Allowing users to join sub-groups, even if they are not a member of the parent group

Samson mentioned the fact that he has a sub-group and wants to invite some users to join the group, even though they are not members of the parent group.

At the moment we have decided to limit sub-group membership to users who are already members of the parent group. The reason we decided to implement sub-groups this way is because once we decide to NOT do this (and allow anyone to join subgroups), it's not really an action we can undo.

So the question is... how do people feel about the idea of opening up sub-group membership to people who are not members of the parent group?

Trello currently allows for this (boards that are owned by organizations can have members that do not belong to the organization).

Yammer, on the otherhand, does not allow you to join groups belonging to networks that you are not a part of. However, I think Yammer deals with this by providing "External Networks" for companies. So non-members can't join a company's internal sub-groups, but they can join one of the company's "external networks".

External networks seems like a bit of overkill for Loomio. I think we could probably handle everything with sub-groups the way Trello does (very explicit/easy-to-understand permission controls on a per-subgroup basis).

How does everyone feel about this? Can anyone think of any negative side-effects from allowing users to join sub-groups even if they are not a member of the parent group?

I'm in favor of having members of sub groups that are not part of the parent group - I could work around it by having a parent group which is a shell (no discussions etc.) and all the action happens in various sub groups so it's not a huge deal either way for me.

I like the idea of subgroups being able to engage people with their technical requirement even if they do not want to be part of main group - not everyobe will thank god/goddess! So long as the majority of subgroup memebrs are memebrs of the wider group I am happy

It's sounding like people are in favor. Can anyone think of any potential negative consequences (red hat!) from going ahead with this decision?

black hat :P Red is emotions and gut response

I find on trello it's a hassle to add people to a sub group and parent group at the same time so having a checkboxon a subgroup invite box for invite to parent as well would be good.

It would also make it hard to see a list of all members in all sub groups and main groups if you ever wanted to broadcast something to all of them - an admin report in the parent group which showed all members and what group they were in would fix that.

I'm not so good at black hat sometimes...

@Jon thanks for bringing this up as a discussion. From your introduction of this topic and on Yammer it should like you all have given it a lot of thought. I can see it adding a little bit more complexity in you development but from a user point of view I think it is more flexibility which is nice. I think as long as you can easily identify if a user does not exist in the parent group then it should be fine.

Digging into implementation here but, these are the scenarios I can see right away

---

Group A: Jon, Joshua, Vivien

Group A-A: Jon, Joshua [remove]
Group A-B: Jon, Joshua, Vivien, Sam [add]

Group A-C: Jon, Joshua, Sam [add and remove]

So Group A is the parent group and the others our sub-groups that are ether a subset or a superset of the parent group. I think what Joshua is saying is that at a certain point you might want Sam to the parent group in an easy way.

I think that can just be handled with something like

User(:id)
UserGroup(:user_id, :group_id)
Group(:id, :parent_group_id = nil)

Gah, too much implementation! Mainly, I think this will be more of a UI/UX problem than a programming problem. And I know in my case it would be useful so I'm sure it would be for other subgroups as well.

Okay, so what I'm hearing is that no one can think of any negative functional impacts from making this change. Just a few additional user-experience considerations.

From a business perspective, if we set a limit on how many people can join your group (say if you've got a subscription for a small Loomio group account), you could potentially create a sub-group and then just invite a whole bunch of members who aren't part of your main group (bypassing your parent group limit). I guess the solution to limit the sub-group size as well.

I could also see situations where you might want your sub-groups to only allow people to join if they are members of your parent group. If this became a requested feature we could add it as an additional permission option to sub-groups.

Can anyone else think of any other potentially negative impacts from making this switch? Making this change represents quite a large structural shift in terms of the way groups are organized (and groups are Loomio's central organizing unit!).

I can't think of any show-stoppers, but it'd be good to know that we've really exhausted the list of potential outcomes, since as I mentioned before, once we make this change it will be hard to undo.

(I want to push a button like: [I'm following this conversation with interest, but I haven't got anything meaningful to contribute at the moment. But don't think you're alone in cyberspace, o contributors!] )

[I want to like the previous comment with extra emphasis than the 'like' button affords me.]

Sounds like nothing's stopping the Jon Lemmon Show, Jon Lemmon! Keen to go ahead and put this into proposal form?

I think we need to consolidate the Loomio focused work teams and enhance those with specialists and resource people in each work area. As long as there are more Loomio members in each subgroup that'specialists' I think having people be members of the group is important to get the expertise we need.