Re: A single font can display ANY alphabet, pictograph, or rune

Hi all,

On Tue, 11 Oct 2005, tomhchappell wrote:
> Hi, all, especially Gary Shannon.
>
> I have been thinking of something like this for quite some time now;
> you've made more progress than I, especially with your alphabetic-
> codings of the various strokes.
>
> -----
>
> [WHAT I DID:  (skip to "curves" if you want totally new ideas)]
>
> I based my ideas on 1) the LED outputs of elevators in Vancouver
> during the World's Fair there and 2) the "imaginary lines" frame
> provided for teaching Chinese script.
... [much snipt]

[YA]
I like your analysis.  However, I think you may need cubic splines
to produce well-formed characters as used in many natscripts eg
Devanagari or Cambodian.


> Most strokes start, and most end, either on one of the lines --
>  "imaginary" or not -- or someplace between two of them (it doesn't
> matter much where, as long as it is clearly between the correct two.)
>
> This could all be simulated by having each glyph be 12 cells high --
> eight cells above the base-line, and four below it.
[YA]
Which is of course many more cells than the two by two
arrangement for which you developed a set of stroke codes ...
My own experiments in phonetic font & script design, many
years ago, led me to conclude that you can't get good
approximations to existing scripts or create a reasonable
facsimile of human handwriting motions with any less than
four or five cells on a side, both horizontally and vertically.

Also, my maternal grandfather taught primary school for
many years in Tasmania, where I grew up.  My mother had
a filing cabinet full of the instruction cards he used in
teaching the arts of drawing, writing, and lettering.  Those
for handwriting all followed the style known as Victorian
Copperplate; they dated from the twenties and thirties.
By the time I started school in the fifties, we still learnt
a Copperplate Italic hand, but with much simplified capital
(majuscule) letters.  The foundation of a good Copperplate
hand was said to be the "pot-hook".  We were drilled (it
may only SEEM) endlessly in producing acceptable pothooks
and slanted verticals for many weeks before we were ever
allowed to write a circle or - wonder of wonders! - a whole
letter ... :-)  The point of this anecdote is that we learnt a
repertoire of a mere handful of strokes, from which we
then learnt to construct a full pair of alphabets, both
majuscule and minuscule.

When I later tried to reproduce these basic copperplate
strokes on a small grid, I found that the smallest grid in
which I got passable results was four cells wide and five
cells high.  Still, the results were nothing to brag about.
In fact, most of these "grid-fit" characters were rather
lumpish and disproportionate.  So I tend to agree with
Tom's larger numbers as a reasonable base resolution.

You've probably all noticed that computer fonts in
anything less than about 20 pixels square are fairly crude.
(Font anti-aliasing does improve the results a bit.)  Since
a cubic curve requires three points to fix it, 20 pixels can
represent at most 9 joined cubics of large curvatures; at
most 5 joined cubics of moderate curvature; or at most
3 joined cubics of subtle curvature.  These figures suggest
that using Tom's two-by-two cell approach, we may need to
make both sides of each cell from six to ten pixels, in order
to satisfactorily approximate the curves we write by hand.

But I do think that Tom's approach - providing a fixed
repertoire of strokes, of given complexity (whether that
be linear or cubic), to combine freely in a fairly small
number of grid positions - may well be very fruitful in
creating a flexible script representation.

How best to generate practical scripts is another matter,
and I think that may benefit from studying the dynamics
of handwriting.  Because dynamics involves both speed
and acceleration, its description needs more than a fixed
grid of cells of fixed size, which is better suited to the
more static process of creating a print typeface from
elements of fixed size.

Regards,
Yahya


--
No virus found in this outgoing message.
Checked by AVG Anti-Virus.
Version: 7.0.344 / Virus Database: 267.11.13/126 - Release Date: 9/10/05