On 2019-01-28 8:58 PM, Richard Wordingham wrote:
> On Mon, 28 Jan 2019 03:48:52 +0000
> James Kass via Unicode <unicode@unicode.org> wrote:
>
>> It’s been said that the text segmentation rules seem over-complicated
>> and are probably non-trivial to implement properly. I tried your
>> suggestion of WORD JOINER U+2060 after tau ( γένοιτ’ ἄν ), but it
>> only added yet another word break in LibreOffice.
>
> I said we *don't* have a control that joins words. The text of TUS
> used to say we had one in U+2060, but that was removed in 2015. I
> pleaded for the retention of this functionality in document
> L2/2015/15-192, but my request was refused. I pointed out in ICU
> ticket #11766 that ICU's Thai word breaker retained this facility. ...
Sorry for sounding obtuse there. It was your *post* which suggested the
use of WORD JOINER. You did clearly assert that it would not work. So,
human nature, I /had/ to try it and see.
It. did. not. work. (No surprise.) But it /should/ have worked. It’s a
JOINER, for goodness sake!
When the author/editor puts any kind of JOINER into a text string,
what’s the intent? What’s the poînt of having a JOINER that doesn’t?
Recently I put a ZWJ between the “c” and the “t” in the word
“Respectfully” as an experiment. Spellchecker flagged both “respec”
and “tfully” as being misspelt, which they probably are. A ZWNJ would
have been used if there had been any desire for the string to be *split*
there, e.g., to forbid formation of a discretionary ligature. Instead
the ZWJ was inserted, signalling authorial intent that a ‘more joined’
form of the “c-t” substring was requested.
Text a man has JOINED together, let not algorithm put asunder.
