14-09-2018, 10:22 PM
There are two implications of this approach to the VMS.
First, character entropies of the "truncated" version will be higher that those of expanded version. This is not the direction which we would like to see. However, values such as h0-h1 or h1-h2 may be also higher for the "truncated" version. (This is the desired direction of change). I did a brief check with the declaration of human rights in Latin, in which I changed all "-tis" and "-ris" endings (42 and 15 occurrences, respectively) with single characters T and R, respectively. And what I observed was change of h0-h1 from 0,45 to 0,54, and of h1-h2 from 0,78 to 0,79.
Second, treatment of certain EVA glyphs as of combination of other glyphs with abbreviation symbols effectively reduces the "core" alphabet. Most probably this will lead to situation when the size of the core alphabet is suspiciously small. What will be left? c, i, e, h, a, l, o and the gallows.
Abjad with positional abbreviations?
First, character entropies of the "truncated" version will be higher that those of expanded version. This is not the direction which we would like to see. However, values such as h0-h1 or h1-h2 may be also higher for the "truncated" version. (This is the desired direction of change). I did a brief check with the declaration of human rights in Latin, in which I changed all "-tis" and "-ris" endings (42 and 15 occurrences, respectively) with single characters T and R, respectively. And what I observed was change of h0-h1 from 0,45 to 0,54, and of h1-h2 from 0,78 to 0,79.
Second, treatment of certain EVA glyphs as of combination of other glyphs with abbreviation symbols effectively reduces the "core" alphabet. Most probably this will lead to situation when the size of the core alphabet is suspiciously small. What will be left? c, i, e, h, a, l, o and the gallows.
Abjad with positional abbreviations?