The Nineteenth Byte

« first day (1302 days earlier) ← previous day next day → last day (3526 days later) »

12:30 AM

For people interested in the stock exchange question: pastebin.com/raw.php?i=tfCu5gnp

Well I'm certainly less interested in it after seeing those scores. I'm just glad my browser uses a monospaced font so I can easily see which long string of digits is the bigger number.

PhiNotPi

1:04 AM

@PeterTaylor That scoreboard seems pretty messed up.

githubphagocyte

1:20 AM

There's something to be said for calibrating a challenge based on some example entries before posting...

PhiNotPi

1:33 AM

The BF Jousting tournament seems interesting.

7 hours later…

New Sandboxed Posts

8:08 AM

A: Proposed Question Sandbox - Mark XIV

Ultmate Tic-Tac-Toe This will be king-of-the-hill. Controller is currently being written.. Tic-Tac-Toe is too simple a game. This time, your programs will play games of Ultimate Tic-Tac-Toe. The Rules of Ultimate Tic-Tac-Toe Ultimate Tic-Tac-Toe is played on the following board: The object...

Martin Büttner

8:30 AM

@githubphagocyte I did some doodling on the plane and came to the conclusion that there's no 1:n tile mapping for subdivisoin after all.

However I try that I get a few duplicate tiles.

I guess I can solve that by amending the vertex neighbourhood information with a direction (i.e. in which direction does tile x extend from this vertex) which shouldn't be a problem because there are only 5/10 distinct possibilities.

2 hours later…

cjfaure

10:41 AM

ludum dare is going well

cjfaure

10:51 AM

Map description is read from a file

above's:

g_md g_md g_md g_md g_md g_md
null null null null null null
null null null null null null
null null null g_el g_er null
null g_el g_md gd_l d_er null
null d_el d_md d_md d_er null

1 hour later…

Martin Büttner

11:58 AM

Re new sandbox policies. In order to keep the number of unposted challenges manageable even as the sandbox gets older, what do you think about something like "Any challenge idea that has not been updated by the OP in a month (say) is up for grabs for the community."

That would allow us to post good challenge ideas that have been abandoned by their OPs in the sandbox.

-> Less sandbox clutter
-> More challenges for main

PhiNotPi

I agree that there should be a policy in which old sandbox entries can be claimed.

Anything in a sandbox before Sandbox X should be fair game at this point.

Martin Büttner

Well yes we can decide something similar for the old sandboxes, too, but I was more talking about stuff that's then already in the new one.

But I think you have a good point that if Doorknob merges all the old sandboxes that post could basically be just like a bag of challenges where people can take some from until its empty.

PhiNotPi

I agree that we could merge all the old sandboxes into one giant grab bag.

There are a lot of old sandbox entries.

At least 100.

Martin Büttner

12:14 PM

yup

PhiNotPi

Which is a decent percentage of the number of questions on the main site (about 5%).

Peter Taylor

Some of them are abandoned because they're duplicates or just not very good and not obviously fixable.

I don't know how many of them are merely "on hold"

(Although I can certainly plead guilty with my elliptic curve cryptography question. Need to sort out some test vectors for it).

Martin Büttner

I've got two KotHs in there, but I'll move them to the new sandbox once it's up.

But I've definitely seen a lot of question that just seem to have been forgotten, or aren't ready yet but certainly fixable.

We can delete the duplicates.

Peter Taylor

On the subject of koth, since I've just set up a Github repo for the stock exchange one I've been thinking: how about editing the tag wiki to say:

Before posting a King of the Hill question it is strongly recommended that you create the test framework and one or two simple example AIs and upload them as a public repository to a site such as Github which will permit competitors to add their AIs and to easily download and build the other competitors.

Martin Büttner

12:30 PM

@PeterTaylor YES!

I'm surprised how few people are actually doing that.

PhiNotPi

12:43 PM

The recent Stock Exchange tournament needed to be tested more thoroughly.

PhiNotPi

1:03 PM

I'm writing a program to use genetic optimization on the BF jousting challenge.

Martin Büttner

Nice :) let me know how it goes!

How are you "mating" and mutating programs?

PhiNotPi

To start off, I am just going to use a template bot and use the genetic algorithm to adjust the numbers (like the size of decoys and stuff).

Martin Büttner

ah I see

PhiNotPi

It would keep the top N contestants and make random variations upon those.

The "score" of each would be its score against the other entries in the actual competition.

But random mating and stuff is probably something on my wish-list.

Martin Büttner

1:21 PM

This got reposted:

Q: Generate Lightning Forks in a 3D Grid

Objective: Overview: Write a program, that given the two inputs number of rays and ray length, will produce a procedurally random array of connected points, containing the ray number, number in sequence and coordinates of each point. Details: The lightning is in the form of rays, which are dete...