Thanks. Good to know that another web based poker service is using it. Displaying the hands is not so much of an issue, TBH, I would rather do that myself. A lot of my plans for PkrSess require a fairly custom hand display, so if the FPDB will parse them into a format I can write to a DB then im sorted.

Btw, pokeit looks awesome

99% sure thats a no. You can only review on the app store if you purchased via the app store. Otherwise it would be pretty easy to fake reviews.

That was the point: I *can* doesn't mean I *should*.

If that comment didn't scare the daylights out of everyone here, I'd be seriously scared.

Yes, this is why I might build it at home for fun, but reference the scary comment.

this is kind of badass: http://www.cnn.com/2012/05/29/tech/w...ame/index.html

Been playing around with the FPDB, its actually pretty awesome. The hand converter is quite modular can be called from an external script with no modification. You just pass it the location of the file to parse and it returns a list of 'hand' objects that can be converted into JSON and returned to the calling script.

it's python, right? is your website written in python or are you calling the python code from something else?

FPDB is written in python, my website is written in JavaScript (using NodeJS).

What I am going to do is create a thin python wrapper around the FPDB hand converter which receives the hand data as an input, sends it to the FPDB hand converter, processes the output to JSON and sends it back to the website.

I can either call the wrapper directly from the JavaScript, or, the python wrapper can operate as an independent application and listen on a designated port. I prefer the latter as it gives a bit more control over how hands are being processed and also separates the hand converter from the website (e.g. the hand converter could run on a different machine).

Edit: The third option is to expose it as a web service. I like this option as I have not seen any other hand converters that provide a JSON output.

Any of you read this on an alternative to A/B testing? (Posted on HN) It was an 'aha' moment for me and is a really great read!
http://stevehanov.ca/blog/index.php?id=132

Surprisingly for me a lot of HN comments just didn't get it, a lot of people criticise it for this and that without realising the same criticisms apply to A/B testing.

Weird, I thought the n-armed bandit problem was pretty well known and most CS students would come across it as some time during their studies?

Juk

A/B testing is crap.
epsilon-greedy is probably what I'd use as well, simple enough.

The multi-armed bandit picture is pretty sweet.

People would probably be better off improving content than worrying over optimizing clickthrough rates given fixed content.

Cool, thanks for that link Gull

Global state is a legitimate requirement though. It's just that doing it through static variables is dangerous for reasons like race conditions, parallel design, etc. We always implement it through an interface, and provide two impls, one through a local singleton pattern and another through some sort of service (db, keystore, cache, whatever). seems to work well.

can you give any more details? xD

thank you sir, i need to read HN more..

I also subscribe to Hacker Monthly, its a monthly print magazine (tho I subscribe to the digital version) of Hacker News articles

http://hackermonthly.com/

got my samsung 7chromos today and it looks very very very sweeet!. thanks for the suggestions guys

If this is out of line plz dont give me another infraction but... does anyone know how to rip audio from a ROM? (more specifically, a M.A.M.E. ROM). I've spent about an hour looking for a sound and I cant find it so this is the only thing I can think of. I'm sure its possible but I'm afraid it might be harder than it sounds.

Well "M.A.M.E. ROM" is a very vague concept. MAME supports many kinds of hardware, which suggests many different audio formats.

There's going to be no general answer to this, even among the hardware types that use standard encoded audio that may or may not be compressed in memory.

probably easiest to just try and capture it on playback

But then I have sounds mixed together >.<

one place to start would be the MAME source code, could be a fun project

Yeah, I think this is prolly your best bet although I think it'll be pretty hard going as they just simulate the hardware and the actual software on the ROM could be using some weird encoding/compression.

Are you trying to get something akin to a tracker track or just the samples themselves?

Juk

Had a quick look and I think your best bet might be to see if you can hack the source to dump the raw sample here:

http://mamedev.org/source/src/emu/sound/samples.h.html

See the start_raw() function.

It depends if there is software mixing going on in the game or not.

If that proves to be a dead end, then see if you can find out what hardware your MAME ROM is running on:

http://mamedev.org/source/src/emu/sound/index.html

Juk

Yeah it is likely encoded in MIDI in the ROM in some way

Thanks guys, now I will be busy again xD

Judt read about this in NY-Times, supposedly 20mb in size, quite extraordinary it went undetected for so long