locator.pheeble.net

[Qara-Xuan Zenith]

We're missing either a direction vector length value, or a way to GET a direction vector length value, or some hidden hint about how to bloody USE the locator...
I know we're missing something. But I don't know what it is we're missing, until we find it. Useful, I know.

Wait, actually... we got this locator from Peter, and very shortly afterward, he gave us another clue. Maybe we should get back to solving that "silent music" riddle; it will lead to either an echo or more info on THIS, I'm thinking.

[The Finch]

What is the Silent Music riddle again?

[Qara-Xuan Zenith]

What is the Silent Music riddle again?

It's here.

[The Finch]

What is the Silent Music riddle again?

It's here.

Thanks. I've guessed a Phonograph.

[WackyMeetsPractical]

There was a brief discussion in the chat room. Here's some of the talking points:
1. Some compass points may be decoys.
2. After recoloring the map, some lines seem to point in the same general direction, but do not intersect at a single point.
3. We could need Pan's or Cheshire's coordinates.
4. We would like to know if two people from the same IP address could get the same compass points.
5. If we had someone in Pittsburgh, we could use their compass points to somehow find the echo.

[S_o_S]

Guys. Seriously.
There are two freaking hundreds of us, and we still only have 12 records, including me. We'll need much more than that to pinpoint the target. I'm not asking you to go meta-knighting - adding yourself would only take a couple of minutes. And now that the coordinates are rounded to 1 digit after the point, it's not even breaches your privacy... too much.

I'm pretty sure I sent you mine, right? Are they up?

[Dryunya]

Guys, are you serious? There is a link in the OP post! Everyone can read the document (at least I think they can).
(Also, yes, you are up.)

[JackAlsworth]

Interesting find; Inukai and I live about 25 miles from each other (his coordinates on the document are wrong, I think; it should be a negative 122), and we have identical compass points.

[S_o_S]

Guys, are you serious? There is a link in the OP post! Everyone can read the document (at least I think they can).
(Also, yes, you are up.)

... Derp

[Dryunya]

Interesting find; Inukai and I live about 25 miles from each other (his coordinates on the document are wrong, I think; it should be a negative 122), and we have identical compass points.

Oh. Indeed. I'll correct it. I guess I need to double-check the ones we have now - talk about derp.

[Guyshane]

Right figured out how to do this. working now

[Scarab]

Has somebody already added me? I think I sent you my co-ords because I was being a useless noob and couldn't work out how to do it. :/

You were added to the map, but now I can't get your exact coordinates... because I can't figure out how it's done on google maps.
If you were to check your location again, I'd be very grateful.

Sure. The map I recieved was this: http://i210.photobucket.com/albums/bb10 ... ba68d2.png

(Sorry for taking so long and being generally useless at co-ordinates.)

[Dryunya]

No worries. We are at a dead end anyway.

UPD: A good idea originated at the chat - we ask John Silver! It is too appropriate to not get any hints from him.

[Sicon112]

Great idea, Dryu! I'll also get Mimsy to test the page to make sure ours are identical.

[Pixelmage]

Continuing efforts! Source code analysis is a possibility so I bring you all the source code.

Main Page
Sub Script - Unrelated

Updated. Thanks to Sophira the second link is no longer an unreadable abomination! Many many many thanks!
Also, she confirmed that it is unrelated, it's a generic code to allow the use of JavaScript, so whatever anaysis we can do should be focused on the main page code.

[Dryunya]

I've said it in chat, and I'll say it here: The *.js files are useless. The location data are sent with this line:

Code: Select all

sendRequest("http://locator.pheeble.twwf.info/locate", params, returnSub);

It is processed by ajax.js, but, like the json-minify.js, it's mostly standard stuff. The thing is, the data are sent to the server, transformed somehow, and then returned. We don't know (and cannot know, unless we hack the server) what happens to them.

We must analyze the data. But I've given up on that long ago. I've done what I could, and I'm waiting for new ideas.

To recap, I tried simple plotting the compasses on the map (they don't really converge in a single location) and calculating a mean vector for every set of directions (makes even less sense). The links to the maps are somewhere in the thread.

Somebody suggested that the directions may be actual directions - that is, go that way some distance, and then go some distance the other way, etc. In that case, for the sake of simplicity, I assume that the distance is the same every time. In that case it's possible to guess that distance. I'm not sure how to automate the process, though.

UPD: WAIT
SHIT
WE SHOULD JUST MAKE A GODDAMN EQUATION

UPD2: ok, elaboration. It should be a system of vector equations. D[1..7] are direction vectors, P is the target point, d is the step distance.
(D1_1 + D1_2 + ... D1_7) * d = P
(D2_1 + D2_2 + ... D2_7) * d = P
...
(Dn_1 + Dn_2 + ... Dn_7) * d = P

n is the number of data entries.

Now that I think about, though, this approach makes no sense, because, for example, the location that's too close can't be reached with steps like these. Maybe the right half should check the distance to a spot instead. I'm not sure. I won't probably try it yet.

[Blurred_9L]

Finally got it to work.

I'll add myself in just a minute. I need to understand the file format

I hadn't noticed but it seems that all but one point NE. That's weird.

[Pixelmage]

How did you get it to work?

[Blurred_9L]

I used my school's internet connection

[Qara-Xuan Zenith]

I just had an idea about the locator.

It was sent to Peter, who uses the internet on his phone.
Has anyone with the kind of phone and data plan that this would work for tried accessing the site on their phone?
I can come up with two possibilities why that would be good:
1. the interface might be different when you get it on a phone; for example, you might see only ONE compass.
2. it's possible that it DOES act as a sort of GPS. What would happen if you started following the first compass point, until it changed on your phone's screen? This could be our answer to the missing direction vector value.

(I could try checking this theory on my phone at some point, but I can only get internet on wifi, not on the mobile network, so step 2 is useless for me.)

[narrativedilettante]

I'll get my sister to visit it on her phone this weekend and report the results.

[Pixelmage]

I tried accessing it on my tablet. But it does not locate me on it either. And it is a non-3G model, meaning I can't use it outside.

[Qara-Xuan Zenith]

Update: I checked it on my phone, but just got the sailor who turns up when it can't get your location.
This might be because my phone's internet doesn't go through the mobile networks, though, and therefore probably isn't able to broadcast my location.
So it's probably still worth checking if you've got a phone that would work with the mobile networks.

[Dryunya]

My phone has 3g (I'm online through it now, lol), but its browser is crappy, so the site doesn't work for me.

Meanwhile, about that equation of mine.
It won't work. The reason is that it is basically the same as my previous approach: it involves finding the sum of the direction vectors. At that point I tried to find the intersections by hand.
Following the vectors has the same result, as anyone who remembers the vector addition will say. So I get a set of vectors, which I try to scale until they all converge in a single point (I actually used a 30x30 square). One doesn't have to be a genuis to understand that it's finding that same intersection point, except analytically instead of graphically.

I probably don't make much sense now.
tl;dr: The "follow every direction for a fixed distance" approach doesn't work, either (under the assumption that this "fixed distance" is the same for every set of vectors - if it's not, there's not enough data to solve the equation system).

Also, I totally wrecked my "mean vector" map. While trying. It was useless anyway, though.
WE NEED MORE BRAINPOWER HERE, PEOPLE! I'M SOOO STUCK!

[Dryunya]

Ok. Mr. A said that there WAS a correct suggestion in this thread, and we didn't take it. I'll compile our suggestions, gimme a minute.
I'll also include the ones who came up with the suggestion first (in the thread).

1. Plot all points on the map and see where the lines converge (Me; got a messy map and nothing more).
2. Just examine the points that the locator returned (Me; didn't check it - it's nonsense).
3. Walk some distance in one direction, then some other distance in the other, etc (Sicon; didn't follow as we don't know the distances).
4. Trying to intersect the lines after summing all the direction vectors (me; didn't seem to work).
5. Plotting the lines on a goddamn globe (eyebones; didn't really try yet).
6. It's a red herring (Qara).
7. The coordinates are not important, only the directions are, somehow (Qara).
8. Try direction sets on the echoes with known locations (Qara; hey, didn't try that one!)
9. Plot the lines end-to-end and see if they make some shape (Qara; they don't).
10. We must apply Pan's "silent music" riddle (Qara).
11. Some compass points may be decoys (Wacky).
12. We could need Pan's or Cheshire's coordinates (Wacky).
13. Make someone in Pittsburgh try it (Wacky).
14. Ask John Silver (Me; he ignored the tweet).
15. №3, assuming that the distances are the same everywhere. (Me) Busted by math. Turned out it's the same as №4.
16. Using the locator as GPS on a phone. (Qara, I think)

ONE OF THESE IS CORRECT.

EDIT: The more likely ones are 5, 8, 13 and 16.