Draft Revisions to UASG014 - Quick Guide to EAI
The attached revisions stem from our session in Redmond in April of this year. Thanks to Mark for applying the edits. Would you please review these and compare them to your notes/recollections. We have a number of other points to consider: The most user-friendly aliasing methods (between the EAI address and the ASCII address that point to the same user mailbox) may be script-specific. UASG should make a proposal for a few scripts and let communities handle the rest. Email service providers should offer support for SPF/DKIM. Old-school Spam filters will be challenged with new scripts. Reputation systems like DKIM are a good mitigation. Franck was going to write up something for this. Mark will follow up with his anti-spam folks (Terry Zink) to make sure that U-Labels are compatible with DKIM. We need to think about digital signatures. A user with aliases will have multiple signatures. Are there any usability issues related to that? Should we include anti-abuse safety issues in our recommendations? We are also looking at this topic with respect to the Quick Guide to Linkification. We are waiting till after the forthcoming M3WAAG meeting in Lisbon before making a decision on this topic, but your thoughts would be welcome. I would welcome feedback by the 10th, please. Don Don Hollander Universal Acceptance Steering Group Skype: don_hollander
Hello Don, On 2017/06/03 19:19, Don Hollander wrote:
We have a number of other points to consider:
The most user-friendly aliasing methods (between the EAI address and the ASCII address that point to the same user mailbox) may be script-specific. UASG should make a proposal for a few scripts and let communities handle the rest.
Or they may be language-specific, or they may be individual preferences of the users. As an example, the fallback for "ü" in German would be "ue", but in French or Spanish, it would be "u". But there might be individual exceptions, where people have different preferences. For most non-Latin scripts, there are at least a few different transliterations or transcriptions, and they may differ by language. Some communities may be able to find a consensus on a single way of how to do it, but others may not. And each user may have their own preference. In addition, some scripts and some languages are not very phonemic (think about English, where the relationship between pronunciation and spelling is quite loose). This creates additional issues when trying to convert to ASCII with some simple rules. The toughest example may be Japanese. There are not really any rules about how to pronounce or transcribe names. The same Kanji spelling may be read in various different ways. Family names may be somewhat easier, because they are essentially a closed set (although with a long tail). But for given names, parents can (and do) make up names at will, and they can just take some Kanji they like, and can assign them a pronunciation they like. Absolutely impossible to cover that with rules, even before we get into the question of what transcription method to use. Regards, Martin.
That's interesting. I've encountered a few names with pronunciations based on archaic soundings of specific kanjis, but never realized that there was so much freedom to assign one's own preference. Then again, I grew up as "swan ZAIR ik", and have dabbled with several alternate pronunciations since. That always seemed like an option since Š and č and aren't familiar to Americans at all.... I'd think Kanji would be more "enforced". /marksv -----Original Message----- From: ua-eai-bounces@icann.org [mailto:ua-eai-bounces@icann.org] On Behalf Of Martin J. Dürst Sent: Sunday, June 4, 2017 8:08 PM To: Don Hollander <don.hollander@icann.org>; ua-eai@icann.org Subject: Re: [UA-EAI] Draft Revisions to UASG014 - Quick Guide to EAI Hello Don, On 2017/06/03 19:19, Don Hollander wrote:
We have a number of other points to consider:
The most user-friendly aliasing methods (between the EAI address and the ASCII address that point to the same user mailbox) may be script-specific. UASG should make a proposal for a few scripts and let communities handle the rest.
Or they may be language-specific, or they may be individual preferences of the users. As an example, the fallback for "ü" in German would be "ue", but in French or Spanish, it would be "u". But there might be individual exceptions, where people have different preferences. For most non-Latin scripts, there are at least a few different transliterations or transcriptions, and they may differ by language. Some communities may be able to find a consensus on a single way of how to do it, but others may not. And each user may have their own preference. In addition, some scripts and some languages are not very phonemic (think about English, where the relationship between pronunciation and spelling is quite loose). This creates additional issues when trying to convert to ASCII with some simple rules. The toughest example may be Japanese. There are not really any rules about how to pronounce or transcribe names. The same Kanji spelling may be read in various different ways. Family names may be somewhat easier, because they are essentially a closed set (although with a long tail). But for given names, parents can (and do) make up names at will, and they can just take some Kanji they like, and can assign them a pronunciation they like. Absolutely impossible to cover that with rules, even before we get into the question of what transcription method to use. Regards, Martin. _______________________________________________ UA-EAI mailing list UA-EAI@icann.org<mailto:UA-EAI@icann.org> https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fmm.icann.or...
participants (3)
-
Don Hollander -
Mark Svancarek -
Martin J. Dürst