On Sun, 16 May 2021 11:26:06 +0000, Bo Jacoby wrote (to 
[email protected] <mailto:[email protected]>):
...  ordinal fractions are more general than arrays.
-----
Comments on "ORDINAL FRACTIONS / the algebra of data" by Bo Jacoby
at: https://www.academia.edu/10031088/ORDINAL_FRACTIONS_the_algebra_of_data

Constricting the symbol set to ten characters requires the introduction of 
extraneous structure in the representation of data having some aspect with a 
greater than ninefold variation.  The same principles could be applied to 16 
hexadecimal digits, or any larger system.  In J, a list of digits, lowercase 
letters, and uppercase letters could readily provide 62 symbols.  Even more 
generally, arbitrary bitstrings of length n provide 2^n distinct values, which 
may be stored individually in arrays, or concatenated for storage and split 
apart for use.

It is not immediately clear how to deal with data which receive two different 
representations in the hierarchy, as when "grandfather" of primary object A is 
also "grandfather" of object B.  It would seem to require acceptance of an 
equation "A01 = B01", which would normally be erroneous.  (This occurs when 
their relationships do not form a tree, but a lattice--or even loops!)
--Art Anger

----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm

Reply via email to