G’day all In another world I heard there’s a major new release of mailman Does this support EAI? IF so, has ICANN deployed it? Best wishes for the new year Don Sent from my iPad
It appears that Don Hollander via UA-EAI <don.hollander@gmail.com> said:
G’day all
In another world I heard there’s a major new release of mailman
Hi, Don. The current version is 3.3.11. From version 2 to 3 was a compete rewrite, since then small incremental changes.
Does this support EAI?
Mailman has done a good job of internationalizing the messages and web pages, but I have no seen any work on allowing EAI subscribers. And honestly, I have no idea what it would mean for a list manager to support EAI in a meaningful way. It wouldn't be hard to allow UTF-8 addresses in subscriptions, but then there's the other 95% of the work. I suppose you could have EAI lists where all of the subscribers have to be on systems with EAI support, and non EAI lists that don't allow any non-ASCII addresses but that doesn't seem very useful. In the forseeable future we are going to have a mix of people on EAI and non-EAI mail systems, and on EAI mail systems some have ASCII addresses and some UTF-8 addresses. I have no idea how you would run a mailing list with a mix of EAI and non-EAI subscribers to exchange mail in a reasonable way and I would be surprised if anyone else did either. EAI is still very much a work in progress. R's, John
Hello Don and John If I understand correctly, just like this mailing list ua-eai@icann.org, it must be able to hold every kind of valid email address and successfully deliver email to all subscribers. That's our expectation. This functionality we have in xgenplus and we have been successfully running it since many years. For an example mailing eaiteam@xgenplus.com will also have EAI addresses and anyone from mailing lists sends an email to this list, it will be delivered to everyone from this list. Naturally, email delivery responsibility is from xgenplus mailing list to each individual email addresses in list everytime the email is sent, and it successfully delivers. Thanks Ajay On January 1, 2025 7:32:24 AM GMT+05:30, John Levine via UA-EAI <ua-eai@icann.org> wrote:
It appears that Don Hollander via UA-EAI <don.hollander@gmail.com> said:
G’day all
In another world I heard there’s a major new release of mailman
Hi, Don. The current version is 3.3.11. From version 2 to 3 was a compete rewrite, since then small incremental changes.
Does this support EAI?
Mailman has done a good job of internationalizing the messages and web pages, but I have no seen any work on allowing EAI subscribers.
And honestly, I have no idea what it would mean for a list manager to support EAI in a meaningful way. It wouldn't be hard to allow UTF-8 addresses in subscriptions, but then there's the other 95% of the work. I suppose you could have EAI lists where all of the subscribers have to be on systems with EAI support, and non EAI lists that don't allow any non-ASCII addresses but that doesn't seem very useful.
In the forseeable future we are going to have a mix of people on EAI and non-EAI mail systems, and on EAI mail systems some have ASCII addresses and some UTF-8 addresses. I have no idea how you would run a mailing list with a mix of EAI and non-EAI subscribers to exchange mail in a reasonable way and I would be surprised if anyone else did either.
EAI is still very much a work in progress.
R's, John _______________________________________________ UA-EAI mailing list -- ua-eai@icann.org To unsubscribe send an email to ua-eai-leave@icann.org _______________________________________________ By submitting your personal data, you consent to the processing of your personal data for purposes of subscribing to this mailing list accordance with the ICANN Privacy Policy (https://www.icann.org/privacy/policy) and the website Terms of Service (https://www.icann.org/privacy/tos). You can visit the Mailman link above to change your membership status or configuration, including unsubscribing, setting digest-style delivery or disabling delivery altogether (e.g., for a vacation), and so on. -- Sent from my Android device with XGen Email.
Don, That's a good news to see some progress. Looking forward to see this implemented and use our EAI addresses. Wishing you all Happy New Year. On January 1, 2025 5:35:22 AM GMT+05:30, Don Hollander via UA-EAI <ua-eai@icann.org> wrote:
G’day all
In another world I heard there’s a major new release of mailman
Does this support EAI?
IF so, has ICANN deployed it?
Best wishes for the new year
Don
Sent from my iPad _______________________________________________ UA-EAI mailing list -- ua-eai@icann.org To unsubscribe send an email to ua-eai-leave@icann.org _______________________________________________ By submitting your personal data, you consent to the processing of your personal data for purposes of subscribing to this mailing list accordance with the ICANN Privacy Policy (https://www.icann.org/privacy/policy) and the website Terms of Service (https://www.icann.org/privacy/tos). You can visit the Mailman link above to change your membership status or configuration, including unsubscribing, setting digest-style delivery or disabling delivery altogether (e.g., for a vacation), and so on. -- Sent from my Android device with XGen Email.
Hi, 3.10 still doesn't include EAI support. I think the only reason is that the changes involved are rather large. If I were the Mailman maintainers I'd plan an uninterrupted week to review and merge that, and that's a big chunk of time for a spare-time project like Mailman. Arnt On 01/01/2025, 01:06, "Don Hollander via UA-EAI" <ua-eai@icann.org <mailto:ua-eai@icann.org>> wrote: G’day all In another world I heard there’s a major new release of mailman Does this support EAI? IF so, has ICANN deployed it? Best wishes for the new year Don Sent from my iPad _______________________________________________ UA-EAI mailing list -- ua-eai@icann.org <mailto:ua-eai@icann.org> To unsubscribe send an email to ua-eai-leave@icann.org <mailto:ua-eai-leave@icann.org> _______________________________________________ By submitting your personal data, you consent to the processing of your personal data for purposes of subscribing to this mailing list accordance with the ICANN Privacy Policy (https://www.icann.org/privacy/policy <https://www.icann.org/privacy/policy>) and the website Terms of Service (https://www.icann.org/privacy/tos <https://www.icann.org/privacy/tos>). You can visit the Mailman link above to change your membership status or configuration, including unsubscribing, setting digest-style delivery or disabling delivery altogether (e.g., for a vacation), and so on.
It appears that Arnt Gulbrandsen via UA-EAI <arnt.gulbrandsen@icann.org> said:
Hi,
3.10 still doesn't include EAI support. I think the only reason is that the changes involved are rather large. If I were the Mailman maintainers I'd plan an uninterrupted week to review and merge that, and that's a big chunk of time for a spare-time project like Mailman.
I think the reason is that there is no way to have a mailing list with both EAI and ASCII subscribers because there's no way to downgrade an EAI message and allow ASCII users to reply. We've been thinking about this for a decade and nothing has worked. I suppose you could add EAI-only lists where you could only subscribe from a system with SMTPUTF8 support but that doesn't seem very appealing. R's, John
My take on EAI adoption and Mailman’s potential role is that, while Mailman has made substantive progress in messages and interfaces internationalizations, achieving full EAI support remains a significant challenge, as evidenced in Mailman 3.10. Managing mailing lists with mixed EAI and non-EAI subscribers introduces complexities that extend beyond enabling UTF-8 addresses, requiring seamless email delivery across diverse environments- a gap not yet fully addressed by current standards. As John pointed out, suggestions like ASCII fallback addresses, while practical, conflict with EAI’s goal of native non-ASCII support and risk suboptimal user experiences, as observed in early EAI trials. Promising alternatives, such as EAI-exclusive lists or validating EAI-readiness at registration, depend on broader ecosystem support, including reliable SMTPUTF8 adoption, highlighting the need for stronger industry-wide progress to reach critical mass. Regards, Dessalegn On Fri, Jan 3, 2025 at 9:35 PM John Levine via UA-EAI <ua-eai@icann.org> wrote:
It appears that Arnt Gulbrandsen via UA-EAI <arnt.gulbrandsen@icann.org> said:
Hi,
3.10 still doesn't include EAI support. I think the only reason is that the changes involved are rather large. If I were the Mailman maintainers I'd plan an uninterrupted week to review and merge that, and that's a big chunk of time for a spare-time project like Mailman.
I think the reason is that there is no way to have a mailing list with both EAI and ASCII subscribers because there's no way to downgrade an EAI message and allow ASCII users to reply.
We've been thinking about this for a decade and nothing has worked.
I suppose you could add EAI-only lists where you could only subscribe from a system with SMTPUTF8 support but that doesn't seem very appealing.
R's, John _______________________________________________ UA-EAI mailing list -- ua-eai@icann.org To unsubscribe send an email to ua-eai-leave@icann.org _______________________________________________ By submitting your personal data, you consent to the processing of your personal data for purposes of subscribing to this mailing list accordance with the ICANN Privacy Policy (https://www.icann.org/privacy/policy) and the website Terms of Service (https://www.icann.org/privacy/tos). You can visit the Mailman link above to change your membership status or configuration, including unsubscribing, setting digest-style delivery or disabling delivery altogether (e.g., for a vacation), and so on.
It appears that Dessalegn Yehuala via UA-EAI <mequanint.yehuala@gmail.com> said:
-=-=-=-=-=- -=-=-=-=-=-
My take on EAI adoption and Mailman’s potential role is that, while Mailman has made substantive progress in messages and interfaces internationalizations, achieving full EAI support remains a significant challenge, as evidenced in Mailman 3.10. Managing mailing lists with mixed EAI and non-EAI subscribers introduces complexities that extend beyond enabling UTF-8 addresses, requiring seamless email delivery across diverse environments- a gap not yet fully addressed by current standards. As John pointed out, suggestions like ASCII fallback addresses, while practical, ...
I wish people would stop saying this. ASCII fallback addresses are *not* practical. The experimental version of EAI had them and they were a failure. Mail programs didn't handle them well, and the two addresses kept leaking into the wrong contexts. And as many people have noted, if everyone has to have an ASCII address anyway why bother with EAI? Regards, John Levine, johnl@taugh.com, Primary Perpetrator of "The Internet for Dummies", Please consider the environment before reading this e-mail. https://jl.ly
Some of the systems integrated well ASCII fallback. However it was with experimental EAI, those newer systems have not created any mess in my tests. But, even if I want to agree with those who says that if you have EAI why do you need ASCII, it is literally because of that ASCII fallback that seems to be working in some systems a decent job. Otherwise, everyone will abandon EAI and just use ASCII, because EAI doesn't work (every time).
If you say so, but I have never seen mixed EAI and ASCII mail systems that really work. People end up using their ASCII addreses because they know they can write to everyone. At one of the ICANN tech days we heard about a community of EAI mail users in Thailand where everyone was writing in Thai. I also gather there are EAI mail communities in India where people are writing in Indian lanauges. That seems a lot more plausible than somehow figuring out one address at a time who you can write to with EAI addresses and who you can't, and trying to figure out impossible combinations of mail to groups of people some who have EAI addresses and some don't. Regards, John Levine, johnl@taugh.com, Primary Perpetrator of "The Internet for Dummies", Please consider the environment before reading this e-mail. https://jl.ly
John, Also, given the limitation I mentioned in my previous email I agree with you that viewing ASCII fallback as a practical approach is unlikely to serve as a long term solution. However, as Dusan explained, it has offered short-term "practicality" in certain scenarios. Instead, as I suggested earlier, approaches like validating EAI readiness at registration or implementing EAI-exclusive lists(with necessary safeguards to avoid exclusion) seem more aligned with the goals of EAI. Regards, Dessalegn On Sat, Jan 4, 2025 at 11:31 AM Dessalegn Yehuala < mequanint.yehuala@gmail.com> wrote:
My take on EAI adoption and Mailman’s potential role is that, while Mailman has made substantive progress in messages and interfaces internationalizations, achieving full EAI support remains a significant challenge, as evidenced in Mailman 3.10. Managing mailing lists with mixed EAI and non-EAI subscribers introduces complexities that extend beyond enabling UTF-8 addresses, requiring seamless email delivery across diverse environments- a gap not yet fully addressed by current standards. As John pointed out, suggestions like ASCII fallback addresses, while practical, conflict with EAI’s goal of native non-ASCII support and risk suboptimal user experiences, as observed in early EAI trials. Promising alternatives, such as EAI-exclusive lists or validating EAI-readiness at registration, depend on broader ecosystem support, including reliable SMTPUTF8 adoption, highlighting the need for stronger industry-wide progress to reach critical mass.
Regards, Dessalegn
On Fri, Jan 3, 2025 at 9:35 PM John Levine via UA-EAI <ua-eai@icann.org> wrote:
It appears that Arnt Gulbrandsen via UA-EAI <arnt.gulbrandsen@icann.org> said:
Hi,
3.10 still doesn't include EAI support. I think the only reason is that the changes involved are rather large. If I were the Mailman maintainers I'd plan an uninterrupted week to review and merge that, and that's a big chunk of time for a spare-time project like Mailman.
I think the reason is that there is no way to have a mailing list with both EAI and ASCII subscribers because there's no way to downgrade an EAI message and allow ASCII users to reply.
We've been thinking about this for a decade and nothing has worked.
I suppose you could add EAI-only lists where you could only subscribe from a system with SMTPUTF8 support but that doesn't seem very appealing.
R's, John _______________________________________________ UA-EAI mailing list -- ua-eai@icann.org To unsubscribe send an email to ua-eai-leave@icann.org _______________________________________________ By submitting your personal data, you consent to the processing of your personal data for purposes of subscribing to this mailing list accordance with the ICANN Privacy Policy (https://www.icann.org/privacy/policy) and the website Terms of Service (https://www.icann.org/privacy/tos). You can visit the Mailman link above to change your membership status or configuration, including unsubscribing, setting digest-style delivery or disabling delivery altogether (e.g., for a vacation), and so on.
On Mon, 6 Jan 2025, Dessalegn Yehuala wrote:
John, Also, given the limitation I mentioned in my previous email I agree with you that viewing ASCII fallback as a practical approach is unlikely to serve as a long term solution. However, as Dusan explained, it has offered short-term "practicality" in certain scenarios.
We'll just have to disagree about that. Can you point to actual working systems that use ASCII fallback? Tiny demos with one or two users don't count. Regards, John Levine, johnl@taugh.com, Taughannock Networks, Trumansburg NY Please consider the environment before reading this e-mail. https://jl.ly
Dear John, Ask Dr Ajay for demo account and see for your self... Or CoreMail, if I remember correctly.
Those let you pick which address to use when sending a single message, but if you want to send mail to both people with EAI and ASCII addresses, or have a useful mailing list with EAI and ASCII users, it does not work. Yes, it can send separate messages. No, people cannot reply in any normal way. I think we're done here. R's, John
6. 1. 2025. 18:58, John R Levine via UA-EAI <ua-eai@icann.org> је написао/ла:
On Mon, 6 Jan 2025, Dessalegn Yehuala wrote: > John, > Also, given the limitation I mentioned in my previous email I agree > with you that viewing ASCII fallback as a practical approach is unlikely to > serve as a long term solution. However, as Dusan explained, it has offered > short-term "practicality" in certain scenarios.
We'll just have to disagree about that. Can you point to actual working systems that use ASCII fallback? Tiny demos with one or two users don't count.
On Mon, 6 Jan 2025, Душан Стојичевић wrote:
I presume Ajay is here... Is this what John said correct for XGEN, Ajay?Personally, I don't remember picking up email addresses, I believe there was EAI checker before sending, but I am really not sure and I don't have a way now to test it. I am apologizing if I am wrong.
It may well probe the remote systems to see if they announce SMTPUTF8, but that does nothing to solve the other problems. Really, all it does is to persuade people that EAI is more trouble than it's worth. R's, John
6. 1. 2025. 22:02, John R Levine <johnl@taugh.com> је написао/ла:
> Dear John, > Ask Dr Ajay for demo account and see for your self... > Or CoreMail, if I remember correctly.
Those let you pick which address to use when sending a single message, but if you want to send mail to both people with EAI and ASCII addresses, or have a useful mailing list with EAI and ASCII users, it does not work. Yes, it can send separate messages. No, people cannot reply in any normal way.
I think we're done here.
R's, John
> 6. 1. 2025. 18:58, John R Levine via UA-EAI <ua-eai@icann.org> је написао/ла: > > On Mon, 6 Jan 2025, Dessalegn Yehuala wrote: > > John, > > Also, given the limitation I mentioned in my previous email I agree > > with you that viewing ASCII fallback as a practical approach is unlikely to > > serve as a long term solution. However, as Dusan explained, it has offered > > short-term "practicality" in certain scenarios. > > We'll just have to disagree about that. Can you point to actual working > systems that use ASCII fallback? Tiny demos with one or two users don't > count.
Regards, John Levine, johnl@taugh.com, Taughannock Networks, Trumansburg NY Please consider the environment before reading this e-mail. https://jl.ly
I already have an “xn—” email address. From: Душан Стојичевић via UA-EAI <ua-eai@icann.org> Sent: Monday, January 6, 2025 16:58 To: John R Levine <johnl@taugh.com> Cc: ua-eai@icann.org Subject: [EXTERNAL] [UA-EAI] Re: Mail man 3.1 It's true that problems like replying from ASCII user to EAI user will create chaos (no delivery for sure). Also, it's true that replying to alias - different address who sent initial email, is tricky and chaotic for mailing software (and users as well). And I also agree there's more issues (sorting and searching through a backup of emails, admin features - banning both email addresses etc). Anyway, thanks for chatting, at least we clarify some ideas, and we saw why we can't have EAI ready mailing software yet. And I still think "xn--..." would be better solution for user (mailbox) name in RFC, instead of SMTPUTF8, BUT we have what we have... Dušan 7. 1. 2025. 00:09, John R Levine <johnl@taugh.com<mailto:johnl@taugh.com>> је написао/ла: On Mon, 6 Jan 2025, Душан Стојичевић wrote:
I presume Ajay is here... Is this what John said correct for XGEN, Ajay?Personally, I don't remember picking up email addresses, I believe there was EAI checker before sending, but I am really not sure and I don't have a way now to test it. I am apologizing if I am wrong.
It may well probe the remote systems to see if they announce SMTPUTF8, but that does nothing to solve the other problems. Really, all it does is to persuade people that EAI is more trouble than it's worth. R's, John
6. 1. 2025. 22:02, John R Levine <johnl@taugh.com<mailto:johnl@taugh.com>> је написао/ла:
> Dear John, > Ask Dr Ajay for demo account and see for your self... > Or CoreMail, if I remember correctly.
Those let you pick which address to use when sending a single message, but if you want to send mail to both people with EAI and ASCII addresses, or have a useful mailing list with EAI and ASCII users, it does not work. Yes, it can send separate messages. No, people cannot reply in any normal way.
I think we're done here.
R's, John
> 6. 1. 2025. 18:58, John R Levine via UA-EAI <ua-eai@icann.org<mailto:ua-eai@icann.org>> је написао/ла: > > On Mon, 6 Jan 2025, Dessalegn Yehuala wrote: > > John, > > Also, given the limitation I mentioned in my previous email I agree > > with you that viewing ASCII fallback as a practical approach is unlikely to > > serve as a long term solution. However, as Dusan explained, it has offered > > short-term "practicality" in certain scenarios. > > We'll just have to disagree about that. Can you point to actual working > systems that use ASCII fallback? Tiny demos with one or two users don't > count.
Regards, John Levine, johnl@taugh.com<mailto:johnl@taugh.com>, Taughannock Networks, Trumansburg NY Please consider the environment before reading this e-mail. https://jl.ly
On Mon, 7 Jan 2025, Mark Svancarek (CELA) wrote:
I already have an “xn—” email address.
All the cool kids do. I'm xn--ls8ha@outlook.com. I must admit that your address is much more sophisticated and tasteful than mine.
And I still think "xn--..." would be better solution for user (mailbox) name in RFC, instead of SMTPUTF8, BUT we have what we have...
It's 40 years too late for that. R's, John
Sun Wukong & Son Goku are notoriously unsophisticated. But, yes, perhaps more tasteful. -----Original Message----- From: John R Levine <johnl@taugh.com> Sent: Monday, January 6, 2025 19:01 To: Mark Svancarek (CELA) <marksv@microsoft.com>; Душан Стојичевић <dusan@dukes.in.rs> Cc: ua-eai@icann.org; mark svancarek <xn--98sy4jmv0a@outlook.com> Subject: RE: [EXTERNAL] [UA-EAI] Re: Mail man 3.1 On Mon, 7 Jan 2025, Mark Svancarek (CELA) wrote:
I already have an “xn—” email address.
All the cool kids do. I'm xn--ls8ha@outlook.com. I must admit that your address is much more sophisticated and tasteful than mine.
And I still think "xn--..." would be better solution for user (mailbox) name in RFC, instead of SMTPUTF8, BUT we have what we have...
It's 40 years too late for that. R's, John
In this discussion we are missing the important point here. In mailing list communication there is no CC, BCC , it's always FROm mailingList to TO id, and hence there is no visibility required for ASCII and eai addresses to each other and hence no complications. It's not a normal one to one or one to many communication, in mailing list , it's receive from one and broadcast to everyone. The original sender server is not really talking to EAI address and will never do in mailing list case. I know it's bit complicated and need more deeper understanding. In one of the ICANN meetings, if required , someone from my team can demonstrate in tech day. Thanks Ajay On January 7, 2025 6:27:33 AM GMT+05:30, "Душан Стојичевић via UA-EAI" <ua-eai@icann.org> wrote:
_______________________________________________ UA-EAI mailing list -- ua-eai@icann.org To unsubscribe send an email to ua-eai-leave@icann.org _______________________________________________ By submitting your personal data, you consent to the processing of your personal data for purposes of subscribing to this mailing list accordance with the ICANN Privacy Policy (https://www.icann.org/privacy/policy) and the website Terms of Service (https://www.icann.org/privacy/tos). You can visit the Mailman link above to change your membership status or configuration, including unsubscribing, setting digest-style delivery or disabling delivery altogether (e.g., for a vacation), and so on. -- Sent from my Android device with XGen Email.
It appears that Dr Ajay Data via UA-EAI <ajay@data.in> said:
-=-=-=-=-=- -=-=-=-=-=-
In this discussion we are missing the important point here.
In mailing list communication there is no CC, BCC , it's always FROm mailingList to TO id, and hence there is no visibility required for ASCII and eai addresses to each other and hence no complications.
In the mailing lists we run, the From: address is the address of the author of the message, or sometimes a version of the address modified to survive DMARC rules. I know a lot of lists now substitute in the list's address, to work around DMARC damage, but the IETF is finally working on fixing that. The usual approach with anti-DMARC'ed lists is to put the original author in the Reply-To header, but of course if the author has an EAI address, you can't sent the message to non-EAI recipients. There is no way for the ASCII recipients to respond to the author, or even to tell who the author was. This situation is a mess and there is no good solution short of the entire world adopting EAI. R's, John
Hi, Yes, I know the problems. However, Michel Bernier has written code. It's on gitlab now, I've looked at it and AFAICT it should handle ICANN's lists acceptably. Arnt On 03/01/2025, 19:34, "John Levine" <john.levine@standcore.com <mailto:john.levine@standcore.com>> wrote: It appears that Arnt Gulbrandsen via UA-EAI <arnt.gulbrandsen@icann.org <mailto:arnt.gulbrandsen@icann.org>> said:
Hi,
3.10 still doesn't include EAI support. I think the only reason is that the changes involved are rather large. If I were the Mailman maintainers I'd plan an uninterrupted week to review and merge that, and that's a big chunk of time for a spare-time project like Mailman.
I think the reason is that there is no way to have a mailing list with both EAI and ASCII subscribers because there's no way to downgrade an EAI message and allow ASCII users to reply. We've been thinking about this for a decade and nothing has worked. I suppose you could add EAI-only lists where you could only subscribe from a system with SMTPUTF8 support but that doesn't seem very appealing. R's, John
participants (9)
-
Arnt Gulbrandsen -
Dessalegn Yehuala -
Don Hollander -
Dr Ajay Data -
John Levine -
John R Levine -
John R. Levine -
Mark Svancarek (CELA) -
Душан Стојичевић