Q for Zomp: Verdurian font

[Serali]

I've turned the diacritic glyphs into zero-width characters: here's the new font. (I've renamed it "Verdurian CSUR" to make it easier to pick the right font for the right encoding.)

The text   now looks like this under Mozilla:

(The initial capital pe letter was munged up because, for some reason, a Greek polytonic font (and probably other fonts) on my system decided to dump some stuff into the Unicode private area... ugh.)

The same text under GIMP (well, after I specified the font to use):

And here's what Philip Newton's page now looks like on my browser:



Leaving aside the totally bogus non-Verdurian characters, there's still a problem: the diacritics tend to be at the wrong places (especially those above small e). Oh well.

WurdBendur: Unfortunately, doing ligatures and stuff within the Unicode private ranges seems to be outside the realm of most text renderers; it seems we'll need something like Graphite.

OMG! Preeeeeeeeetty Scripty!

[bi]

zompist: it seems that there are no characters in the CSUR proposal for the "ĭ" glyph ( ) and the dash "-". What's the correct way to handle them?

[zompist]

zompist: it seems that there are no characters in the CSUR proposal for the "ĭ" glyph ( ) and the dash "-". What's the correct way to handle them?

IIRC there are a couple slots open in the CSUR code block-- you can fit ĭ in there somewhere, I don't think it matters where. The dash too if there's room.

[WurdBendur]

WurdBendur: Unfortunately, doing ligatures and stuff within the Unicode private ranges seems to be outside the realm of most text renderers; it seems we'll need something like Graphite.

A good number of modern text editors are starting to handle OpenType for ligatures, kerning and diacritic placement, so it's doable. It'll just be a little while before they work out the stupid (for example, automatic placement of diacritics in the PUA doesn't work in several programs, like Word, even if the OpenType code specifies these.)

It might be possible to simply make a set of precombined characters and treat them like basic ligatures, which are pretty well supported.

[bi]

zompist:

Thanks... there's not much space left (only 7 unassigned codes), so instead of using up two more codes for "Ĭ" and "ĭ" separately, I decided to assign a single code for the "inverted mole" diacritic, at U+E229 -- so "Ĭ" will be U+E204 U+E229, and "ĭ" will be U+E234 U+E229. The output doesn't look very pretty though.

Then the dash "-" goes to U+E22A. I think this assignment sort of preserves the symmetry between U+E200--U+E22F and U+E230--U+E25F (38 codes for letters, then 4 for diacritics, then 6 for signs).

Again, I've updated the font file.

* * *

WurdBendur: Hmm... for some reason, even after I created some "Required" ligature lookups for e.g. U+E223 U+E256, I still can't get any program to resolve them into the precomposed glyph. (GIMP uses Pango if I'm not wrong, and it doesn't work there either. Argh.)

[pne]

FWIW, I've updated the Luke page to remove the Latin 'B' 'L' '!' characters (I've replaced them with the proper Verdurian characters now).

I've turned i-breve into plain (Verdurian) i -- I believe this is common practice in Verdurian texts anyway.

I've kept the hyphen as ASCII for now

[bi]

pne: Thanks! The "-"s still come up bogus on my browser, but I guess that can't be helped.

[bi]

Finally... I got around to using Graphite to add rules to resolve letter + diacritic pairs into precomposed glyphs (unfortunately, this requires one to use a Graphite-capable program like SILA to see the precomposed characters; the diacritics will still be placed wrong on stock browsers).

Here's the updated font, and this is what it looks like under SILA.

These are the Graphite rules I added:

Code: Select all

#define ps postscript
#define comp(c1, c2, r) ps(c1) ps(c2) > ps(r):(1 2) _
#define lenge(c1, r) comp(c1, "lenge", r)
#define vuae(c1, r) comp(c1, "vuae", r)
#define mole(c1, r) comp(c1, "mole", r)
#define lengc(c1, r) comp(c1, "lengecuzea", r)
#define elom(c1, r) comp(c1, "elom", r)

Bidi = false;

table(substitution)
        // u
        lenge("Uverd", "Ulenge");       lenge("uverd", "ulenge");
        vuae ("Uverd", "Uvuae");        vuae ("uverd", "uvuae");
        mole ("Uverd", "Umole");        mole ("uverd", "umole");
        lengc("Uverd", "Ulengecuzea");  lengc("uverd", "ulengecuzea");
        elom ("Uverd", "Uelom");        elom ("uverd", "uelom");
        // a
        lenge("Averd", "Alenge");       lenge("averd", "alenge");
        vuae ("Averd", "Avuae");        vuae ("averd", "avuae");
        mole ("Averd", "Amole");        mole ("averd", "amole");
        lengc("Averd", "Alengecuzea");  lengc("averd", "alengecuzea");
        // o
        lenge("Overd", "Olenge");       lenge("overd", "olenge");
        vuae ("Overd", "Ovuae");        vuae ("overd", "ovuae");
        mole ("Overd", "Omole");        mole ("overd", "omole");
        lengc("Overd", "Olengecuzea");  lengc("overd", "olengecuzea");
        // e
        lenge("Everd", "Elenge");       lenge("everd", "elenge");
        vuae ("Everd", "Evuae");        vuae ("everd", "evuae");
        mole ("Everd", "Emole");        mole ("everd", "emole");
        lengc("Everd", "Elengecuzea");  lengc("everd", "elengecuzea");
        // i
        lenge("Iverd", "Ilenge");       lenge("iverd", "ilenge");
        vuae ("Iverd", "Ivuae");        vuae ("iverd", "ivuae");
        mole ("Iverd", "Imole");        mole ("iverd", "imole");
        lengc("Iverd", "Ilengecuzea");  lengc("iverd", "ilengecuzea");
        elom ("Iverd", "Ielom");        elom ("iverd", "ielom");
        // y (i breve)
        vuae ("Yverd", "Yvuae");        vuae ("yverd", "yvuae");
        // kodh
        vuae ("Kodh", "Kodhvuae");      vuae ("kodh", "kodhvuae");
        // chen ismae
        vuae ("Chenismae", "Chenismaevuae");
        vuae ("chenismae", "chenismaevuae");
endtable;

[Serali]

Awesome work!