Martin -- completely agreed. But I do know that the experts here at Microsoft (Murray Sargent as the key one) are still actively working on this. For our products, the ability to detect and properly display IRIs, IDNs, and EAIs is very much expected by our users. The current subscription version of Word has the best example of our current thinking. It may be necessary to specify a full protocol (mailto:) in order to trigger recognition of some EAIs at this time. The @Martin J. Dürst notation is certainly also currently implemented though only within the Outlook product at this time (I believe). Note that much of our model grew out of our early support of the file: protocol dating from Windows 95 allowing folder and file names in any writing system. And, please do not hear this as Microsoft is trying to set a standard. That's not what I mean at all. Murray serves on the Unicode groups that work with the BiDi algorithm and this remains a passion area for him. (Prior to my current role as a globalization specialist, I was the program manager lead for our text foundation systems in Office so am reasonably familiar with our end-to-end text storage, editing, and rendering stacks). -Stuart -----Original Message----- From: Martin J. Dürst [mailto:duerst@it.aoyama.ac.jp] Sent: Tuesday, April 4, 2017 6:59 PM To: Stuart Stuple <stuartst@microsoft.com>; Paul Borokhov <borokhov@apple.com>; Don Hollander <don.hollander@icann.org> Cc: Andrew Sullivan <ajs@anvilwalrusden.com>; ua-eai@icann.org Subject: Re: [UA-EAI] arabic1.arabic2@arabic3.arabic4 On 2017/04/04 06:41, Stuart Stuple via UA-EAI wrote:
I believe for plain text with no fancy Unicode values, the rendering is well-defined because the period and ampersand are both neutral and assume directionality from the surrounding characters. See https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwww.w3.org%2....
Add Unicode directionality controls and you get a wide variety of orderings – but that’s as true of English as it is of Arabic. See https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fblogs.msdn.....
When I last heard (several years ago), the Unicode consortium was still discussing whether to update the BiDi guidance to account for these cases.
"Hasn't decided on how" is definitely correct as far as I know, but "still discussing" is probably not, because I don't know of any ongoing discussion. It wouldn't be impossible to update the Unicode Bidi algorithm (which is a very clear algorithm, not just guidance) except for the fact that it's not clear how. Mail addresses, domain names, and IRIs (the version of URIs/URLs that contain non-ASCII characters) are somewhat difficult to identify in text. And then there are other artifacts that would need similar treatment, such as hash tags, the @somebody notation, and so on. And such notations keep coming up, and may also disappear. And there are cases where it's quite obvious to a human reader (assuming the reader can read Arabic or Hebrew or whatever other RTL script is in use) what the text says, but that's not necessarily easily put into a simple and straightforward (although already rather lengthy) algorithm. Regards, Martin.