Thanks for your reply James, and really, your last line is crucial.
> I think the
> reason you are doing this is going to be useful in determining how you
> should mark it up.
> However, thinking about it, why do you need to record that it is in
> IPA? What is the end use that knowing that will be put to?
I think that this procedure will be most useful long before
publication, insofar as it indicates an area of text that needs to be
carefully proofread as OCR (at least our OCR) is incapable of
rendering IPA well. So it will serve to mark chunks that need further
While I can see there might be certain benefits of tracing the use of
IPA in a corpus of ling. texts for the history of linguistics, this in
itself would not justify the labor of tagging IPA -- were it not for
the fact that it would be needed to be marked anyway during the
proofreading stage. But since we have to mark it anyway, why not mark
it in a meaningful way?
What is clear to me is that the phonetic transcriptions are not part
of the meta-text about language (nor are the square brackets
> Is it that you are worried about font support? Is it that
> the text is going to be read aloud by machine? Or is it simply that
> you want it styled differently in a print rendering?
Yes, font support is certainly a concern, although I imagine that by
the time we reach publication it will be less of one given the speed
at which Unicode is becoming standard. Still, we have had a number of
issues with utf-8 encoding already -- especially with MS Word -- that
we have never completely understood.
The idea of the text being read aloud by machine is an added benefit
that I had not thought about.
Thank you very much for the suggestion that I look at SSML and CSS3.
I will do so.
[<q xml:lang="fr" alphabet="ipa">trodœkɛstjɔ̃</q>]
<q rend="PRE [ POST ]" xml:lang="fr" alphabet="ipa">trodœkɛstjɔ̃</q>
with unicode nice and mashed up now through several different systems!
(In retrospect, I recognize that it should have been a closed /o/ in
"'trop' de questions"... for the phonetics gurus out there who might
have been irked :-)