From: verdy_p (verdy_p@wanadoo.fr)
Date: Tue Aug 18 2009 - 15:46:40 CDT
"Karl Pentzlin"
> When discussing this, it should be also taken in consideration that
> some letters which were Greek in first line, are used now within Latin
> orthographies of living languages.
> As far as African orthographies are concerned, these letters usually are
> encoded as Latin letters with separate codepoints, like
> U+0251 LATIN SMALL LETTER ALPHA or U+0269 LATIN SMALL LETTER IOTA.
Is that true for the African capital Latin Esh letter, which really looks like a capital Greek Sigma (but is very
distinct of it, and must not be confused with its lowercase counter part (that looks too much like Latin letter o
with a hook used in Vietnamese) ?
The Latin small esh is part of IPA, given that it does not use any capitals (however it does uses small capitals).
IPA if it was considered as a script by itself (and fully encoded as such) would be strictly monocameral,
contrasting with Latin
The problem is that now we effectively have some languages that have adopted IPA symbols as part of their regular
Latin orthography, and they now want also their associated capitals (the same also happened temporarily with a few
Cyrillic-written languages, until these letters were added to Cyrillic itself, and the use of IPA/Latin letters
derecommanded). Unfortunately, the case mappings are stabilized and can't be added in the UCD if the capitals are
added, except through language-specific tailorings.
You cannot make now the IPA script really distinct from the Latin script itself, except for a few symbols that may
be encoded distinctly when the Latin letters are ambiguous. but even in this case, you can still avoid the complete
disunification.
Using variation selectors with Latin letters (or even Greek) for IPA purists is a good solution for me (and it has
many practical advantages compared to new separate symbols), including for full-text searches in plain-text,
including in legacy documents already encoded, or for font designers pressed by users that want a working single-
font solution for texts mixing Latin and IPA, or for typography purists; we do have the working tools for them now).
But it's true that for the Greek letters that were borrowed into IPA, I would prefer that IPA remains fully self-
contained within the Latin script (and unified with it); and so that these IPA Greek symbols effectively get encoded
separately as Latin letters.
So variation selectors may effectively be used in IPA only after Latin letters, if this is needed for IPA typography
purists, UNLESS there are contrasting uses between letter forms that are naturally unified in the regular Latin
script (like the two forms of 'a', that should require, in my opinion, a separate encoding of the IPA "Latin small
letter a with handle" contrasting cleanly with the IPA "Latin small letter alpha" to really avoid the rendering
ambiguities that VS's do not solve cleanly after the regular "Latin small letter a", for example when IPA is used to
exhibit French phonology in dictionnaries that have been created in plain text and rendered on the web with
unpredictable fonts, even when CSS is used):
This position is exactly the same position adopted for letter-like mathematic symbols, that were added for the same
reasons that the use of VS do not solve cleanly, as the results of fonts and renderers used by the recipient of the
encoded texts are unpredictable and unenforcable but still perfectly correct for rendering normal texts (notably in
HTML or even in HTML+CSS, where the enforcement of fonts and other rendering styles also creates a usability problem
that can be solved cleanly ONLY if the text is not ambiguous as plain-text only, without the added styles).
This archive was generated by hypermail 2.1.5 : Tue Aug 18 2009 - 15:48:35 CDT