Jorge's error-tolerant encoding "ECC"

The ECC column is an error-tolerant encoding that I used to prepare the label location map. (The name "ECC" has no particular meaning.)

To produce this encoding, the Voynich characters are broken down into separate pen strokes. Then certain stroke sequences are collapsed back into single characters (not necessarily honoring the original character boundaries), omitting several details that are either too prone to error, that seem to be meaningless calligraphic variations, or that seem to have little semantic value.

In particular, the ECC encoding ignores

• ligatures, such as the difference between Friedman's [T] and [CC], or [S] and [2C];
• the horizontal line in "intruding" gallows, such as [HZ], [PZ], [DZ], [FZ];
• the left "eye" on the gallows, that distinguishes [H] from [D], and [P] from [F];
• the connectivity between the `c' and `i' strokes that distinguishes [A] from [CI];
• any additional `i's after three consecutive `i' strokes, in particular the difference between [AM], [AN], [AIM], [CIN], [CIIN], etc.
• extension of strokes below the baseline, in particular the difference between [G] and [A].
• the difference between [O] and [A].

The ECC encoding uses only the letters 8coqHPemrkij. Here is the approximate correspondence between Frogguy and ECC:

    Frogguy:  o    a    9    4    8    ig  
    --------  ---- ---- ---- ---- ---- ----
    ECC:      o    o    o    q    8    8   
            
    Frogguy:  x    ix   iix  iiix
    --------  ---- ---- ---- ----
    ECC:      e    e    e    e   
            
    Frogguy:  e    e'   c    t    s    v    iv   iiv  iiiv
    --------  ---- ---- ---- ---- ---- ---- ---- ---- ---- 
    ECC:      c    c    c    c    c    m    m    m    m    
            
    Frogguy:  2    i2   ii2  iii2 r    ir   iir  iiir 
    --------  ---- ---- ---- ---- ---- ---- ---- ---- 
    ECC:      r    r    r    r    r    r    r    r    
            
    Frogguy:  k    ik   iik  iiik eg  
    --------  ---- ---- ---- ---- ----
    ECC:      k    k    k    k    cj  
            
    Frogguy:  qp   lp   eQPt eLPt
    --------  ---- ---- ---- ----
    ECC:      H    H    cHc  cHc 
            
    Frogguy:  dj   fj   eDJt eFJt
    --------  ---- ---- ---- ----
    ECC:      P    P    cPc  cPc 

Here is the approximate inverse mapping from ECC to FSG:

    ECC   FSG  
    ----  ---  
    8     8, 7    
    H     H, D       
    P     P, F       
    o     A, G, CI, O   
    om    AM, AIN, CIIIL, AN, CM, CN, OM, ...
    or    AR, CIR, OR, OIR, O2, A2, ...
    ak    AIK, CIIK, OK, ...
    c     C        
    cHc   HZ, DZ      
    cPc   PZ, FZ      
    co    CA, CG, CCI, TI, SI
    cc    T, S, CC     
    cj    6        
    e     E, IE, IIE, IIIE
    i     I        
    k     K, IK, IIK, IIIK
    m     M, N, IM, IL, L
    q     4        
    r     R, IR, IIR, IIIR, 2, I2, II2, III2

Note that the inverse mapping is (on purpose) quite ambiguous.

Leaving the cj code for FSG 6 was a minor mistake; I shoudl have mapped it to 8, as I did for FSG 7.