This should do it:


group @Group1 [zero];
group @Group2 [zero.2];
group @Group3 [zero.3];
group @Group4 [zero.4];
group @Group5 [zero.5];

feature ContextualAlternates1 calt {
# Contextual Alternates
lookup Lookup1;
lookup Lookup2;
}

lookup Lookup1 {
# Referenced by feature "ContextualAlternates1"
context (@Group1) @Group1;
sub 0 Lookup3;
context (@Group2) @Group1;
sub 0 Lookup4;
context (@Group3) @Group1;
sub 0 Lookup5;
context (@Group4) @Group1;
sub 0 Lookup6;
}

lookup Lookup2 {
# Referenced by feature "ContextualAlternates1"
context (@Group5 @Group4) @Group1;
sub 0 Lookup7;
context (@Group2 @Group5) @Group2;
sub 0 Lookup7;
context (@Group4 @Group2) @Group3;
sub 0 Lookup7;
context (@Group1 @Group4) @Group4;
sub 0 Lookup7;
context (@Group5 @Group1) @Group5;
sub 0 Lookup7;
context (@Group3 @Group5) @Group1;
sub 0 Lookup7;
context (@Group2 @Group3) @Group2;
sub 0 Lookup8;
context (@Group1 @Group2) @Group3;
sub 0 Lookup8;
context (@Group4 @Group1) @Group4;
sub 0 Lookup8;
context (@Group3 @Group4) @Group4;
sub 0 Lookup9;
}

lookup Lookup3 {
# Referenced by lookup "Lookup1"
sub zero -> zero.2;
}

lookup Lookup4 {
# Referenced by lookup "Lookup1"
sub zero -> zero.3;
}

lookup Lookup5 {
# Referenced by lookup "Lookup1"
sub zero -> zero.4;
}

lookup Lookup6 {
# Referenced by lookup "Lookup1"
sub zero -> zero.5;
}

lookup Lookup7 {
# Referenced by lookup "Lookup2"
sub zero -> zero.2;
sub zero.2 -> zero.4;
sub zero.3 -> zero;
sub zero.4 -> zero.5;
sub zero.5 -> zero.3;
}

lookup Lookup8 {
# Referenced by lookup "Lookup2"
sub zero.2 -> zero;
sub zero.3 -> zero.4;
sub zero.4 -> zero.3;
}

lookup Lookup9 {
# Referenced by lookup "Lookup2"
sub zero.4 -> zero.5;
}

I've provided some more information along with the FontCreator project file here:
http://forum.high-logic.com/viewtopic.php?f=4&t=4797

Thank you for saying "pseudo".

hhp

Yes, I think that is very likely.

To be sure, just open the font I made with FontCreator and see how it looks in your font editor.

You can download the font from here:
AlmostRandom

Let us know your results.

Unfortunately I can not open the .fcp file with Glyphs app. The coding seams to be a bit different in FontCreator.
Can you explain a bit more in detail what happens in lookup1 and lookup2? I don't understand the function "context" and "sub 0". It seems like the problem is something with this part of the code:

lookup Lookup1 {
  # Referenced by feature "ContextualAlternates1"
  context (@Group1) @Group1 ;
  sub 0 Lookup3 ;
  context (@Group2) @Group1 ;
  sub 0 Lookup4 ;
  context (@Group3) @Group1 ;
  sub 0 Lookup5 ;
  context (@Group4) @Group1 ;
  sub 0 Lookup6 ;
}Lookup1;

lookup Lookup2 {
  # Referenced by feature "ContextualAlternates1"
  context (@Group5 @Group4) @Group1;
  sub 0 Lookup7;
  context (@Group2 @Group5) @Group2;
  sub 0 Lookup7;
  context (@Group4 @Group2) @Group3;
  sub 0 Lookup7;
  context (@Group1 @Group4) @Group4;
  sub 0 Lookup7;
  context (@Group5 @Group1) @Group5;
  sub 0 Lookup7;
  context (@Group3 @Group5) @Group1;
  sub 0 Lookup7;
  context (@Group2 @Group3) @Group2;
  sub 0 Lookup8;
  context (@Group1 @Group2) @Group3;
  sub 0 Lookup8;
  context (@Group4 @Group1) @Group4;
  sub 0 Lookup8;
  context (@Group3 @Group4) @Group4;
  sub 0 Lookup9;
}Lookup2;

Maybe I'm missing something, but would it not be simpler to say

sub zero from zero-class

Maybe it is:

 "sub zero by lookup 3" 

I don't really understand the logic behind the code, which makes it hard to find the mistake.
I've never seen a code with "context", and the "sub 0" is also weird:

context (@Group1) @Group1 ;
  sub 0 Lookup3 ;

any hints?

[FontForge code, font from erwindenissen ]

lookup caltContextualAlternatesinLatin {
lookupflag 0;
sub [\zero ] [\zero ]' by [\zero.2 ];
subtable;
sub [\zero.2 ] [\zero ]' by [\zero.3 ];
subtable;
sub [\zero.3 ] [\zero ]' by [\zero.4 ];
subtable;
sub [\zero.4 ] [\zero ]' by [\zero.5 ];
} caltContextualAlternatesinLatin;

lookup SingleSubstitutionlookup6 {
lookupflag 0;
sub \zero by \zero.2 ;
sub \zero.2 by \zero.4 ;
sub \zero.3 by \zero ;
sub \zero.4 by \zero.5 ;
sub \zero.5 by \zero.3 ;
} SingleSubstitutionlookup6;

lookup SingleSubstitutionlookup7 {
lookupflag 0;
sub \zero.2 by \zero ;
sub \zero.3 by \zero.4 ;
sub \zero.4 by \zero.3 ;
} SingleSubstitutionlookup7;

lookup sub_chain_caltlatn_0 {
lookupflag 0;
sub [\zero.4 ] [\zero.5 ] [\zero ]'lookup SingleSubstitutionlookup6 ;
subtable;
sub [\zero.5 ] [\zero.2 ] [\zero.2 ]'lookup SingleSubstitutionlookup6 ;
subtable;
sub [\zero.2 ] [\zero.4 ] [\zero.3 ]'lookup SingleSubstitutionlookup6 ;
subtable;
sub [\zero.4 ] [\zero ] [\zero.4 ]'lookup SingleSubstitutionlookup6 ;
subtable;
sub [\zero ] [\zero.5 ] [\zero.5 ]'lookup SingleSubstitutionlookup6 ;
subtable;
sub [\zero.5 ] [\zero.3 ] [\zero ]'lookup SingleSubstitutionlookup6 ;
subtable;
sub [\zero.3 ] [\zero.2 ] [\zero.2 ]'lookup SingleSubstitutionlookup7 ;
subtable;
sub [\zero.2 ] [\zero ] [\zero.3 ]'lookup SingleSubstitutionlookup7 ;
subtable;
sub [\zero ] [\zero.4 ] [\zero.4 ]'lookup SingleSubstitutionlookup7 ;
subtable;
sub [\zero.4 ] [\zero.3 ] [\zero.4 ]' by [\zero.5 ];
} sub_chain_caltlatn_0;

feature calt {
script DFLT;
language dflt ;
lookup caltContextualAlternatesinLatin;
lookup sub_chain_caltlatn_0;
script latn;
language dflt ;
lookup caltContextualAlternatesinLatin;
lookup sub_chain_caltlatn_0;
} calt;

The following code works in FontForge and gives the same behaviour as the lookups in AlmostRandom.otf.

lookup contextofone {
  lookupflag 0;
    sub [\zero  ] [\zero]'  by [\zero.2] ;
    sub [\zero.2] [\zero]'  by [\zero.3] ;
    sub [\zero.3] [\zero]'  by [\zero.4] ;
    sub [\zero.4] [\zero]'  by [\zero.5] ;
} contextofone ;

lookup contextoftwo {
  lookupflag 0;
    sub [\zero.4] [\zero.5] [\zero]'    by [\zero.2] ;
    sub [\zero.5] [\zero.2] [\zero.2]'  by [\zero.4] ;
    sub [\zero.2] [\zero.4] [\zero.3]'  by [\zero  ] ;
    sub [\zero.4] [\zero  ] [\zero.4]'  by [\zero.5] ;
    sub [\zero  ] [\zero.5] [\zero.5]'  by [\zero.3] ;
    sub [\zero.5] [\zero.3] [\zero  ]'  by [\zero.2] ;
    sub [\zero.3] [\zero.2] [\zero.2]'  by [\zero  ] ;
    sub [\zero.2] [\zero  ] [\zero.3]'  by [\zero.4] ;
    sub [\zero  ] [\zero.4] [\zero.4]'  by [\zero.3] ;
} contextoftwo ;

feature calt {
  script DFLT;
     language dflt ;
      lookup contextofone;
      lookup contextoftwo;

  script latn;
     language dflt ;
      lookup contextofone;
      lookup contextoftwo;
} calt;

In the font AlmostRandom.otf, the zeros are numbered from 1 to 5; the default zero contains a 1; without any substitution lookups you would see zeros numbered as follows:

   111111111111111

With the first lookup, that uses a single character as left context, you get

   123451234512345

With the second lookup, that uses two characters as left context, you replace the first occurrence of 451 by 452 and advance the cursor to get

   1234522|34512345

and then 522 becomes 524 and so on.

You end up with

   123452415321435

and that sequence is then repeated indefinitely.

For more details, cf http://typophile.com/node/19625

Michel

Thanks, @MichelBoyer. I finally understand the concept.

I have one more problem. I managed to randomize the zero's with 10 alternates and it works, as long as I type the zeros in one line. When I start a new row, it starts the repetition from 1 again.

So in one row the pattern looks like this: 0,1,2,3,4,5,6,7,8,9,10 - 1,10,00,9,2,8,3,6,4,7,5 - 9,6,0,2,5,7,10,4,1,8,3

But when I have a linebreak, it starts over again and makes this:
1,2,3,4,5,6,7,8,9,10 - 1,10,00,9,2,
1,2,3,4,5,6,7,8,9,10 - 1,10,00,9,2,

I would like it to be like this:
0,1,2,3,4,5,6,7,8,9,10 - 1,10,00,9,2,
8,3,6,4,7,5 - 9,6,0,2,5,7,10,4,1,8,3 etc...

This is what my code looks like with 10 additional alternates:
lookup calt_1 {
lookupflag 0;
sub zero.00 zero.00' by zero.01 ;
sub zero.01 zero.00' by zero.02 ;
sub zero.02 zero.00' by zero.03 ;
sub zero.03 zero.00' by zero.04 ;
sub zero.04 zero.00' by zero.05 ;
sub zero.05 zero.00' by zero.06 ;
sub zero.06 zero.00' by zero.07 ;
sub zero.07 zero.00' by zero.08 ;
sub zero.08 zero.00' by zero.09 ;
sub zero.09 zero.00' by zero.10 ;
} calt_1;

lookup calt_2 {
lookupflag 0;
sub zero.09 zero.10 zero.00' by zero.01 ;
sub zero.10 zero.01 zero.01' by zero.10 ;
sub zero.01 zero.10 zero.02' by zero.00 ;
sub zero.10 zero.00 zero.03' by zero.09 ;
sub zero.00 zero.09 zero.04' by zero.02 ;
sub zero.09 zero.02 zero.05' by zero.08 ;
sub zero.02 zero.08 zero.06' by zero.03 ;
sub zero.08 zero.03 zero.07' by zero.06 ;
sub zero.03 zero.06 zero.08' by zero.04 ;
sub zero.06 zero.04 zero.09' by zero.07 ;
sub zero.04 zero.07 zero.10' by zero.05 ;

sub zero.07 zero.05 zero.00' by zero.09 ;
sub zero.05 zero.09 zero.01' by zero.06 ;
sub zero.09 zero.06 zero.02' by zero.00 ;
sub zero.06 zero.00 zero.03' by zero.02 ;
sub zero.00 zero.02 zero.04' by zero.05 ;
sub zero.02 zero.05 zero.05' by zero.07 ;
sub zero.05 zero.07 zero.06' by zero.10 ;
sub zero.07 zero.10 zero.07' by zero.04 ;
sub zero.10 zero.04 zero.08' by zero.01 ;
sub zero.04 zero.01 zero.09' by zero.08 ;
sub zero.01 zero.03 zero.10' by zero.03 ;

} calt_2;

any ideas?

If your aim is to fill a page with only variants of zero, I see no OTL solution because no contextuality is carried from one line to the next. When pseudo-randomness is used to mimic handwriting for instance, then it is possible to use some character further away in the current line as a "seed" for a "pseudo-random process" so as to avoid repetition. There is an example in Nick Shinn's thread http://typophile.com/node/46521.

Michel