Re: Phoneme winnowing continues

Hi!

"Mark J. Reed" <markjreed@...> writes:
> On Fri, Jun 06, 2003 at 09:57:00PM +0200, Henrik Theiling wrote:
> > "Mark J. Reed" <markjreed@...> writes:
> > > Henrik wrote rubbish again:
>
> I wrote no such thing.  You know, false accusations of slander are
> themselves slanderous. ;-)
HAHAHA!  Sorry, the initial > should probably have been deleted...
:-) (-:

> > Notice that 21 = k and 22 also = k?  22 should be = x.
> > (This also explains 21+22=kx)
>
> You mean the way 15=N and 21=k explains 15+16=Nk?  For some reason
> I didn't really feel safe making that generalization . ;-)
HAHA!  Hmm, yeah, that's a bit of an inconsistency isn't it?  But I
did not want a bent, so that is not 15+21, but 15+16.

Well, I actually meant analogous to 17 = t, 18 = s, 17+18 = ts.

Ok, in the first syllable, because there is no /ts/, the theoretical
combination for /t/+/s/ is pronounced /n/...

It *is* quite logical, really. %-)

> Also for recognition, I would think.  Without the connectivity it's
> hard for the eye to create the logical "glyph space" around the strokes
> to identify what part of the space the strokes are in, or possibly -
> depending on the arrangement of adjacent glyphs - even which glyph
> they're part of.
I have actually checked that.  It was one of the design goals to make
the eye catch and recognise a word as a whole.  The writing has a very
high recognition rate.  When I show someone a text and tell them to
search *THAT* character, their eye (not only mine) finds equal
characters immediately.  And the brain *does* create boxes, 'glyph
space', around each character easily, too.  So I think I'm satisfied
with that issue so far.

To optimise into this direction, there are very few strokes to make
the characters less messy.

**Henrik