I disagree that the question is this simple. It is not just a font issue.
Yes, it is.
It is a matter of the writing system being used.
The writing system used by Irish is the Latin script. Older orthography differs from modern orthography, regardless of whether the font has Roman glyphs or Gaelic glyphs.
For instance, since the spelling reform of the 1940s, Irish has represented lenition in a fundamentally different way than in the "traditional" system--i.e., a posterior 'h' vs. the diacritical dot.
This is well-known to all of us. It is an orthographic issue, not an issue of character encoding.
Thus, the digraph <0062>+<0068> (i.e., "bh") represents the same conceptual object as <1E03>.
We know.
Note that, if a selection of Irish text is set using one convention or the other, problems with spell checkers will occur UNLESS there is some metadata that indicates the writing system.
Script codes of ISO 15924 could be used to flag this, if people were interested in implementing it. Microsoft markets an Irish spell-checker (for modern orthography), but only on the PC platform, which is a bit of a blow to those of us who typeset in Irish. Free spell-checkers for the Mac are available on my web site. No spell-checkers exist for pre-Caighdeán orthography.
Marion's question [...] implies that "dotless i" and "i" are not the same character because the latter DOESN'T EXIST in the traditional writing system.
No, it doesn't. The question shows an ignorance of the character/glyph model and the facts of the development of writing in Irish. The dot on a Latin letter "i" is not a diacritical mark, it is a feature of some font styles. "Brian" and "Bržan" are the *same* name in Irish, with the *same* spelling. This is a different thing from Irish "sín" vs Scottish Gaelic "sìn" which are *different* spellings -- both differing from the word "sin". There is not, nor ever has there been, a third distinction between "sin" and "sžn".
Therefore, it's not a question of what font the document creator chooses; it's a matter of what system is chosen.
You mistake orthography and glyph choice with character identity. "Dotless i" as a *character* is used only in Turkic languages, has nothing to do with Irish, and never has.
To answer your question:
In the context of a document using traditional Irish orthography (which does not contain "i"), how can "dotless i" be preserved in plain text?
It may be preserved by the use of fonts without dots on the "i". It should not be preserved by spelling Irish with the letter used in Turkic language orthography, unless you don't want to spell-check or sort the data correctly.
--
Michael Everson * * Everson Typography * * http://www.evertype.com
