blob: 88cb3faebdd58fa271941ac9e34696077729a1bb [file] [log] [blame]
digraph {
graph [outputorder=edgefirst];
node [shape="record", fontname="Noto Sans Mono SemiBold", fontsize=15];
edge [fontname="Verdana", fontsize=12,labeldistance=7.5 ];
fontname="Verdana";
ranksep=0.02; nodesep=0.5;
subgraph {
ranksep="0.02 equally";
preprocessing[style=filled,fillcolor="lightgreen",fontname="Verdana",label="Glyph pre-processing"];
orthographic[style=filled,fillcolor="lightblue",fontname="Verdana",label="Orthographic Unit Shaping"];
reordering[style=filled, fillcolor="lightcoral",fontname="Verdana",label="Reordering group (USE)"];
topographic[style=filled,fillcolor="lightgoldenrod",fontname="Verdana",label="Topographical Features‡"];
typographic[style=filled,fillcolor="lightpink",fontname="Verdana",label="Typographic Presentation"];
positioning[style=filled,fillcolor="lightsalmon",fontname="Verdana",label="Positioning"];
preprocessing->reordering->orthographic->topographic->typographic->positioning;
}
decision1 [shape="diamond", label="Script\ndirection?",fontname="Verdana"];
rvrn->decision1;
ltrfeatures [label="{ltra|ltrm}", fillcolor="lightgreen",style="filled"];
{
rtlfeatures [label="{rtla|rtlm¹}", fillcolor="lightgreen",style="filled"];
}
{
rank=same;
fracfeatures [label="frac²|numr³|dnom⁴", fillcolor="lightpink",style="filled"];
fracnotes [fontname="Verdana",shape=plaintext,label=<<table border="0" cellborder="0" cellspacing="0">
<tr><td align="left" rtlm is scoped to characters with a Unicode mirroring property</td></tr>
<tr><td align="left" frac is scoped to numr + the slash + dnom</td></tr>
<tr><td align="left" numr is scoped to all decimal numbers before a U+2044 FRACTION SLASH.</td></tr>
<tr><td align="left">⁴ dnom is scoped to all decimal numbers after a U+2044 FRACTION SLASH.</td></tr>
</table>
>];
}
rand [fillcolor="lightpink",style="filled"];
decision1 -> ltrfeatures [label="Left-to-right"];
decision1 -> rtlfeatures [label="Right-to-left"];
decision1 -> fracfeatures [label="Other"];
ltrfeatures -> fracfeatures;
rtlfeatures -> fracfeatures;
fracfeatures->rand;
decision2 [shape="diamond", label="Script?",fontname="Verdana"];
{rank=same; HARF [label="{Harf|HARF}"]; notes;}
rand -> trak -> HARF -> decision2;
commonfeatures [shape=none,label=<<table border="0" cellspacing="0">
<tr>
<td border="1" bgcolor="lightsalmon">abvm</td>
<td border="1" bgcolor="lightsalmon">blwm</td>
<td border="1" bgcolor="lightgreen">ccmp</td>
<td border="1" bgcolor="lightgreen">locl</td>
<td border="1" bgcolor="lightsalmon">mark</td>
<td border="1" bgcolor="lightsalmon">mkmk</td>
<td border="1" bgcolor="lightpink">rlig</td>
</tr>
</table>>
];
decision3 [shape="diamond", label="Script\ndirection?",fontname="Verdana"];
BUZZ [label="{Buzz|BUZZ}"];
BUZZ -> commonfeatures -> decision3;
horizontalfeatures [
shape=none,label=<<table border="0" cellspacing="0">
<tr><td border="1" bgcolor="lightpink">calt <font face="Verdana">(not Hangul)</font></td></tr>
<tr><td border="1" bgcolor="lightpink">clig <font face="Verdana">(not Khmer)</font></td></tr>
<tr><td border="1" bgcolor="lightsalmon">curs</td></tr>
<tr><td border="1" bgcolor="lightsalmon">dist</td></tr>
<tr><td border="1" bgcolor="lightsalmon">kern</td></tr>
<tr><td border="1" bgcolor="lightpink">liga <font face="Verdana">(not Khmer)</font></td></tr>
<tr><td border="1" bgcolor="lightpink">rclt</td></tr>
</table>>
];
vert [label="vert",style=filled,fillcolor="lightpink"];
decision3 -> horizontalfeatures [label="Horizontal"];
decision3 -> vert [label="Vertical"];
discretionary [label="User-selected\ndiscretionary\nfeatures",fontname="Verdana"];
horizontalfeatures -> discretionary;
vert -> discretionary;
decision2->stch;
BUZZ;
subgraph shapers {
subgraph cluster_arabic {
bgcolor="lightyellow"
label="Arabic, Syriac";
stch [ style="filled", fillcolor="lightgreen",label="stch"];
ccmplocl [ style="filled", label="ccmp|locl", fillcolor="lightgreen"];
arabicfeatures [label="isol|fina|fin2|fin3|medi|med2|init", style="filled", fillcolor="lightgoldenrod"];
arabicfeatures2 [label="rclt|calt", style="filled",fillcolor="lightpink"];
rlig[style="filled",fillcolor="lightpink"];
mset [fillcolor="lightpink",style="filled"]
stch->ccmplocl->arabicfeatures->rlig->arabicfeatures2->mset;
}
mset->BUZZ:n;
subgraph cluster_hangul {
bgcolor="lightyellow"
label="Hangul";
hangulfeatures [label="ljmo|vjmo|tjmo", style="filled",fillcolor="lightgoldenrod"]
}
hangulfeatures->BUZZ:n;
subgraph cluster_indic {
label="Indic";
bgcolor="lightyellow"
// Preprocessing
loclccmpindic [label="locl†|ccmp†",style=filled,fillcolor="lightgreen"];
node[style=filled,fillcolor="lightgreen"];
nukt [label="nukt†"];
akhn [label="akhn†"];
loclccmpindic->indic_reorder_1->nukt->akhn;
indic_reorder_1[label="Initial reordering", fontname="Verdana",fillcolor="lightgrey",shape=ellipse,style=filled]
// Orthographic
node[style=filled,fillcolor="lightblue"]
rphf [label="rphf⁵"];
rkpf [label="rkpf†"];
pref [label="pref⁶"];
blwf [label="blwf⁷"];
abvf [label="abvf⁸"];
half [label="half⁹"];
pstf [label="pstf⁸"];
vatu [label="vatu†"];
cjct [label="cjct†"];
akhn ->rphf -> rkpf -> pref -> blwf -> abvf -> half -> pstf -> vatu -> cjct;
// Typographic presentation
indic_typographic[style=filled,fillcolor="lightpink",label="init|pres|abvs|blws|psts|haln"]
indic_reorder_2[label="Final reordering",fillcolor="lightgrey",fontname="Verdana", shape=ellipse,style=filled]
cjct->indic_reorder_2->indic_typographic;
notes2 [fontname="Verdana",shape=plaintext,style="",label=<<table border="0" cellborder="0" cellspacing="0">
<tr><td align="right">⁵ rphf is scoped to pre-base ra+halant sequences</td></tr>
<tr><td align="right">⁶ pref is scoped to the two glyphs after the base; outputs are reordered</td></tr>
<tr><td align="right">⁷ blwf is usually scoped to the whole syllable, except in Telugu and Kannada where it is post-base</td></tr>
<tr><td align="right">⁸ abvf and pstf are scoped to post-base</td></tr>
<tr><td align="right">⁹ half is scoped to pre-base</td></tr>
</table>
>];
indic_typographic -> notes2 [style=invis];
}
subgraph cluster_khmer {
label="Khmer";
bgcolor="lightyellow"
khmerbasic [style=filled,fillcolor="lightgreen",label="locl†|ccmp†|pref†|bwlf†|abvf†|pstf†|cfar†"]
khmerother [style=filled,fillcolor="lightpink",label="pres|abvs|blws|psts"]
khmerbasic -> khmerother -> khmerclig;
khmerclig [label="clig",style=filled,fillcolor="lightpink"];
}
subgraph cluster_myanmar {
label="Myanmar";
bgcolor="lightyellow"
loclccmpmyanmar [label="locl†|ccmp†",style=filled,fillcolor="lightgreen"];
rphfmymr [label="rphf†",style=filled,fillcolor="lightblue"]
prefmymr [label="pref†",style=filled,fillcolor="lightblue"]
blwfmymr [label="blwf†",style=filled,fillcolor="lightblue"]
pstfmymr [label="pstf†",style=filled,fillcolor="lightblue"]
myanmarother [label="pres|abvs|blws|psts",style=filled,fillcolor="lightpink"];
reorder_myanmar[label="Reordering", shape=ellipse,style=filled,fontname="Verdana"]
loclccmpmyanmar -> reorder_myanmar-> rphfmymr -> prefmymr -> blwfmymr -> pstfmymr -> myanmarother;
}
subgraph cluster_use {
label="Universal Shaping Engine"
bgcolor="lightyellow"
use_preprocessing [style=filled, label="locl†|ccmp†|nukt†|akhn†", fillcolor="lightgreen"];
// Reoredering
rphfuse [label="rphf¹⁰", style=filled, fillcolor="lightcoral"];
prefuse [label="pref¹¹", style=filled, fillcolor="lightcoral"];
// Orthographic
orthographicuse [label="rkrf†|abvf†|blwf†|half†|pstf†|vatu†|cjct†", style="filled", fillcolor="lightblue"];
topographicaluse [label="isol|init|medi|fina", style="filled", fillcolor="lightgoldenrod"];
typographicaluse [label="abvs|blws|haln|pres|psts", style="filled", fillcolor="lightpink"];
reorder_use[label="Reordering", shape=ellipse,style=filled,fontname="Verdana"]
use_preprocessing -> rphfuse -> prefuse->orthographicuse ->reorder_use -> topographicaluse -> typographicaluse;
notes3 [fontname="Verdana",shape=plaintext,label=<<table border="0" cellborder="0" cellspacing="0">
<tr><td align="left">¹⁰ Outputs are reordered as category R</td></tr>
<tr><td align="left">¹¹ Outputs are reordered to before base</td></tr>
</table>
>];
typographicaluse -> notes3 [style=invis];
}
}
indic_typographic->BUZZ:n;
typographicaluse->BUZZ:n;
khmerclig -> BUZZ:n;
myanmarother -> BUZZ:n;
decision2->hangulfeatures;
decision2->loclccmpindic;
decision2->khmerbasic;
decision2->loclccmpmyanmar;
decision2->use_preprocessing;
decision2->BUZZ [label=" Hebrew, Thai,\n Lao, other"];
notes [fontname="Verdana",shape=box,label=<<table border="0" cellborder="0" cellspacing="0">
<tr><td align="left">
<b>Indic</b> scripts are: Bengali, Devanagari,
Gujarati, Gurmukhi, Kannada,
Malayalam, Oriya, Tamil,
Telugu
</td></tr>
<tr><td align="left">
<b>USE</b> scripts are:
Adlam, Ahom, Balinese, Batak, Bhaiksuki, Brahmi, Buginese,
Buhid, Chakma, Cham, Chorasmian, Dives Akuru, Dogra, Duployan,
</td></tr>
<tr><td align="left">
Egyptian hieroglyphs, Elymaic, Grantha, Gunjala Ggondi, Hanifi Rohingya,
Hanunoo, Javanese, Kaithi, Kayah li, Kharoshthi, Khojki,
</td></tr>
<tr><td align="left">
Khudawadi, Lepcha, Limbu, Mahajani, Makasar, Mandaic, Manichaean,
Marchen, Masaram Gondi, Medefaidrin, Meetei Mayek, Miao, Modi,
</td></tr>
<tr><td align="left">
Mongolian, Multani, Nandinagari, Newa, Nko, Nyiakeng Puachue Hmong,
Old Sogdian, Pahawh Hmong, Phags Pa, Psalter Pahlavi, Rejang,
</td></tr>
<tr><td align="left">
Saurashtra, Sharada, Siddham, Sinhala, Sogdian, Soyombo, Sundanese,
Syloti Nagri, Tagalog, Tagbanwa, Tai Le, Tai Tham, Tai Viet,
</td></tr>
<tr><td align="left">
Takri, Tibetan, Tifinagh, Tirhuta, Wancho, Zanabazar square,
</td></tr>
</table>>]
footnote[fontname="Verdana",label=<<table border="0" cellborder="0" cellspacing="0">
<tr><td align="left">† Feature is scoped to each syllable</td></tr>
<tr><td align="left">‡ All topographic features are scoped based on topographic position</td></tr>
</table>>];
notes3->footnote[style=invis];
}