Re: Mark-Driven Script Categorisation

From: Philippe Verdy <verdy_p_at_wanadoo.fr>
Date: Fri, 18 May 2012 03:15:44 +0200

So all this looks like a legacy, coming from ages where Unicode was
not very developed. I wonder what is the effective origin (and time)
of this Lao keyboard. If it comes from a mechanical typewriter layout,
it would be interesting to see how the glyph looked like : did it
really have serifs (like most mechanical typewriters had) ?

Are serifs used in Lao prints ? How does it work when a multilingual
Lao text mixes Latin words in Lao text : are these "x" used as Lao
vowels bearers styled differently ?

If so, it would be highly preferable to encode it as a plain Lao
letter. The Lao keyboard layout is then just an old legacy that could
be revised to use a better assignment. Otherwise the solution would be
that Lao fonts define a special subsitution rule when it is used as a
bearer, in order to offer compatibility with the legacy, so that it
looks the same as the multiplication sign with which it was visually
unified on the keyboard, even if a more modern text encoding would use
an effective Lao letter rather than the multiplication sign or the
Latin x, for use as a vowel bearer. In that case, it would collate at
the beginning of the Lao alphabet, but after vowel marks.

Another solution would also be to encode it as a cross-script symbol
for SE Asian scripts, except that it would still be considered a
letter in its properties, and possibly elligible in identifiers
(including as the initial).

For now the Latin letter x just breaks every word breakers (or Lao
would require a special rule like Japanese to allow mixing scripts,
plus various adjustments for text parsers and in renderers, as well as
in fonts that combine Lao, Basic Latin and ASCII punctuation and
symbols).

2012/5/18 Richard Wordingham <richard.wordingham_at_ntlworld.com>:
> On Thu, 17 May 2012 23:16:10 +0200
> Philippe Verdy <verdy_p_at_wanadoo.fr> wrote:
>
>> > OK, OK.... So this looks like there's an 'x'-like letter in the Lao
>> > script. But why should it be the Latin letter with all its allowed
>> > variations, its dual case, its cursive joining, its serifs ? May be
>> > the letter x was chosen because it has the property of a letter and
>> > not a mere symbol. But then this means that the Lao script needs a
>> > new letter of its own.
>>
>> Correction: the Shift state of this keyboard is used only for some
>> rare Lao letters, but the rest is filled by punctuation and other math
>> symbols. When looking at the keyboard map, I actually don't see an x
>> letter but effectively a multiplication sign (just like there's a plus
>> sign two rows above).
>
> That's what I originally thought - but every implementation I've
> investigated comes up with the letter 'x'. I have seen 'x' with serifs
> in lists of Lao vowels, but I'm not sure that the Lao vowels were
> actually typed. I must say 'x' works very well as a bearer for SE
> Asian combining marks.
>
> I originally raised the issue because I am not sure what having a
> letter reclassified as common would do to the definition of CLDR
> tailorings. The Garshuni writing system could have an interesting one
> - Syriac consonants with Arabic marks - but I suspect Arabic vowels
> being primary ignorables prevents any ridiculous results.
>
> Richard.
>
Received on Thu May 17 2012 - 20:17:44 CDT

This archive was generated by hypermail 2.2.0 : Thu May 17 2012 - 20:17:45 CDT