Hello again,
I just want to know if there's a SDK or just some docs planned to help about creating mods and hacks ?

If it's any help, I have extracted (semi-automatically) DTDs for the level XML files (level, scene and resource scripts); I haven't really done anything with them yet, but these may be useful to people who want to edit their own levels.

(See attachment)

I also figured out the profile data file, which I'll share if anyone has a legitimate use for it.

I've watched them (thanks for these by the way)
In fact, if I find the motivation, I may try to create a level editor. I put a may since I haven't found the motivation yet :)

I received a request for this, so I am adding (my interpretation of) the format of the pers2.dat file. Additions and corrections are welcome.

[pre]
The pers2.dat file is encrypted in the same way as the data files, except the
plaintext is padded with zero bytes. The character encoding used is unknown;
most likely it is the native character set of the operating system.

At the highest level, it contains a sequence of strings, terminated by the
sequence ",0." (without quotes). Each string is encoded with a decimal integer
signifying its length, followed by a comma, followed by the string contents.
For example, "foo" would be encoded as "3,foo" (quotes for clarity only).

The strings in the data file describe a dictionary, with alternating keys and
values. Keys in use are:

- "countrycode": values are two-letter ISO 3166-1 country codes, or EU for the
                European Union, or XX for unknown.
- "fullscreen": indicates whether the game runs in full-screen or in windowed
                mode; value is either "true" or "false"
- "mrpp": most recent player profile; value is a decimal integer between 1 and 3.
- "profile_0"/"profile_1"/"profile_2": the three player profile slots.

Obviously, the player profile strings are the most interesting. These strings
consist of a number of fields, seperated by comma's. The first four fields are:
  - player name
  - player flags (see below)
  - total play time (in seconds)
  - number of levels played/completed

The player flags field is an integer that combines several bits:
  1: online play is enabled
  2: World of Goo Corporation has been unlocked
  4: World of Goo Corporation has been destroyed
  8: the whistle has been found
  16: the Terms & Conditions have been accepted (this unlocks Deliverance)

So a player that has finished the first two chapters, for example, would
have flags set to 2|8 = 10 or 1|2|8 = 11 (here '|' indicates bitwise-or).

Then follows, for each level:
  - level id (same as the directory name in res/levels/)
  - greatest number of balls collected
  - least number of moves used
  - least time taken (in seconds)

Usually, the number of level descriptions is equal to fourth field (total
number of levels played), but sometimes there is more level data. I haven't
looked into exactly why this is; maybe these are levels that have been played
but not yet completed.

Then, the World of Goo Coporation data follows, which is a string prefixed by
an underscore character (_) which is how it can be distinguished from more
level data.

It may be empty if WoGC has not been unlocked yet. Otherwise, it contains first
the location of balls, and then the configuration of strands (connections
between balls). All data is seperated by colon characters (:).

Each ball description consists of six fields; for example:
    b:Drained:-61.88:211.98:-0.96:1.75
The fields are:
  - "b" (fixed)
  - "Drained" (fixed)
  - x-coordinate as a decimal number
  - y-coordinate as a decimal number
  - unknown decimal number
  - unknown decimal number
The unknown numbers may be horizontal/vertical walking speed or something like
that.

Each strand description consists of seven fields; for example:
    s:Drained:288,20,9.000,140.0,0
The fields are:
  - "s" (fixed)
  - "Drained" (fixed)
  - first endpoint (zero-based index of a ball in the list above)
  - second endpoint (same as above)
  - "9.0" (fixed): may be connection strength
  - decimal number: usually close to 140, may be length of strand but this can
                    also be derived from its endpoints.
  - "0" or "1": indicates whether or not a ball was absorbed to create this
                strand, which is released if the strand is destroyed.

Next, the online player key follows, which is prefixed by an underscore; the
player key is 32 characters long and consists of lower-case hexadecimal digits.
If the player has never connected to the internet, the key is empty.

Since this key is used to authenticate players online, it should be kept
private. I propose to take an MD5 hash code of the string and use that instead,
so players can be identified without compromising their scoreboard standing.
For example: "d41d8cd98f00b204e9800998ecf8427e" would be transformed to
"74be16979710d4c4e7c6647856088456". If you do not want players to be globally
identifiable, use an HMAC instead.

The final field is a decimal integer, representing the number of newly collected
goo balls since the player last visited World of Goo Corporation. This number is
displayed on the main screen.
[/pre]

I've setup a project on sourceforge to help organizing knowledge about the file structure a little while ago:
See http://sourceforge.net/projects/wogedit

I'm currently focusing on trying to visualize level data correctly (level.bin is mostly done, I'm now focusing on scene.bin).

Feel free to contribute to the wiki of the project:
http://apps.sourceforge.net/mediawiki/wogedit/index.php?title=Main_Page

Thanks Soultaker, this was very useful.

Jonas sent me his profile which fails to parse according to this algorithm (http://pastebin.com/f54eba97a). Here's my analysis.

[quote author=Soultaker link=topic=643.msg4700#msg4700 date=1225388207]
[pre]Obviously, the player profile strings are the most interesting. These strings
consist of a number of fields, seperated by comma's. The first four fields are:
[snip]

Then follows, for each level:
  - level id (same as the directory name in res/levels/)
  - greatest number of balls collected
  - least number of moves used
  - least time taken (in seconds)

Then, the World of Goo Coporation data follows, which is a string prefixed by
an underscore character (_) which is how it can be distinguished from more
level data.
[/pre]


I have found that:

- for most profiles, after the levels, a level named "0" indicates the end of levels, with no balls/moves/time data. After that is the tower, beginning with _.

[pre]...HangLow,22,6,170,[=12pt]0[/],_b:Drained:-488.46:11.32:-1.08:0.00:...[/pre]

- for some players, this 0 doesn't exist. Instead is a 1 followed by a level name, followed by the tower. I'm not sure what the 1 indicates (perhaps the number of fields following), nor what the level name means. I do know that the level doesn't appear anywhere else in the list so maybe it's the level they're currently attempting.

[pre]...ObservatoryObservationStation,0,12,96,[=12pt]1,HTInnovationCommittee[/],_b:Drained:-551.97:811.85:-0.87:-0.44:...[/pre]

I think for now I will code it so that if it receives an integer when expecting a level name, it skips over that many fields to reach the tower. Normally 0 but in this case 1, and possibly may have other values.

-davidc

Sorry for resurrecting this ancient thread, but I figured this information belonged with the rest of the file format information above:

The mysterious list of level names is the list of levels that have been skipped. Makes sense to keep them separate, since you can't have completion stats for them.

Edit: Also, in the tower data:

[quote author=D0SBoots link=topic=643.msg11125#msg11125 date=1238282680]
The mysterious list of level names is the list of levels that have been skipped. Makes sense to keep them separate, since you can't have completion stats for them.


Nice find. Was I correct in assuming that the number that precedes them is the counter of skipped levels then?

-davidc

[quote author=davidc link=topic=643.msg11128#msg11128 date=1238285026]Nice find. Was I correct in assuming that the number that precedes them is the counter of skipped levels then?

Yes, sorry I didn't make that clear. I was a wimp, so in my file it goes, "...,3,UpperShaft,ThirdWheel,HTInnovationCommittee,_b:Dra..."

I'm fiddling with the numbers in WoGC, because I'm really interested to know the role of the rest of those fields. I'll edit more stuff in to my first post to keep things compact.