Below is the list of all rooms of figure spaces. I hope I got them right, I didn't double check. These writings must have a meaning, I'm expecting something in the spirit of KG's levels. The decoded language may be English or Swedish.
Gile Jyvy
Ba Gepehasy
Hevore 0
Lady Tufe
Govu Letahe
Teta Minu
Somy Luru
Vafo Byfu
Jejobo Redy
Fegely
Mofa Dabusu
Jabahifymo
Hevore 1
Soko Jyja
Vapame
Falu Kohe
Falu Kohe (Kybiga Nofa)
Tibemy
Hovynofo
Ruhele
The words seem to have vowels and consonants in order. So, the transformation probably preserves them.
Nifflas' comment may imply that there actually isn't a strict rule of transformation, but some more general logic.
Here are my blind guesses:
Nofa = Mode (this one's got good chances to be correct)
La Vupene = La Mulana (off the top of my head and unlikely the right one. If there is a strict rule of transformation, this one is wrong)
I don't think they are anagrams, at least not for English words.
Take Hevore O. It's only English anagrams are: Veer Oho, Veer Ooh, Ever Oho, Ever Ooh, Hoe Rove, Hoe Over, Hove Roe, Hove Ore, Eve Oh Or, Eve Ho Or, none of which would make sense.
Then I thought each may be an independent riddle. For example, "Hevore O" might mean remove the "O" from "Hevore", leaving "Hevre", the hebrew word for a group of friends.
However, Nifflas is usually more methodical than that.
I thought maybe there might be a simple algorithm that is swapping letters around, so I set O equal to I and A, the only english one-letter words.
HEVORE O:
8-5-22-15-18-5-15
O=A:
TOHADO A
20-15-8-1-4-15-1
O=I
BYPILY I
2-25-16-9-12-25-9
None of which make sense.
Then I remembered that it isn't "Hevore O," but "Hevore 0," and that I'd been wasting my time. Sort of.
I did at least prove that they are not english anagrams...
I think I was onto something with the algorithms, though.
Forgot to post this.
I continued my failed work with simple algorithms a bit, this time with Ruhele.
First is setting R equal to every letter and the sequence that follows.
Second is displacement by near-by keys.
A third idea I had was to plug these into that type game nifflas made a while ago, but I never got around to it.
R U H E L E
----------------
A D Q N U N
B E R O V O
C F S P W P
D G T Q X Q
E H U R Y R
F I V S Z S
G J W T A T
H K X U B U
I L Y V C V
J M Z W D W
K N A X E X
L O B Y F Y
M P C Z G Z
N Q D A H A
O R E B I B
P S F C J C
Q T G D K D
R U H E L E
S V I F M F
T W J G N G
U X K H O H
V Y L I P I
W Z M J Q J
X A N K R K
Y B O L S L
Z C P M T M
----------------
R U H E L E
----------------
E Y G W K W
T I J R A R
As you can see, they both failed.
No simple letter-swap algorithm exists. (English)
Displacing with near-by letters isn't complete, though things get confusing when you go vertical instead of horizontal. It would seem this is also incorrect.
I'll try more complex algorithms soon :P
Actually, thibthib figured it out :)