Michael,
> >Michael Everson wrote:
> >
> >>Did you not read my draft paper proposing the solution for this
> >>feature of this script?
> >
> >Can't find it.
>
> http://www.evertype.com/standards/iso10646/pdf/tengwar-vowels.pdf
> http://www.evertype.com/standards/iso10646/pdf/tengwar.pdf
Maybe I haven't read these carefully enough, but it appears to
me that the analysis you provide in tengwar-vowels.pdf (which I
find myself in agreement with) doesn't match your statement about
the vowels (tehtar) in tengwar.pdf, where you claim the tehtar
are not combining marks, that the logical order is the same for
both Quenya and Sindarin modes, and that display is a matter of
picking out ligatures that ligate the tehtar with preceding base
letters in Quenya and following base letters in Sindarin.
>
> >In the CSUR there is a reference to smart keyboarding,
> >but if logical order trumps base-before-combining (which in practice
> >would mean saying that Tengwar vowel signs aren't combining at all,
> >but subject to andatory language-dependent ligatures), then what
> >we need is smart rendering (i.e. OpenType vel sim.)
>
> So I think, if we want underlying logical phonetic order.
It seems to me that the best approach to the encoding of Tengwar
would be to treat the tehtar as ordinary combining marks (which
in Cowan's terminology would be base-before-combining trumping
logical order). This gives a simple rendering model, avoiding
all the hassle of ligation. One then extends the notion of "logical
order" to mean something different for Sindarin than for Quenya.
For Quenya mode, logical order is just C V . C V . C V . and like a good
abjad, you place the vowelling mark over the consonant of its
syllable. If you need a syllable-closing consonant, you just treat
it like another defective syllable, with no tehtar:
C V . C . C V . C V .
For Sindarin mode, logical order means something different. The
syllables are treated as inherently CVC, but represented as ([C1 C2] V),
where the vowel is pronounced between C1 and C2, but marked on C2.
If you need an open syllable CV, you just omit the C1: ([ - C2] V).
To *read* Sindarin correctly, you have to see the syllables, and
bring forward the tehtar to pronounce before the consonant to which
it is applied. The *type* Sindarin correctly, you wouldn't need any
smart keyboarding -- you'd just need the same sense of syllabic
structure. Type C1 C2 V . C1 C2 V ., just the way the text looks.
This approach keeps the rendering simple, avoids the need for
any odd ligaturing or extensive use of ZWJ, keeps the keyboarding
simply related to the visible rendition, and makes the tehtar just
ordinary combining marks. Given that we are dealing here anyway
with a constructed script of limited (if passionate) use, it seems
that that would be the better approach to Unicode encoding than
trying to maintain some archetypal logical order that applies
equally to both Quenya and Sindarin. Let the complexity reside
instead in special filters that people could build to transform
text from Quenya mode to Sindarin mode, if they choose.
--Ken
This archive was generated by hypermail 2.1.2 : Thu Jan 03 2002 - 19:22:05 EST