On Mon, 26 Aug 2002, Doug Ewell wrote:
> Jungshik Shin <jshin at mailaps dot org> wrote:
> > disappointed. At first, I was intrigued by their claim that it
> > supports Hangul Jamos. I've seen some false claims that Hangul
....
> > that claim. It just treats them as 'spacing characters' instead of
> > combining characters. Basically, it's useless except for making
> > Unicode code chart (so is Arial MS Unicode.)
>
> This is one of those cases where the verb "support" is so flexible that
> it loses meaning.
IMHO, their use of 'support' (for Hangul Jamo and various
S&SE Asian scripts) goes beyond the flexibility limit I perceive the
word to have.(well, then I'm not a native speaker of English while
you're ....) Wouldn't you return the product (or request refund) if you
purchased it (instead of downloading it free) believing that it supports
South and Southeast Asian scripts? I would at least put '*' next to that
'yes' mark and qualify what I mean by 'support'. What would you do if
it's your program?
> UniPad does include glyphs for individual jamos as
> well as precomposed Hangul syllables, which is more than most
> non-Korean-specific TrueType fonts can offer.
Which is still far from supporting Hangul Jamos.
I'm not sure if you can say that including _spacing_ Hangul Jamo glyphs
in their *bitmap* font is more than most non-Korean-specific TTF's can
offer. Because I know only four non-Korean specific TTF/OTFs striving
to be 'pan-Unicode' (or pan-BMP) fonts and at least two of them
(Arial MS Unicode and CODE2000. Cyberbit and Titus may or may not.)
include _spacing_(i.e. useless) Hangul Jamo glyphs as well as glyphs
for precomposed Hangul syllables. Besides, James Kass wrote to me that
next release of CODE2000 would include glyphs for Hangul Jamos that can
be overstruck over each other to form syllables. Simple overstriking
works by making glyphs for leading consonants, vowels and final consonants
occupy disjoint portions of a square cell. The result is not very pretty
but is similar to the way old mechanical Korean typewriters worked and
is certainly a good beginning. Thai script can be supported in a similar
manner (ftp://ftp.nectec.or.th/pub/thailinux/cvs/docs/thaisupp/thaisupp.html).
> But it does not provide
> any mechanism for combining jamos into syllables, which of course is
> required for proper handling of Korean. Again, I don't know of any
> other mainstream Windows tools or fonts that can do this either
> (although I'm sure there are Korean-specific tools that can).
Uniscribe, MS IE 6 and MS Office XP count as mainstream
Windows 'tools', don't they? As for fonts, you're certainly right. I
have yet to see a single OTF with the proper support of Hangul Jamos.
(MS Word 2000 came with a couple of Korean fonts with about 2000? Middle
Korean precomposed syllables in PUA.) Anyway, according to what Seuk
Soo SUNG with MS Korea wrote to this list and to me off the line about
a year ago, Uniscribe shipped in Office XP(and Windows XP. perhaps with
MS IE 6 as well) can do that just fine _provided_ that OT fonts with
necessary gsub/gpos and other tables for Hangul Jamos are present on the
system. (Yudit manages to do the same with fonts I'd call 'precursors'
to those fonts.) With those fonts and the newest Uniscribe, MS IE 6 and
MS Word 2002 can handle over one million syllables (made out of Hangul
Jamos). Unfortunately, those fonts are only shipped with _Korean_ version
of MS Office XP. I misunderstood what he wrote about fonts and purchased
US Office XP for the sole purpose of testing it mistakenly assuming that
any language version of Office XP comes with at least one such font.
> > Then, I found its claim that it supports 300 languages(scripts). Wow !
> > Does it properly support various South and Southeast Asian scripts?
> > Again, it does not. It treats combining characters as spacing
> > characters.
> UniPad never claims to support 300 scripts. I'm not even sure there are
> 300 scripts. Probably half of the 300 "supported languages" are written
> with the Latin script.
> But again, Jungshik has a good point that true
> "support" for Devanagari, Khmer, etc. really does imply shaping and
> combining behavior, similar to what UniPad already provides for Arabic.
All right. 300 languages. Anyway, I'd never put 'yes' marks
(I believe you wouldn't either) for languages like Devanagari,
Thai, Bengali, Tibetan, and Lao if my program does not support
reordering/shaping/combining for scripts used to represent those
languages. Suppose some people living in a remote village of Laos and
Tibet stumbled upon SC Unipad web page and were thrilled to find that
it supports their languages. How much would they be disappointed if they
downloaded it (via a painstakingly slow link) and found that it did not.
> 300 scripts. Probably half of the 300 "supported languages" are written
> with the Latin script.
As we all found out in a recent thread on Middle English,
even supporting Latin alphabet based languages requires supporting
non-spacing combining characters, which SC Unipad can't as of 0.99. To be
fair, it appears to be able to render a sequence of 'base char + combining
char' with the glyph for the corresponding precomposed character.
> > it works in MS Windows as well as in Unix/X11.
>
> I haven't downloaded it yet, so I haven't seen whether this is true. I
> have my doubts, however, based on release notes like the following:
Let's put aside the view of the author of Yudit on MS Windows. You can
believe me(maybe, you've already tested it by now). It works exactly
the same way on both Windows and Unix/X11 as far as Unicode editing is
concerned and supports a large subset of languages SC Unipad claims it
supports but doesn't.
> > lacks support for your script/language and you can code, you may be
> > able to add it yourself either for yourself or with the author's help
> > as I did for Hangul Jamos.
>
> "If you can code" is a big stumbling block for anyone who is not a
> programmer.
It is, but it is not a bigger stumbling block than having to
wait indefinitely after asking (or even begging if it's spoken/used by
only a small number of people so that invisible hands don't work favorably
for it)) developers of closed-source programs to support one's script.
Jungshik
This archive was generated by hypermail 2.1.2 : Tue Aug 27 2002 - 05:49:01 EDT