VMs: RE: Qoteedy nine

["GC" <glenclaston@xxxxxxxxx>]

Nick wrote:

> Unless you know what is being written in the alphabet,
> statistics may not
> be completely helpful. For example, if (as I suspect)
> it is largely
> comprised of a *number system* (written in a kind of
> shorthand-stylised
> Roman numerals) with additional letters to decline/conjugate as
> appropriate, then it might very well have the kind of
> low entropy stats
> profile we've come to recognise.

I guess I'm having a problem considering Roman numerals because of
the size of the "alphabet".  Although my recording is up to some
78 variations, a count readily yields the fact that 23 common
glyphs are top runners, something I consider to be too high a
number for Roman numerals.  (I consider <in> and <iin> separate
and distinct glyphs, along with their variations.)  These 23 occur
thousands of times, and after them the counts of variants drops
off sharply.  I also consider that although there are a great
number of variants (some of which may be attributed to
penmanship), none of these seem to raise the glyph counts on any
given page well above the number 23, 24 at the outside.

So I guess my problem with a Roman numeral tachygraphy is this -
if speed of writing is the goal, why use a character set as large
as an alphabet, and why limit the number of variations of these
characters on any given page?  A tachygraphic system has much of
its speed in variations of base symbols, (28 symbols is the
smallest recorded base set in stenography)while declension and
other language artifacts are usually omitted and later
reinterpreted through context.

I'd also point out that the length of the "words" would preclude
each character being a word and the breaks being a sentence.  It's
more likely to me that they are indeed words of varied length and
varied spelling.  I'd point to the labels as an example, where we
tend to think that some are names of plants and others names of
stars.  These are unlikely to be entire sentences, more likely
single words.

Based on the glyph counts, the number of unique glyphs per page,
and the 'word length' observed in the VMS, I can't help but think
we're looking at a medium of information encoding that is
character-to-character and word-to-word, written top to bottom and
left to right.

Does your view of Roman numerals answer these observations, and if
so, can you give me a glimpse of how it would theoretically
operate?

GC