OK, it turns out the characters were entered thru Word - insert symbol, which apparently used the presentation forms b block.
I used the ICU transliterator to normalize with NFKC and convert the strings to the Arabic block equivalents, which seemed to work nicely. The Egyptian now displays correctly (to my untrained eye) on Opera 6, Netscape 6, Mozilla 1.1 and IE. The claims that NS and Opera did not yet fully support bidi undermined fixing this sooner, which is a shame. They seem to support bidi well enough for the purposes of these examples. I am not sure why IE worked so well, and wonder if it normalizes first internally. (Kill one rumor, start another...) Using NFC normalization wouldn't have converted these. As excepting these submissions by email is not much different from accepting them over the web, it is going to be an problem for some applications to determine if the use of compatibility forms by some users is intentional or not, with the subsequent problems for searching, matching, etc. I am not sure which display is correct now for the presentation block forms. This will also be a problem if it varies by browser. Oh well, I have some nice examples for the normalization section of the Web I18n tutorial now. I'll need to find a validator or other tool that will warn me when presentation forms are used, and/or be able to convert the file. Maybe if ICU accepted files... ;-) Only the html file is fixed. I'll update the pdf later. Also Cherokee was added, fyi. tex Tex Texin wrote: > > For some reason the Egyptian example reverses under the Opera browser, > but the other RTL scripts do not. > Can someone spot a reason this would occur? > > http://www.i18nguy.com/unicode-example.html -- ------------------------------------------------------------- Tex Texin cell: +1 781 789 1898 mailto:[EMAIL PROTECTED] Xen Master http://www.i18nGuy.com XenCraft http://www.XenCraft.com Making e-Business Work Around the World -------------------------------------------------------------
