The Nineteenth Byte

@MartinBüttner Hey thank you! I already knew that article, and I have a starting implementation based on that too, but I want to see what people can come up with. I'm going to put reference to that article too as example, but I'll probably make it to be more about some kind of terrain generation based on continental drift or something like that... We'll see how it comes out

PhiNotPi

2:11 AM

Okay, so I'm still floating around on the idea of doing a KOTH.

I think I don't like the mechanic of locking submissions from edits.

It has some interesting features, strategy-wise, but as a participant I feel out of the game after submitting my bot.

Calvin's Hobbies

The edit rule was only because of the team thing, to prevent older answers from copying the behavior of their more optimized teammates.

I'm not sure how useful it really was and I wouldn't add the restriction for an all-vs-all koth

PhiNotPi

Working together to build bots is definitely the core of this type of game.

Calvin's Hobbies

Ah, ok

PhiNotPi

Slightly unrelated, I feel bad for those three or four people who have completely broken bots as their final submission.

I think Stack Snippet KOTHs are the future, though.

Calvin's Hobbies

Heh, yeah. Though there are a surprising number of low-rep answerers. I wonder if any are sock puppets.

PhiNotPi

2:26 AM

I need to figure out what I'm doing for a KOTH next.

The team-based thing is cool.

But doing even/odd would result in the exact same teams.

feersum

better decide how many binary digits of id you can use before it becomes too biased!

PhiNotPi

Look at the best scorers of Red vs Blue and find a way to get us all on the same team?

Idea for an implementation of a diplomatic procedure:

There will be two types of communications:

"I will vote for X" and "I will vote with X"

I think there can be as few of two rounds of negotiations, although this can be increased to whatever number you want.

Each round, players can send up to one message to each player. They are allowed to say conflicting things to different players. For the second half of the first round, they receive all communications.

This would be repeated to make a second or third round.

Then, players vote for a single option. The option with the most votes wins (or loses depending on what type of game it is). Tiebreakers can be either random or determined by giving each bot a turn to be the tiebreaker.

At the end of voting, you can reveal all the votes so that bots can see when promises are broken, or you could keep that a secret.

Calvin's Hobbies

3:11 AM

Whuh?

PhiNotPi

3:30 AM

@Calvin'sHobbies Yes, that's about a completely different topic.

I was talking about this: chat.stackexchange.com/transcript/message/20929249#20929249

PhiNotPi

3:52 AM

It was probably confusing that I was talking about two entirely different things.

luser droog

4:37 AM

Hey, everyone. I've written up my idea for multiple numeric types in my APL interpreter. groups.google.com/d/topic/comp.lang.apl/KSwo40-eLKI/discussion

Basically, it's fancy tagged-integers. Small integers encode as themselves (immediate value), larger integers and doubles encode as two array indices to lookup the value in tables.

It all appears very straightforward to me now, but of course it took a lot of time to make it that simple. :)

I now have several features which should make inca3 very competitive with J. Extra parens at the edges of expressions can be omitted: '3)+(4' <==> '(3)+(4)'. Single-letter variables may be strung together in expressions with no intervening whitespace: 'a=2' 'b=+' 'c=3' 'abc' <==> '2+3'. And APL functions have single-byte encodings.

And the first weekend of the Musical, The Mystery of Edwin Drood went well (I'm playing the acoustic Bass). I've made fewer mistakes each night, but I'm missing more and more cues as I get more absorbed in the story. :) Plus the cute brunette in the bustier is not helping my concentration.

Transcript for