Proposed guidelines for discussion of proposed Changes

Project Discussion
22

+1 on this idea when it looks like a particular topic is getting too hot. I think it encourages folks to think more about their next post and potentially gets them to read more of the thread if they’re not up to date.

I think breakout topics can be helpful. Maybe as well as updating the topic level post with breakout topics you also reply to the thread with the latest list of topics each time one is added. That way you have something at the top and bottom to indicate the structure of the conversation.

For example, you start a thread and it has two breakout topics from the start. Those just live at the top. Then conversation happens and there’s need for a third breakout topic. You may the new thread, add it to the top post, and make a new post saying “here is the updated list of breakout topics.” Now someone new to the convo knows about the breakouts and someone coming back to the discussion also knows about the breakouts. Before the folks who already started participating wouldn’t have caught the updates to list.

One thing I’m not so sure of us moderators pulling posts around into different topics. I know the idea is to reduce the number of posts in the main thread when it gets too long, but then you lose the context in which the post was made. Also, if the poster is not aware of this as a norm, they may feel slighted or like moderators are overstepping a boundary in shuffling their opinions around. It may be better to leave the main threads alone but advertise that the breakout threads exist for more focused discussion. Let people go to those threads on their own rather than pulling them into it.

One more suggestion I would make is to consider having some kind of timeout period? So maybe when a discussion is moving too quickly we have like 2 days of discussion, then 1 day of locking the thread so everyone can catch up, then 2 more days of discussion. If we can help pace the conversation for the community, I think that could reduce the number of posts overall and let people catch up instead of feeling like they’re reading a never ending thread. If there’s an option where a thread can be locked but people can still leave reactions, I think that’s even better. And the last post in the thread should be the one that announces the recess, joined by a similar notice in the top post.

As much as we want everyone to have read every single comment, I don’t think that’s a reasonable expectation outside of the proposal owner and FESCo. Even in a mailing list nest conversations can be hard to follow. I personally end up glossing over those just as much as I had to gloss over the telemetry proposal. Happens on Reddit it. TL;DR is just a fact of life. Not to say that we shouldn’t try to keep discussions from unfurling into something unproductive, but sometimes it is what is is.

23

I missed this one earlier. I think this is a good compromise. Hopefully time will remove the emotion involved and much of the back and forth.

However, it will conflict with your idea of multiple posts instead of combined posts.

3 Likes
24

I wonder if the making changes based on the premise that the situation was out of control or in need of improvement is the the best thing so soon. An open and inclusive process and related discussion is always going to be messy. Not unlike the contract that exists in so many other situations: with freedom comes responsibility. It usually mostly works but never works perfectly.

Easy for me to say as someone with zero moderation responsibilities, but it appears that the community remains whole and functional. Me thinks that’s a good thing!

1 Like
25

My hope with putting this expectation clearly in the initial guidelines is that it won’t be surprised, and will be more inclined to follow to the right place naturally. (Especially easy if we close the main thread!) And, most of the moving around was either in initially establishing the breakouts or shortly after. I think that if we create them much more early on (including pausing to organize things — a suggestion @dalto should get credit for, by the way), it’ll be far less confusing and maybe even become natural. I’d like to test that theory. :classic_smiley:

I feel that the result of this would have been a horrific, messy “main” discussion, which FESCo and the Change Owner (and, you know, me) would have still had to try to digest. But… maybe that’s just how it has to be.

26

Yeah, I’ve moved “Clarifications” up in the list to reflect this.

27

It seems — at least from initial feedback — my thoughts on moderation, and on repetition and etc., make sense to most people. Hopefully, to FESCo as well. But the breakout flow is more controversial. I’d like to be intentional about what we do next, and I’d like @fesco to please choose our approach for the next round.[1]

Because I Can, I’m going to make this a Discourse poll limited to FESCo members. I’ll set it so the person who voted is visible to all. FESCo folks, if for some reason you don’t want to use this, you can just comment instead. I’m going to follow this with the same poll, open to all. This is multiple choice — select all that you’re okay with, and we’ll go with the top. (Or else just tell me via normal FESCo voting procedures, of course.)

FESCo poll
  • Pre-create breakouts as I have proposed above (including a few topics we can anticipate, like “dialog design”)
  • As I have proposed above, but combine Support / Oppose in to one[2]
  • Create only a few topical initial breakouts (again e.g. “dialog design”)
  • Start with a few topical breakouts, but pause after ~50 posts for moderators to create new ones and organize things
  • Do not create any breakouts explicitly, and see how things go — without Slow Mode. Moderators won’t move posts [3]
  • Do not create any breakouts explicitly, and see how things go — with Slow Mode. Moderators won’t move posts [3:1]
  • Something else (please tell me!)
0 voters
“Everyone” poll — take this as what do you prefer?, not just what would you accept?
  • Pre-create breakouts as I have proposed above (including a few topics we can anticipate, like “dialog design”)
  • As I have proposed above, but combine Support / Oppose in to one[2:1]
  • Create only a few topical initial breakouts (again e.g. “dialog design”)
  • Start with a few topical breakouts, but pause after ~50 posts for moderators to create new ones and organize things
  • Do not create any breakouts explicitly, and see how things go — without Slow Mode. Moderators won’t move posts [3:2]
  • Do not create any breakouts explicitly, and see how things go — with Slow Mode. Moderators won’t move posts [3:3]
  • Something else (please tell me!)
0 voters

  1. Or, I know how things work, so “tell me something entirely different, and we’ll try that!” is on the table too :classic_smiley: ↩︎

  2. and also probably combine “Improvements” and “Alternates” into one, so there are fewer breakouts overall ↩︎ ↩︎

  3. except perhaps those which are completely off topic. ↩︎ ↩︎ ↩︎ ↩︎

1 Like
28

Speaking of polls, we have some options which we might use, and it’s worth discussion if we should.

As demonstrated above, we can create polls limited to particular groups. I don’t think we should generally have a FESCo one like this for the topic itself, but we could create sets of polls like this where one is open to everyone and another is restricted to, say, the “fedora-contributor” group — everyone who has agreed to the FPCA and who is in at least one other group.

We could also restrict to “Trust Level 1” — I’ve made it so everyone in @fedora-contributor is automatically elevated to that level[1], so it includes that plus people who have put in a minimum level of time participating in the forum.[2] This would filter out drive-by signups just to vote — for better or worse. (I think worse, actually, but I wanted to mention the possibility.)

Also, note that trust level[3] is included in the admin-downloadable poll results. I said that I wouldn’t use that for anything on this last poll, so I won’t — but we could say something different in the future. (That is, we could report on the results with the tl0 responses separately, if we so desired…)

  1. to avoid the problem of “old hand at Fedora, trying Discourse, hit by anti-spammer / troll protections, becomes frustrated and annoyed”. ↩︎

  2. the exact parameters aren’t public, because that would just be a playbook for spammers to follow ↩︎

  3. I presume at time of voting, but I don’t actually know. ↩︎

29

Speaking of polls, I think polls are very important!
The poll should be created as soon as possible.
IMO there should be a “main” poll that most accurately represent the current proposal. The first UNOFFICIAL poll about OPT-OUT metrics proposal was a great example of this.

Such poll should be pinned and linked to the proposal thread.

There could be other “reference polls” too.

1 Like
30

I’m in support of polls too, but I do not want to get people into the mindset that they’re a form of referendum or vote. People are elected to FESCo for their thoughtfulness and expertise, and I don’t want to reduce that leadership role to “rubber stamp a popular vote” — or subject FESCo to pressure in that direction. We don’t have a defined voting constituency, and there are plenty of other reasons it’s not scientific.

I don’t want to get back into discussing whether I think that initial poll was ideal (I am certain it was well-intentioned!), but I do think it highlights the difficulty of getting the wording right in polls. That’s the primary reason I suggest we discuss before creating polls.

Also using this as an example, even when a proposal indicates inflexibility in some area, it’s useful to gauge community interest in alternatives. The Change Owner may be open to something they hadn’t considered, or may be convinced by arguments after all. Or, someone else may step up to back a change based on an alternative, if it’s popular enough.

Also, I don’t want to mandate the creation of a poll. The vast majority of change proposals are uncontroversial, so this would just be extra work for no clear benefit.

2 Likes
31

And “reference polls” could be used exactly for this scenario.

1 Like
32

I think the breakout topics were terrible. They fractured the discussion and led to lots of repetition, making it harder to follow. Creating them in advanced seems like a further step back.

1 Like
33

That’s indeed a great idea and useful filter to mitigate “rejection sign-ups” of users who don’t possess sufficient understanding of how Fedora works and what it is. Illustrating the reward (voting capability) this way, even of small contributions (such as that for TL1), might also encourage users to get a little deeper into Fedora (and thus gather understanding of what “it” is / how “it” works). I think even little interactions on discourse can already make a difference in this respect.

34

I also think the break outs fostered some repetition, and given the interrelations of the topics, it was not possible to keep these topics separated so that they sometimes developed both redundant and interdependent. This again fostered repetition and made it harder to follow.

Yet, I am not a fan of splitting only “completely off topic”. Once something is “sufficiently off-topic” to not create interrelated/interdependent discussions (this is also a matter of perception of course), I suggest to split it. In short, I would do it as we do it on other topics, but even more radical enforce splitting when something develops off-topic. However, break outs are not off-topic but related, so this is what I would avoid in future.

Nevertheless, we have to keep in mind that any proposal creates a context on itself, and proposals can have different backgrounds/outreaches and thus create different contexts in which we have to moderate. I could imagine that there are types of proposals where break outs make sense (some proposals might incorporate dividable sub-proposals). Therefore, I am not sure if it makes sense to create a “standardized” decision for all future proposals. (Unfortunately, predicting what makes most sense in a given case will not become easier I guess).

I tend to remain an opponent of break outs in “average” cases - however, it might make a difference to invest more time in deriving appropriate break outs / sub-topics in advance: I think it was not foreseen what we would be dealing with, and thus, there was no time to respond but only to improvise. Planning the categorization/separation of topics more precisely in advance might also make a difference.

1 Like
35

I think this is a good way to make a barrier on the user-base side of fedora. And as already said, to filter.

In the other hand I would also welcome it see it to be applied also for developers who come from other projects alias from the mailing-list. Discourse is definitely a new medium and also the audience is much more mixed. I found it very courageous from @catanzaro jumping in to the cold water with his first topic to bring it up here. Not in a negative way, he showed with his topic that we do have limits with discourse and that we need to adjust, and find new ways to work together.

The point I try to make is, that when a user hast to be TL3 for example he can change tags and set additions to the title and even move a topic into an other category/section. This means an exchange with a bigger audience is needed.

The TL3 Limitation would also give the possibility to talk just with the persons who are behind discourse doing all the administrative work, breakout topics, dealing with flags etc. This smaller group probably would have given valuable information while make a poll just within this group.

While working in first level support, I was an interface/firewall/translators for second and third level supporters like developers and engineers. They do definitely speak an other language. Sometimes we also just had to keep them the back free while they had to implement new systems/environments.

36

As everyone has probably seen by now, we are experimenting with using Fedora Discussion for the community feedback and review portion of our Change Process. This makes the process more open and transparent to all. That’s great — we want more visibility and participation. But, as we’ve just seen, conversation can quickly get unmanageable.

With something big and controversial, this could happen no matter where or how we have the conversation. We need clear guidelines so that discussion is productive, in line with our community values, and easy to follow and participate in. Some of these guidelines will be similar to those we already have for our mailing lists, and others need to be new for the new format.

I propose the following. I’d like to adopt this as soon as possible, for the anticipated upcoming revision to the recent controversial proposal. Of course, we should continue to learn and adjust as we go forward.

I think it may be important to not over engineer things.
This is a very important topic to many people

The first part of the readme looks great IMHO.

Breakout Topics

So, I read things here via the email interface.

It’s worth noting that the ‘breakout’ threads didn’t affect email at
all. I see everything as one big long 1000+ post thread. :wink:
I don’t know if thats a bug or expected.

IMHO, breakout topics should be for things that are substantially
different than the proposed change. Like if someone starts discussing
the change process itself or has some other idea sparked by this change.
Trying to break apart ‘kinds’ of discussions for the change seems doomed
to me, people will reply to particular things and likely not stay on one
topic. Also, general feedback then becomes harder to see where it goes.

It might be worthwhile to ask our discourse upstream what they would
suggest or look at other larger communities how they handle things.

We want everyone to be heard, but many posts repeating the same thing actually makes that harder. If you have something new to say, please say it. If, instead, you find someone has already covered what you’d like to write, consider simply giving that post a :heart: instead of reiterating. You can even do this by email, by replying with the heart emoji or just “+1”.

Is there any way to tally / show a overview of those?
Like “this post got 12 :heart:’s?”

I’m not too keen on polls. I guess that could be another good way to get
more feedback, but writing poll questions is… not at all simple. It’s
super easy to skew things by wording poorly or just not including some
popular options. I’d personally just leave polls to the change
submitter. If they want to gather info about something they could do so.
Having moderators write up polls seems… dicey.

My 0.000002 cents.

37

If you address my comment, you probably misunderstood it.
Just to clarify, If a requester pretends to do a request he has the possibility to do a poll to find out what the narrower group of TL3/TL4/Moderators&Admins thinks about, if this will rise a chaos or if this would be smooth to handle etc.

But for this, he needs to gain first TL3 in discourse what he gets automatically if he just socializes, helps other professionals and or gives his personal meaning based on his professional knowledge.

So we also see that he commits to fedora because he has a relation to the community and not just with his employee who pays him to do so (I just added this point because it was highlighted that Fedora is not RH ).

38

I am not. I was answering Matthew’s orig post that started this thread.
:wink:

1 Like
39

For what it’s worth, I think at least moving opt-out topic helped. That was most contentious, and had

  • most of the cases where people came to be counted — that is, to write their personal stance on that particular topic, even without adding any specific new points or information — and
  • most of the back-and-forth arguing.
40

That’s a good idea. I will bring it up in the “community” section there.

1 Like
41

That’s shown below each post on the website. You can also pull it from (for example)

https://discussion.fedoraproject.org/t/proposed-guidelines-for-discussion-of-proposed-changes/85904/36.json (look for reactions).