotten a whole lot of resubmissions though

Mogri, when's the deadline on your puzzle? And do you want them sent to you via PM this time?

Ten days out, so... next Thursday? I won't be especially strict on it.

I don't care about PM versus in-thread, but judging by how everyone else has opted for PMs, that seems to be the consensus preference.

On the off chance that you've forgotten about this challenge, consider this your reminder. I'll take your submissions through Friday, so don't feel overly rushed.

I only have two entrants so far. Is this puzzle harder than I gave it credit for?

I haven't had time to play SpaceChem recently.

Ditto. Work is blowing me up lately.

I do plan to tackle all the challenges at some point, so in order to avoid spoilers I haven't been hanging out in this thread much.

I'm still good for my next turn as puzzlemaster, though.

so, i don't think this is conceptually too crazy, but it does seem to require some proficiency with the flip flops (on top of all the other tools) simply due to how crowded the area around the lasers is by necessity. i have a solution and unless i can think of other improvements really soon i'm just sending it in tonight since i won't have any more time tomorrow either

By the way, spineshark, what were the results for your puzzle?

i've had family in town again all week (uhh, i didn't know they were coming until they were almost here), so tonight was the first time i've had a chance to make any time for this at all. sometime this weekend i'll try to record video of all the solutions.

failing that, i guess we'll probably end up with something more like the first time. anyway! i think 5 people sent in a total of 12 different solutions? yours came in last, unfortunately

welp

And I thought mine was pretty decent, too.

SpaceChem is hard.

to be fair, it took me a lot of tries to get a better solution than yours (not that i was counting myself). my goal was to make something straightforward in the most obnoxious way possible

Just wanted to make a post saying that I've got my next puzzle ready to go and will be posting it tomorrow night. I think it's got a good balance of difficulty to it. Hopefully we can see some puzzle results before then!

ok, i'm not sure of a good way to get the kind of video i want (the programs i was using all turned out incredibly poor quality) so i guess i'm just posting everyone's pictures, oh well! i'll do better next time, sorry!

(at least this way i don't spoil my probable next puzzle)

in fifth place, aturtledoesbite!

276 cycles, 29 symbols. this one has some strong features that show up in some of the other solutions, like the use of simultaneous commands to fuse the helium in just two cycles and the use of smaller loops to pass off one of the hydrogens to the bottom.

in fourth place, Mogri!

251 cycles, 32 symbols. this one also has some interesting features, like the two flip-flop chains and that the two waldos clear each others' outputs.

in third place, Red Hedgehog!

at 247 cycles, this barely edges out Mogri. the blue flip-flop chain causes some really interesting behavior and creates an unusual sequence of heliums and hydrogens, while the red waldo just runs around dropping off hydrogens the whole time. really cool stuff!

in second place, DFalcon!

this is a really charming solution with extensive parallelism; the two waldos run extremely similar routes such that they stay perfectly in sync to get inputs and fuses at the right times for each other. each produces a sequence of one helium first, followed by transporting two hydrogens via the flip-flops. it's an extremely clever use of the second one to reverse course; i really love this one!

but this challenge's winner is MrBlarney!

216 cycles. i'd like to point out again that this solution is in some ways really really similar to turtle's, with the small blue loop being used to deliver the hydrogen the rest of the way from the red waldo while simultaneously fusing the helium atom. i also love that this one produces the helium last, which saves just a few cycles at the end (because it doesn't need to be moved as far as the hydrogen), it's a really nice little detail that it seems many of the other solutions struggled with.

great work from everyone, i really enjoyed seeing these, even for such a simple concept!

I'm afraid I'll have to go the screenshot route as well.



3rd: spineshark
326 cycles, 47 symbols

Though he's in last place by default, spineshark's solution is about what I was expecting from this puzzle: break oxygen into two borons, break a boron into heliums (helia?), break a helium into hydrogens, then fuse the hydrogens into the boron and the helium to get your outputs.




2nd: DFalcon
172 cycles, 65 symbols

DFalcon creates magnesium (20) and calcium (12) as intermediates. He then splits each atom twice to create the outputs. It's a clever solution that might be the fastest of our group if it were able to be more efficient.




1st: MrBlarney
153 cycles, 32 symbols

MrBlarney's solution makes use of what the SpaceChem community calls a "catalyst." In short, a catalyst is something that's reused as a step for creating multiple solution molecules but is not part of the solution itself.

The blue waldo takes the inital oxygen and splits it, depositing one of the beryllium atoms to the side where it remains unused for the rest of the reaction. The other one is split into heliums, and blue takes one of those and deposits it, too, never to be touched again. The other helium is the catalyst: it gets combined with oxygen to create neon (10). The neon is split into borons, one of which is used as an output. The other boron is split into lithium and helium. The lithium is the other output, and the helium is reused as a catalyst.

If that's confusing, maybe putting our Atomic Arithmetic in number form will make it easier to read:
1. 8 / 2 => 4 + 4*
2. 4 / 2 => 2 + 2*
3. 8 + 2 => 10
4. 10 / 2 => 5 + 5**
5. 5 / 2 => 3** + 2
Repeat steps 3-5.

*Discarded
**Output


While DFalcon's solution might have been fastest if it weren't for space constraints, MrBlarney's catalyst approach achieves a similar result with a far smaller symbol count. Certain puzzles are difficult or impossible without a catalyst, and it's rare to see a catalyst solution to a puzzle with an obvious non-catalyst solution. Congratulations to MrBlarney for both the win and the unconventional approach!

Whew, finally got some top marks. It's a good thing I went back to my Nucleosynthesis solution before the deadline to make another submission. My previous solution had a really nifty flip-flop timing and a tiny symbol count, but would have only gotten second place, just ahead of Red Hedgehog.

Time for me to finish preparing my puzzle; it'll go up on the hour.

Since some control flow was not meant to occupy two dimensions, I'm glad there was a better solution on Mogri's than my 65-symbol monstrosity. So thanks, MrBlarney, for making the world a little brighter.

(the handoff solution was cool too)

#7: 3-Hexene

I've noticed that we haven't had a least-symbols puzzle yet. One puzzle type that lends itself well to least-symbols is polymer puzzles. Repeating the same actions multiple times, then breaking from that pattern once the required length has been achieved - it's a simple structure. But then I decided to add a little twist to it:


This puzzle introduces a new reactor element: the Sensor. There's also a waldo symbol associated with the sensor, the Sense Atom. Normally, the sensor is inactive, as is the Sense Atom symbol. But if an atom hovers over the sensor element, all Sense Atom symbols with a matching chemical element become active, and waldos that move over the symbol will be redirected in the direction pointed to by the Sense Atom symbol.

You can import this into ResearchNet with the following code:
Solution Criteria: Least symbols, tiebreaker least cycles.

Solution Deadline: Solutions are due to me via PM by Friday, August 7, 12:00am PDT / 3:00am EDT / 7:00am UTC. Since I've already had time to test and look over the puzzle and will need to verify solutions as they come in, I won't be counting myself in the rankings.

One week into the challenge and I've got all of... one solution. Still a few more days to put something together!

i actually don't know how to do this...haha :\

i'll try and try again tonight

Hmm, I had hoped that I didn't make an overly difficult puzzle. It felt to me like it was about the same difficulty as Atomic Arithmetic (at least, if you approached it from the single-oxygen breakdown solution).