GNOME Bugzilla – Bug 112467
Hangul opentype module : Un series fonts
Last modified: 2007-06-20 20:06:36 UTC
Now that we have a GPL'd opentype font for Korean thanks to PARK Won-kyu. We need to write hangul-ot module. Perhaps, we can give up fixing bug 95708 and move here some of codes from the patch for that bug. Opentype fonts from MS for Korean (I haven't seen one, but judging from their description given at http://www.microsoft.com/typography/otfntdev/hangulot [1]) are problematic because their handling of Jamo sequences depend on whether corresponding precomposed forms are assigned separate codepoints or not. It's possible to write hangul-ot in such a way that it can be used with both Un-series OT fonts and MS OT fonts, but it'll complicate things (I have little clue as to why MS decided to do things this way...) [1] The page had apparently been up over a year as of early April, but no link to the page (as far as I know) had been available so that nobody outside MS had known its existence. Korean MS Office XP was reportedly shipped in summer of 2001 with Korean opentype fonts.
It looks like to me that Microsoft was trying to solve a different problem ... instead of trying to allow for small fonts for modern Hangul, Microsoft was trying to solve the problem of rendering old Hangul. I think the reason why small fonts weren't a goal for Microsoft is that for TrueType outlines, you can use the TT interpreter to make a glyph out of smaller glyphs. Without seeing the spec for the Un fonts, it's hard to say if supporting both them and the Microsoft OT spec is going to be a problem or not.
I'm afraid I failed to get through to you that my purpose of wanting to support Hangul OT fonts is exactly the same as that of MS. I'm doing this(and the predecessor of this bug, 95708) to support pre-1933 orthography Korean. Modern Korean is more or less (although not 100%) well supported by 'dumb' truetype fonts. 'Un-series' fonts have glyphs for 11,172 precomposed syllables and other characters in KS X 1001 in addition to OT layout tables (GSUB) for Hangul Jamos.
I forgot to give the URL for UnBatang font. It's available at http://www.i18nl10n.com/fonts/UnBatangOdal.ttf Anyway, I plan to work on this in about two weeks. > if supporting both them and the Microsoft OT spec is going to > be a problem or not. It's possible, but certainly makes things complicated because of MS's stupid decision to treat 'precomposed' Jamo clusters with separate code points (modern and obsolete/archaic) from Jamo clusters(modern and obsolete/archaic) without separate code points. > Microsoft was trying to solve the problem of rendering > old Hangul. This does not justifiy their hard-to-understand decision that I described above. See Appendix B of their spec. It's not correct nor easy to do things their way. They had to go extra lengths to do the *wrong* things. Moreover, Uniscribe (as it's shipped) doesn't render a decomposed sequence of Jamos (e.g. U+1100, U+1161) the same way as the corresponding precomposed syllable (U+AC00), which is a clear violation of the Unicode standard. > you can use the TT interpreter to make a glyph out > of smaller glyphs. Un-seris fonts take advantage of this (composite glyphs) as well for 11,172 precomposed syllables reducing its size compared to that of Baekmuk fonts (that do not use composite glyphs.)
Created attachment 20174 [details] [review] v 0.1 patch (not working)
This is the first attempt (patterned after Noah's patch for bug 101079). It's not working because it doesn't even detect UnBatang's OT ljmo/vjmo/tjmo features. I just included the patch for hangul-fc.c so that the patch for bug 95708 (except for tables-ext-jamos.i and hangul-fc.c) has to be applied as well.
Bumping to 1.6.0 since we don't have a finished patch yet. Can you merge the useful parts of bug 95780 into this patch since that has been WONTFIX'ed?
Any updates here?
If this has not been an issue for the past four years, I imagine we don't need it anymore.