Re: Recognized PostgreSQL External Communications Channels - Mailing list pgsql-www

Hi

FWIW, it would be much easier to comment if you included the text inline in the email. With that said, some thoughts/questions below...

I have this in a Google Doc, if you like I can send you the link.

Intentionally I did not include the link here, for a variety of reasons.

On Tue, 15 Jul 2025 at 22:03, Andreas 'ads' Scherbaum <ads@pgug.de> wrote:

Hello,

one of the results of the PGConf.Dev 2025 Community Summit is
"Communication Channels" (1) and how to officially recognize such channels.

Attached is a draft for a

"Recognized PostgreSQL External Communications Channels"

Policy. The meeting in Montreal discussed listening recognized channels
on the postgresql.org website.

1) At the moment, we have taken the position (rightly or wrongly) that any communication channel that includes multiple PostgreSQL people is automatically subject to the CoC. There have been CoC incidents centered around private communications and at least one Telegram channel which was setup by a community member for general chatter on any subject. Having a list of "recognised" channels will - whether we like it or not - change that. It will implicitly set a boundary on what channels the CoC applies to. That could effectively exclude the Telegram channel, or private communications which one could argue are effectively a mailing list created through use of a persistent set of addressees. 

I'm not suggesting that's a bad thing - in fact, I think we do need some boundaries to prevent overreach and ensure that people know where the CoC applies and where it does not (part of the issue in the Telegram case was indeed whether or not the CoC should apply). For example, I have often made the point that my 50th birthday party included multiple PostgreSQL community members, as well as people who have nothing to do with PostgreSQL at all. Arguments have been made that the CoC would apply to that gathering, which might have meant that the community members in attendance would be prevented from making a joke that would have been considered perfectly acceptable in a pub in the UK but that the other attendees would have been able to make, either because they didn't care about or know of the CoC, or because it was recognised that it wouldn't apply to them as they weren't part of the community.

My point is that we need to recognise that this proposed policy and the resulting list may have far wider reaching consequences that initially envisaged, and to be mindful of that.

Correct.

The discussion started about where to find community (approved/recognized) channels and platforms.
More specific, what is necessary in order to list such platforms on the project website.
This is different from automatically including the CoC on all channels where community members are present.

2) Some platforms do not allow multiple owners/administrators. Does that mean they cannot be recognised?

Good question. Got an example?
This is a draft and can cover such platforms, I'd like to understand how this looks like.

3) If an owner/administrator steps down, does the channel automatically become un-recognised? Perhaps a grace period is required?

Yes, makes sense.

4) I find the way the doc talks about owner/administrators and then moderation a little confusing, to the point that I had to read it a couple of times until I realised it wasn't talking about 3 different groups of people. Perhaps the terms owner and moderators would be better? That would likely solve my point 2 above in some cases as well, where platforms allow one owner but multiple moderators.

It might be three different groups,

the owner of a service,
an administrator with global permissions,
a moderator who can only moderate a specific channel.

It does not apply to each and every service though, which indeed makes it complicated.
Slack can have all three, Telegram can not. As example.

5) I think the terms of service section needs some thought. As written, if a service explicitly allows (for example) hate speech, then that means we have to allow it in the PostgreSQL channel too. I think this section needs to state instead that the most restrictive terms must apply. 

Was thinking about this for a long time, but from the other end: can't really force users to violate the terms of service, if that collides with the CoC.
You make a good point though.

Open for suggestions to find a middle ground.

6) Although there is the universal get-out clause at the top allowing the core team to not recognise at will (kudos for keeping the proper spelling there :-) ), I wonder if we should also have an explicit clause stating that we will not recognise channels on platforms that clearly are not appropriate for the project, for example, a platform primarily known for extreme political discussion.

That makes sense, will include it.


Thanks for the feedback!

-- 				Andreas 'ads' Scherbaum
German PostgreSQL User Group
European PostgreSQL User Group - Board of Directors
Volunteer Regional Contact, Germany - PostgreSQL Project