Spacechem

Nov. 1st, 2011 02:21 pm
jack: (Default)
[personal profile] jack
Space Chem is really awesome. Mobbsy first recommended it to me.

There are two little robot arms that manouver atoms around a grid to connect them together into molecules, which you control by putting instructions (turn, pick up, bond, etc) onto the screen.

Each level you have to produce something specific (eg. you get CH4 and H2O and have to make CO2 and H2) and in later levels you can connect several reactors together to make more complicated things.

At first, you just have move, pick up, drop, bond, input and output instructions, but later on you get flip-flops (which can be used to build more complicated loops), sensors (which branch to a different route if the corresponding square has a certain sort of atom in), fusers (which fuse two atoms into one with the combined atomic number) and so on.

The chemistry is at sort of GCSE level: the molecules are real molecules, with the correct chemical composition, but all the molecules are represented schematicly with right-angle single and double bonds, of up to a set number of bonds per molecule, rather than a more realistic 3d electron-shell based simulation. Which, um, is about right for me, although I imagine it may annoy some people.

However, it's a surprisingly really good introduction to programming! The constraints you're under are somewhat artificial (any real program would have variables and conditionals sooner), but the feeling of building a complicated system with the pieces given is really, realistic :)

Eventually I realised synchronising the two manipulators was almost too realistic, it started to feel like work :) It's not just a matter of timing them to make them accurate, because each manipulator can be slowed down if a molecule is slow leaving the reactor or a new molecule is slow entering, so they have to be synchronised with special instructions, and what works on one page where the inputs come in predictably fails miserably where one of the inputs comes from another part of the system and may be slow.

It's a spiritual successor to the Codex of Alchemical Engineering which was a similar idea but based on manipulating classical elements in a pseudo-alchemical way. Which was really interesting, but often annoyingly clunky -- manipulators would tesselate badly and often collide, and the rules of how the elements were modified felt a little arbitrary. Whereas Space Chem massively updates the interface to make it feel "swooshy" and it's a lot easier to use: I think it's both more powerful and less clunky.

Date: 2011-11-01 03:06 pm (UTC)
ptc24: (Default)
From: [personal profile] ptc24
A chemist's view: you have to laugh a bit about all of these very reactive radicals that you're passing around, and the way that reactions break or make bonds but rarely do both. However there's something about the synthesis planning - once you get past the obvious pipelines and into cases where you actually have to make non-obvious decisions about what reaction a reactor is going to do - that feels weirdly compelling even though all of the details are wrong.

Also, can I just mention a nuclear warhead with two whole plutonium atoms in! I suppose you have to have two otherwise you couldn't have a chain reaction.

Anyway, it's a good game.

Date: 2011-11-01 05:10 pm (UTC)
From: [identity profile] alextfish.livejournal.com
Yay. I did enjoy it a lot. It did indeed feel like a nice successor to Codex of Alchemical Engineering.

I seem to have quite a slow-and-steady programming style, judging by the graphs it gives you after each level; my number of cycles spent tends to be higher than 80% of those logged. I think I use the sync instructions quite a lot. I guess that's what they get from a professional quality engineer :P

I didn't mind the right-angled approximations to the molecules; what did irritate me is the way that all bonds are treated as covalent, so it says "Sodium and chlorine and hydrogen each take one bond, so you can make HCl or NaCl or H2 or Cl2... or Na2 or NaH. But not Na3, no way."

Date: 2011-11-01 05:33 pm (UTC)
ptc24: (Default)
From: [personal profile] ptc24
NaH is a perfectly sensible compound, as sensible (and as ionic) as NaCl, it got mentioned a fair bit as a reagent in synthetic organic chemistry lectures.

Na2 - well, google for "sodium dimer". It's out there, and seems to have something to do with lasers.