Re: Suppressing Ligation of Spacing Marks

inserting some zero-width word joiner or disjoiner should work with this...
But if you see a dotted circle, you need to encode some zero-width space as
the base holder for the combining vowel sign following it.

However I wonder if fonts accept zero-width holders for combining vowels,
they could still assume that there's no matching base consonnant and thus
insert another base dotted circle.

There's no consensus across script for using the same null-base holder
acting as a pseudo-consonnant for vowels encoded after them (e.g. Hangul
has its own jamo holder for this because of its specific algorithmic
composition, but some other scripts also use such null holders for their
own orthography).. In Alphabetic scripts, the ZWNJ should work.

But in Indic scripts we are all depending on the capability of renderers to
support specific scripts with only specific subsets of base letters and
every other character outside this subset will trigger the insertion of a
dotted circle glyph, and ZWJ/ZWNJ is already specific for being used in
script-specific clusters for some distinctions (notably to control how
parts of clusters are subgrouped ...)

You'll need to "bug" the maintainers of the renderer if they forgot
necessary cases described earlier for the script when it was initially
approved for encoding.

2016-11-08 10:09 GMT+01:00 Richard Wordingham <
richard.wordingham_at_ntlworld.com>:

> Should it be possible to suppress the ligation of a base character and
> a visually following spacing mark in plain text?
>
> The example I have in minf is the sequence <U+1A36 TAI THAM LETTER NA,
> U+1A63 TAI THAM VOWEL SIGN AA>. It may be desirable to suppress the
> ligation because both ligands have subscript consonants. However, if
> I write <NA, ..., ZWNJ, SIGN AA, ...>, the Universal Shaping Engine
> decides that the ZWNJ triggers a new syllable, and inserts a dotted
> circle before SIGN AA. (The dotted circle after SIGN AA results from a
> failure to read the proposal for the Lanna script as it was then
> called.)
>
> Richard.
>
>
Received on Tue Nov 08 2016 - 20:42:54 CST