After looking some more then it seems you are also interested in the squares marked 'x'?

The region labelling algorithm solves your problem of deciding what is inside/outside of the level (even the isolated regions in the first example), but how do you want to deal with the 'x' - are they supposed to be labelled 'x' or just left as spaces?

Juk

LOL, it's turned into another symbol now!

Juk

Oh, I think I see what the 'x' are now: are they just example squares that failed in your old "pass over in 4 directions" algorithm? If so, then the above will work fine.

One thing that would make it easier would be if you knew a single pixel that WAS going to be contained within the level: that way you could run a flood fill from that point onwards and get round the problem of isolated regions outside the level like the 'x' squares marked in your first example (is all the '$' or '.' squares going to be inside the level for sure?). If you don't know this then I think your stuck using the first algorithm and possibly just doing something like excluding all but the biggest non-border region (the "flood fill from the outside" idea wouldn't find those tiny disconnected regions you marked with 'x' in your first diagram).

Juk

I just assumed it was something to do with the US keyboard layout: on their keyboards they have a '#' (ie: a hash symbol) above the '3' on the 3 key, but we have a '£' (ie: pound symbol):



Not sure if this also effects the code-pages somehow so in some software it may well end up pound signs as hashes (pretty sure I have opened stuff and seen this happen before - possible it's even something like notepad vs wordpad and one does it correctly and the other doesn't?).

Juk

lol is that where it really came from? i've always called # the 'pound sign' (i'm form usa), but when i was living in Australia, they always called it 'hash'. i would call someone and the voice mail would say, "leave a message and then press hash", and i'm like wat?

I thought % was hash?

Thanks for the details juk, I'll look it over in the morning.. no condition to use my brain for good atm

octothorpe ftw

(srsly though i call # 'hash' in all programming contexts.)

shebang is my favorite (obv not just the # but w/e)

eta: when spelling out passwords, i say "bang" for ! and "hash" for #. confuses all but the biggest of nerds

standard ldo (she-bang too).

If someone was reading something out to me at work on the telephone and they were saying things like "123!123" I would LOL if they went "one two three bang one two three" I wouldn't have a clue what they were talking about!

so you want them to say "one two three exclamation point one two three"? that's pretty noob imo. and btw i've heard non-american folks using these terms, so you can't just cry brit here.

how do you (expect them to) pronounce '#'?

I say "Exclamation mark/Exclamation" as does everyone I know! '#' is simply hash. Honestly never ever heard the term 'Bang' used!

On telephone systems the computer voice always says "Followed by the hash key" meaning #.

We bang in Boulder, she-exclamation doesn't sound right...

Redis + Ohm is the stone cold nuts. BuhBye Mongo (until I need dymamic fields)

I ended up using the flood fill algorithm. I just found where the '@' symbol was indicating the player start position inside the walls and went from there filling the empty spaces inside with _'s, which turns out to be simpler for what I want anyways, and a lot faster than what I had been doing.

Thanks for the tip, I'm not very studied on algorithms and generally invent the wheel myself before finding out there's a better, common solution

Can I get someone to post a raw Poker Tracker hand history ITT, Please.
Any hand would do.

NP - I'd only come across it before because of it's use in script recognition (ie: the "total regions" can be used as a feature to distinguish between letters like 'B', 'O' and 'U' and the "ratio of the biggest to the smallest region" can distinguish between letters like 'A', 'P', 'D', etc).

Juk

I put Google +1 on my site, but have found out it seems to be tracking mouse moves, anyone know anything about this? I'm sure it's harmless, but it seems a bit deceptive. Their privacy policy doesn't mention this, and also there is no mention of it on the web.

http://stackoverflow.com/q/6667544/356635

Saw something clever today and wondered about its security.

A design firm we're considering wrote up a proposal for us, and sent us a link like:

www.firm.com/estimates/<20 digit long random alphanumeric string>

which linked to our companies personalized estimate. There was even a button to formally accept the estimate. Was a nice feature but seemed possibly insecure.....

Theoretically links like that are secure imo, on the condition it only exists in your inbox, the problems come when someone copies the link somewhere, a crawler finds it (this does happen every now and then) or someone intercepts the email. If it's for nothing important in particular (like a link to unsubscribe from a newsletter) it's fine imo but for other things it's a bit :S

If it's md5/sha-1 or whatever, seems fine.

mouse,

when you started trying arbitrary char*[20] in the url, did you find other quotes for other companies?

were any of them better than the quote for your company?

I probably would have run a PHP script that did md5/sha-1'd some company names or something if I could identify the string as a common hashing algorithm.

It would teach them to salt their hashes!

I think that's totally OK, assuming the hash really is generated securely.

Google does something like that for google docs, and I think google voice and probably other services.

Edit: the URL request is visible to anyone on the network between you and the design firm's website, but so is any unsecured HTTP traffic. If it's sensitive enough to need encryption, then the site should be HTTPS and the URL will also be encrypted (an eavesdropper will only see the request for the domain name www.firm.com, because you have to ask your DNS server for it).