Nameless Voice on 21/6/2010 at 22:39
Quote Posted by Al_B
Nice - just done that one myself. It's not quite the end, though.
Nope, there are three more bonus levels. I still have to attempt the last one.
After that, I was thinking of posting up a list of the number of parts I used for each solution.
demagogue on 21/6/2010 at 22:55
Sheesh, they reward you with more levels.... :erg:
I can't figure out how to let the system know it's reached the last dot (or when it reaches the first printed dot), since as far as it knows it's just another arbitrary dot (e.g., in robostilts, start green, end yellow). Unless (this just occurring to me) you let it run out then start computing after the first gray arrow... Is that what you have to do? But then you'd have to build in a memory for the arrangement of all dots it's already gone through! Hmm, deserves a test, at least.
Edit: Or you could print an exit-key-code right out of the gates. (Tests.) It worked! Not 100% robust, but it passes.
Edit2: WTF!!! How does one fucking blue count as beginning *and* ending in the same color?! Grrr ! !
Edit3: Well, at least it's easy to make a special case just for it.
Nameless Voice on 21/6/2010 at 23:43
I use a green dot to indicate the end of the string, myself. One green writer at the very start, then a green reader if you need to go through it more than once.
You could just as easily use yellow instead of green, it makes no real difference.
demagogue on 22/6/2010 at 02:11
I was actually complaining more about that being part of the rule. If you interpret the rule more like a sales contract than a logician (where separate terms imply additivity and should specifiy when they don't), then one blue would count as a "fail", and my machine is perfectly sound for that. It ended up being a wrong interpretation of course, but rather than rebuild the whole thing, it was easier just to rig it to pass that one case that would otherwise , under its logic (properly), have been a fail. So the design is sound, just for a slightly different (& wrong) interpretation. On the plus side, anyway, my machine passes in about 1/2 the time as the solution you posted (recursions are tighter; 2 more pieces, though), so I can't feel too bad about it. :)
Nameless Voice on 22/6/2010 at 04:18
My original machine looked mostly like dvrabel's. After some optimisation, though, I managed it in 19 parts and 0:44 seconds. >.]
Here are some of my statistics. How did others compare?
RoboToasters: 3 parts
RoboCoffee: 3 parts
RoboLamp: 8 parts
RoboFish: 4 parts
RoboBugs: 11 parts
RoboCats: 11 parts
RC Cars: 8 parts
RoboCars: 7 parts
RoboStilts: 9 parts
RoboBears: 15 parts
MiliDogs: 8 parts
Androids: 12 parts
RoboRockets: 7 parts
RoboPlanes: 7 parts
RoboMecha: 19 parts
Soliders: 7 parts
RoboTanks: 27 parts
Robo-Children: 21 parts
Police: 52 parts
Teachers: 20 parts
Rocket Planes: 17 parts
Officers: 39 parts
RoboSpies: 8 parts
Judiciary: 67 parts
Politicians: 30 parts
Academics: 31 parts
Generals: 39 parts
Engineers: 25 parts
____________________
Seraphim: 31 parts
<s>Ophanim: 132 parts</s> (The Malevolence Machine does not approve)
Metatron: NOT COMPLETED
Ones marked in red are ones that I haven't gone back and optimised yet.
Shadowcat on 22/6/2010 at 07:38
Is there any way to export the stats, or does one just have to check it all manually?
I'm sure some of my early solutions are horrendous. I had vaguely intended to go back and optimise them a bit, now that I have the hang of the game, but I still have the dreaded Metatron to complete. I looked at it yesterday, but was struggling to fit things into the available space. I might have another crack at it tonight and see if I can't sort it out!
Nameless Voice on 22/6/2010 at 09:26
I ran every single machine manually and noted down the stats.
Shadowcat on 22/6/2010 at 12:57
God damn it; I thought I was done. It passed their tests over and over again, but I just threw some of my own input at it to make absolutely certain, and succeeded in breaking it. D'oh.
Shadowcat on 22/6/2010 at 13:50
W00t. Properly sorted, now :)
It turned out that I'd dealt with an earlier issue in a very iffy manner, but now it's entirely robust -- not to mention noticeable smaller and cleaner looking than before, due to having had a bunch of conveyors ripped out :) -- and it's passing everything I've thrown at it.
[ATTACH]518[/ATTACH]
I can trivially scrunch up a couple of sections and reduce that to 55 conveyors (116 parts). I rather doubt that I'll be doing anything more to this one, however.
