>>There are superscript International Phonetic Alphabet characters which were
>not included to support any particular character sets so far as I know, but
>phonetic entities like aspiration (h). We (Finland, Norway, Ireland) are
>preparing a proposal for Finno-Ugric Phonetic Alphabet support which >contains
rather a lot of superscript, subscript, and small-capital letters, >and whose
semantics are completely different from the plain letters, and >must be
distinguished from them in plain text (for lexicographical
^^^^^
>searching, etc.).
>Why in plain text? This is an obvious application for developing an XML tagging
scheme or some other form of markup.
Representing an organisation that makes heavy use of phonetic transcriptions,
and being in the position of supporting hundreds of linguists that work with
this stuff, I can assure you that the last thing they want is to have their
phonetic/phonemic data be XML-tagged. Just as you wouldn't want this sentence to
be encoded as follows:
<uc>j</uc>ust as you probably <contr>would not</contr> want this sentence to be
encoded as follows:
You can do any process on the latter that you could on the original, but the
algorithm needs to be modified. If you're doing it once, that's fine. But if you
frequently make up new processes, you'd probably rather just have the plain text
rather than have to parse XML in addition to dealing with the plain text. And
this little example is tame in comparison with what would be involved with
phonetic representations.
Peter
This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:20:48 EDT