Re: IPA bugs in Windows?
|From:||And Rosta <a.rosta@...>|
|Date:||Saturday, February 16, 2002, 3:37|
> One of the features of the Tirelhat dictionary / word usage guide that I'm
> slowly putting together is the pronunciation of each word in IPA. I thought
> it would be cool to include the pronunciation on the English side as well.
> But then I noticed a U+019E LATIN SMALL LETTER N WITH LONG RIGHT LEG in one
> of my pronunciations where I had put a syllabic N [n=]! It's odd that this
> obsolete symbol would show up anywhere. It certainly doesn't make sense for
> _English_ syllabic N as in "button" (American pronunciation), which is a
> different thing entirely from the Japanese sound it was originally designed
> for. The _really_ odd thing is that as far as Microsoft Word knew, it still
> _was_ an "n" with a vertical line under it! And deleting the "n" makes the
> vertical line show up normally.
> Okay, I thought: I can live with syllabic "n" being displayed as this
> obsolete character, since the internal representation is correct. But the
> real evil of this bug came up when I tried to represent the Tirelhat
> affricates. Since there's a distinct difference in length between
> affricates and clusters in correctly pronounced Tirelhat ("c" should be
> around 120 ms long, compared with 160 ms for a "ts" cluster that might come
> up in a compound word), I thought it would be nice to use the U+0361
> COMBINING DOUBLE INVERTED BREVE (following the IPA recommendation). But
> when a document containing the sequence 't' U+0361 's' is saved and
> reloaded, ***the 't' is displayed as a lowercase u with grave accent and
> the U+0361 vanishes!*** Again, deleting the 't' makes the tie bar show up.
> And all of this is happening with the Thryomanes font, which is entirely a
> normal TrueType font: it doesn't have any TrueType Open tables for glyph
> substitution and positioning or any such thing! (It happens with other
> fonts too, but most of them don't even have a tie bar in the character
Using Word 2000 under Windows 98, and the Livagian version of Thryomanes,
both combinations appear on screen correctly after saving and reopening.
Can I ask about the latest bug to annoy me? In Livagian Thryomanes, a
word containing italic <ij> will be truncated after the <i> when
printing. Likewise, in the (rather lovely) Garamond that comes with
Win 98, a word containing <ly) (I think it is) will be truncated
after the <l> when printing. Everything looks fine on screen. Does
anyone know where the problem might lie? Since these are my two favourite
fonts, the problem is rather annoying.