Print

Print


David J. Peterson wrote:
> Roger:
> <<
> One difficulty with fonts (and the reason I haven't made many) is the 
> problem with ligatures, contextual variations, marks that need glyph 
> positioning tables, etc. I do have an OlaeUni font which follows the 
> CSUR submission (ftp://ftp.io.com/pub/usr/hmiller/fonts/OLAEUNI_.TTF), 
> although there are a couple of issues with it that would need OpenType 
> features to implement, and OpenType doesn't work with Private Use 
> scripts. Neesklaaz requires complex script processing, even beyond what 
> OpenType tables provide, so my Neesklaaz font only works with SIL Graphite.
>  >>
> 
> Not 100% sure what all that means, but how's this:
> 
> <http://dedalvs.free.fr/writing/ConlangUnicode.ttf>
> 
> I think the characters should show up (they're all in the right
> place).  There are two issues with OlaeUni font:
> 
> (1) There are a lot more characters than I included because you've
> put them in non-private use areas, where other characters from
> other scripts currently sit.  Was that intentional?

The precomposed characters probably should exist in the font without a 
specific encoding. (It looks like they're in the Cyrillic range, 
probably because the default font mapping includes Cyrillic pages and I 
didn't bother to zero out the Unicode values of these extra characters).

> (2) My understand is that these ligature characters don't work
> the way you want them to.  For example, on a Mac, I can "add"
> an acute accent to an "e", but all I'm really doing is typing in a
> code that tells the computer to display the "e-acute" cell.  The
> glyphs you have that display over glyphs already typed should
> work, but I don't *think* Unicode is supposed to be used to,
> say, type an "n", and then type a "g", and then have it automatically
> produce an "ng" ligature.  In order for that ligature to be available,
> it must have a codepoint, and currently it doesn't.

Olaetian has different writing styles, some with an "ng" ligature and 
others without (like the "fi" ligature in English). What may have been a 
mistake when I wrote the CSUR description is not allocating a separate 
code point for the "kc" ligature, which is used in some words but not 
others. I get around that by writing "" for "kc" when the ligature 
should be used (with the "silent e" diacritic over the silent "c").

Other writing systems have more difficult rendering issues. Kazvarad has 
letters with long horizontal strokes that can run into adjacent letters; 
there are short forms of the letters that are used in that case. 
Neesklaaz letters have different forms that attach to preceding or 
following letters, and even ligatures of diacritics, which (to make 
things even trickier) change their positioning depending on what letter 
they're attached to.

> I've installed the test version of the font on my machine, and
> it doesn't seem to be working.  That is, I tested out one glyph
> by putting it on a webpage.  I have unicode encoding up; the
> glyph is at codepoint F072; I typed & # x F 0 7 2 ; onto the
> page (no spaces); I have the font installed; I opened up the
> webpage, and I get a square box.  Have I missed a step?

You might need a meta tag in the head of your page (I've left the angle 
brackets off so that HTML mail readers don't attempt to interpret this 
as HTML):

META HTTP-EQUIV="content-type" CONTENT="text/html; charset=UTF-8"