Mailing list migration

[vjudeu]

Either way, I think the forum could be a good alternative. And maybe someone could create a bot, where you send an email, and the bot then writes the post.

I think GitHub is a better alternative. Because you can make comments, just by sending them via e-mail. So, this functionality is already there, when it comes to GitHub. Of course, it has some drawbacks, for example moderation, that should be applied before some message will be included, and not after that.

In general, when I look at all proposed solutions, I think GitHub is the most promising one, because development is already here, you have all Pull Requests, you have all source code, and all BIPs. Everything is there, except full mailing list support. But of course, it is not 100% bulletproof solution, but only "not that evil as some others".

was it viable?

Well, the closest thing that exists on Bitcointalk, is probably something like Private Messages. But still, you have to log in, so the protocol to use Private Messages is not the same as in e-mails. If you could just send a Private Message via bitcointalk, without logging in, just by filling all needed fields, and without establishing a connection, that could work.

And maybe someone could create a bot, where you send an email, and the bot then writes the post.

The bot should not write the post automatically. It should be accepted by some moderator first, and then published. This is how mailing lists work, and that functionality should be preserved. Because you don't want a Mallory, that would send some spam to the bot. And you also don't want a Mallory, that would know the whole list of participants.

[joker_josue]

And maybe someone could create a bot, where you send an email, and the bot then writes the post.

The bot should not write the post automatically. It should be accepted by some moderator first, and then published. This is how mailing lists work, and that functionality should be preserved. Because you don't want a Mallory, that would send some spam to the bot. And you also don't want a Mallory, that would know the whole list of participants.

Logically I made a minimalist description of what this bot did.
It would only automatically post a message if the sender was on the authorized list. The rest would have to be approved in advance.

Anyway, I think there are several factors that need to be evaluated and several possibilities.

[TryNinja]

Either way, I think the forum could be a good alternative. And maybe someone could create a bot, where you send an email, and the bot then writes the post.
@TryNinja, was it viable?

It should be possible if it's one bot creating every post (so I don't need to get someone else’s password). But I honestly don't see the devs using it.

[DaveF]

- They create a self-moderated thread.

This would not work. Spam control needs to be proactive rather than reactive, or else you end up filling everyone's email inbox with hundreds of daily spam posts before they are trashed.

Indeed, and I think that is partly why the mailing list has been able to be useful for so long.

For a while, the mailing list was greylisted - anyone could send to the list, but people who hadn't posted before would have their mails held in a moderation queue before they were sent. Apparently issues with mailman2 have made the mailing list now a moderated only list, although they still permit a lot of stuff and it's basically just manual greylisting. The key thing is the fact that emails from new people are held while emails from "approved users" go through pretty quickly. This cuts down on the spam.

However any system which requires pre-approving users before they can send anything will discourage many people from posting. With email, you don't know that your email might get moderated before you send it, but anything that requires approval before being able to even send will just result in fewer new people participating.


The fact that email is universal is extremely beneficial. New people do post to the mailing list, and sometimes they have good ideas. I strongly prefer a mailing list for these technical discussions since it has a low barrier of entry. Everyone has an email, and everyone knows how to use one. There's no need to make a new account with yet another password. No need to have yet another website that you have to login to in order to read or respond to posts. The barrier of entry is low, and that's what we want so that fresh ideas can come in. Of course we need proactive moderation so that it doesn't become all spam. Mailing lists are perfect for that.

Of course the last bit that makes everything hard is who is going to host it, where is it going to be hosted, and how will it be paid for. Running a mail server is not easy. It's extremely easy to get blacklisted and then the server is useless. The mailing list also has thousands of subscribers and apparently the existing self-hosted mailing list software just is not capable of dealing with it, or have significant other drawbacks. It's not just "a bit of modification", and just because there are developers involved doesn't mean that all software problems are solvable. mailman2 and mailman3 are complex projects with lots of issues, and the people working on Bitcoin aren't working on fixing those software since they're interested in working on Bitcoin. Even if the software were functional, who would host? Who would pay for servers? These are questions that are yet unanswered, and often have unsatisfactory solutions. Even now, it isn't the Linux Foundation that hosts the list. It's OSUOSL, a non-profit that helps open source projects. But they too have limited time and money.

Some points and just my view on a lot of them and not in order (sorry)

1) Running a wide open mail server is difficult. Running one for just a mailing list is somewhat simpler as if you are keeping the list clean and only allowing approved people to post will cut down on a lot of spam

2) The cost I really don't see as an issue. Although there is a cost, it can be done for under $1000 a year, which while it is not an insignificant sum it's not earth shattering.
2a) I can see a lot of places wanting to give a discount for getting bit a free advertising for supporting BTC

3) If you want to keep with mailman there might be more coding but there are other free and commercial options out there. I know a lot of people will balk from buying a piece of closed source commercial software to do this but time is money and if spending some money now saves time and effort then it's not wrong to do it.

4) Administration would be an issue, BUT I would think there are some people out there that would want to help. I could be 100% wrong on this, but with the number of people supporting BTC I can't see no one helping.

5) Although a royal pain in the ass you are not locked into 1 option. If you go with product A and discover it sucks....you then try B. I keep seeing the attitude of well this sucks but we are sticking with it from a small 2 person business to major 1000+ employee places and it just boggles my mind. Fine it didn't work....move on.


-Dave

[DooMAD]

Much like the mailing list itself, the discussion here about potential migration just kinda... stopped.  There seemed to be a lot of suggestions, but no clear sense of direction.  Do we at least have some sort of leading contender or front-runner yet?  I got the sense that it's a time-sensitive matter, so I would have thought there might be a greater sense of urgency about it.

[joker_josue]

Much like the mailing list itself, the discussion here about potential migration just kinda... stopped.  There seemed to be a lot of suggestions, but no clear sense of direction.  Do we at least have some sort of leading contender or front-runner yet?  I got the sense that it's a time-sensitive matter, so I would have thought there might be a greater sense of urgency about it.

Well, this turns out to be urgent for the holders of these mail listings.
Have they said anything about this matter? Without their feedback, it is also difficult to debate the ideas presented much further.

[garlonicon]

Much like the mailing list itself, the discussion here about potential migration just kinda... stopped.

It is not surprising. If you compare the discussion today, with the discussion in the past, then you note it is much different, than it was. And you also note, that people moved to other places. For example, there are IRC channels, there are GitHub discussions, and also note that the winter holidays are coming.

Another thing is that if you want to see new messages, then someone should accept them. Which means, even if there is some traffic, waiting to be published, you can see it a day or two later, because things are moderated.

Do we at least have some sort of leading contender or front-runner yet?

Not really, and there is another possible outcome: the list could just silently die, without any replacement. This is also possible, and this will not be the end of the world, because participants know their e-mails, and can exchange some messages directly, without any list, and another list can be formed from scratch, when there will be a need for that. Because if there is little or no demand, then it is normal to have little or no supply.

I got the sense that it's a time-sensitive matter

It is not. It is e-mail based, and also moderated list, where it is normal to post something, and have your message rejected, or published two days later. It is not like this forum, where you write a post, and where it is immediately fetched by some bots, and where all statistics are recalculated based on that, and where people can see the first version, even if you edit or remove it.

Have they said anything about this matter?

Not really. I don't think there is anything under the hood, that you cannot see publicly. I think the main idea is to let it die, and create something new, only if there will be any need to do so. And I think it is a good approach, because there are many ways to communicate, and because if someone would want to talk to any participant on the list, the e-mail address is usually public, so if you want to know, what John Doe thinks about that, then you can just ask him directly.

[DooMAD]

Not really, and there is another possible outcome: the list could just silently die, without any replacement. This is also possible, and this will not be the end of the world, because participants know their e-mails, and can exchange some messages directly, without any list, and another list can be formed from scratch, when there will be a need for that. Because if there is little or no demand, then it is normal to have little or no supply.

As someone who isn't a direct participant, I like to periodically read the archives.  If only to get a sense of what developers are thinking and what their attitudes are.  The level of demand from casual observers like myself may be more difficult to ascertain, compared to the demand from those actively posting messages.  I feel the transparency of it is quite important, so would very much like to see that aspect preserved.  Plus the mailing list tends to be more conversational than the equivalent discussions on GitHub, focusing on ideas rather than implementation, which makes it more accessible to those of us who aren't coders.  It would be a shame to see it go.