In message <[log in to unmask]> Francois Lachance <[log in to unmask]> writes: > In reference to Jean-Luc's question about language and transliteration, > would it be tag abuse to consider the "lang" attribute? > > That is, declare within the TEI Header something like > > <language id="romgr">Greek in romanized transliteration </language> > <language id="gr">Greek in greek characters</language> > > > and then at the appropriate spot in the file > > <w lang="romgr"> ... </w> The lang attribute is not available in P5. We have gone over to using xml:lang, as documented in the CH chapter of P5. > > > Alternatively, if several languages represented in the file have > transliterated forms... > > <w lang="gr" ana="translit"> ... </w> > <w lang="jp" ana="translit"> ... </w> > > where any instance a transliterated language segment can be referenced > via a link to a note on transliteration. > > <note id="translit">Romanized transliteration </note> > > This of course could accommodate references to several types of > transliteration schemes (e.g. Chinese in Wade-Giles or in Pinyin; > Tibetan in Roman or Chinese script). > > > BTW, the example in draft P5 doesn't yet address multiple transliteration > schemes at work in one document. > Do you mean that there isn't an example of their use? The spec certainly permits it, as you show in your examples below. > See > http://www.tei-c.org/P5/Guidelines/ref-language.html > > <langUsage lang="en-US"> > <language ident="en-US" usage="75">modern American English</language> > <language ident="i-az-Arab" usage="20">Azerbaijani in Arabic > script</language> > <language ident="x-lap" usage="05">Pig Latin</language> > </langUsage> > > > However, P5 in Section 4.1 "Language Identification" gives some > examples of languages, scripts and and transliteration schemes ... > > zh-Latn-x-pinyin (Chinese transcribed in the Latin script using the Pinyin > system) > > > so this may be more appropriate: > > > <language id="gr-Latn">Greek in romanized transliteration </language> > <language id="gr">Greek in greek characters</language> > ... which was my original suggesiton to J-L.