Re: [i18n-activity] UAOA is not clear about justification and scope

• From: Khaled Hosny via GitHub <sysbot+gh@w3.org>
• Date: Tue, 03 Oct 2017 20:35:15 +0000
• To: public-i18n-archive@w3.org
• Message-ID: <issue_comment.created-333970075-1507062901-sysbot+gh@w3.org>

I don’t think the issue UAOA is trying to fix is limited to Quran. Take the very common mark sequence <U+0651 ARABIC SHADDA, U+064E ARABIC FATHA>, if you normalize it (NFC or NFD) it will come out in the reverse order:
```python
[hex(ord(u)) for u in unicodedata.normalize("NFC", "\u0651\u064e")]
['0x64e', '0x651']

[hex(ord(u)) for u in unicodedata.normalize("NFD", "\u0651\u064e")]
['0x64e', '0x651']
```

This means fonts and/or layout engines need to work around this, otherwise any process applying normalization to the text will break the order of the marks and subsequently their positioning, else they will be technically in violation with the spec since canonically equivalent strings are not treated the same.

-- 
GitHub Notification of comment by khaledhosny
Please view or discuss this issue at https://github.com/w3c/i18n-activity/issues/496#issuecomment-333970075 using your GitHub account

Received on Tuesday, 3 October 2017 20:35:04 UTC