Talk:Special Interest Groups

About this board

Quiddity (WMF) (talkcontribs)
Daniel Kinzler (WMDE) (talkcontribs)

Kevin mentioned a "Technology Sunsetting working group".

And for completeness, don't forget to mention the front end standardization group, even if it's already mentioned as an example in the prose.

Tgr (WMF) (talkcontribs)
CPettet (WMF) (talkcontribs)

One of the best organized and long tenured groups I can think of is Scrum of scrums. I don't know that they have any interest in the branding or subtle changes it may take to be a "SIG" but it's regular, with public minutes, and a well defined charter and meeting framework.

Whether a group is a SIG or not is up to the group I think. In time we'll know if the standard has legs because everyone will find it useful.

KSmith (WMF) (talkcontribs)

I don't view SoS as a SIG, because a) attendees rotate, and b) they are pretty much told to go, as opposed to attending out of professional interest.

For me, the Team Practices Guild and Product Manager Salon could fall into the SIG category.

My views are still forming on all of this. I have my own preconceptions about what a SIG might be, which are different from others (and from this proposal). I do think the main focus should be: Are there groups for whom these guidelines are helpful? Are there tweaks (over time) that would make them more helpful, or to more teams?

CPettet (WMF) (talkcontribs)

Makes sense @KSmith (WMF), I'm glad you are engaged in this conversation. Your point of view is useful.

Reply to "Possible/Proposed SIGs"

Do SIGs have authority?

Daniel Kinzler (WMDE) (talkcontribs)

Is the expectation that SIGs would be granted some authority (and by whom)? Is the expectation that SIGs decide and enforce issues within their scope? Would that be the rule, or the exception?

BDavis (WMF) (talkcontribs)

This is going to depend on the SIG I think. For example, the Wiki Replicas Special Interest Group will have authority in that it will inform the work of the Cloud Services team.

We lack central governance for the Wikimedia FLOSS projects, so there is no unified central authority to appeal to to designate powers to SIGs broadly. Authority by "doing" is certainly possible.

CPettet (WMF) (talkcontribs)

I don't think the question of SIG's having authority is directly answerable. SIGs are not meant to be a model for command and control or lack of it. They are a format for collaboration. Some of the important parts of the SIG model at the moment: agreement to publish public minutes so that ongoing and future conversations can benefit from historical openness, a place to document public groups of specific focus in case there is a shared area of concern, a designated point of contact to reach the self-defined group, and regular and expected meetings for internal and external collaborators.


  • SIGs could be thought of as a mixture of a "Committee" and "Task Force".

May be confusing I think.

SIGs are not an embodiment of authority other than natural authority. A SIG with Bryan and I that is focused on something we both are de facto driving would have authority. A SIG over something we are not in turn would not. We would struggle to express our reasonable stakeholder interest I suspect. I think that leans into "SIGs with authority over their scope of interest" are the rule, in that otherwise it gets hazy on direction. Informational and user-group oriented SIGs are useful and especially for social issues make sense to me. In this case, the idea of stakeholder is not at all tied to authority in the common way, but the format and surfacing of issues is worthwhile and I imagine the group would be self-selecting for derived value.

The model is held to different standards in various organizations, but in essence it's more of a guide on how to run a court proceeding than it is an edict on the powers and authority of a judicial branch of government. That's a clumsy inch deep analogy but it's the one I can think of right now :)

Reply to "Do SIGs have authority?"
Tgr (WMF) (talkcontribs)

Nitpick: it makes more sense to define inactivity based on time interval than number of meetings. For a group that meets every three months 6 meetings feels like an overly long time to recognize it as defunct.

CPettet (WMF) (talkcontribs)

It was intended to be both. The wording now could use some help.

"If 6 or more meetings in a row are missed or do not have a quorum of a majority of members present, the group will be considered inactive."

Intention: If meetings do not occur for 6 months or 6 meetings whichever is shortest a SIG will be considered inactive.


- It is better to be generous and enforce marking groups as inactive publicly rather than conservative and lack consistency

- There is no cost for reactivation but activity and lack of activity are both public signals worth consideration

- We can revise over time. This is a first pass smell test standard.

Reply to "Inactivity"
Daniel Kinzler (WMDE) (talkcontribs)

During the last TechCom meeting, we briefly discussed this document, and wondered whether TechCom counts as a SIG under the framework provided here. I think it does, though it's worth pointing out that TechCom's relationship with the CTO makes it somewhat special.

Was the intention to include TechCom in this framework? Would TechCom count as a good example of a SIG? Or is it rather an odd, border-line case in your opinion?

BDavis (WMF) (talkcontribs)

TechCom could certainly choose to identify and operate as a SIG. There was no intent in the description of this framework to force any existing group to adopt the process. There is also no intent to force any future group to adopt the framework. Instead the intent to to provide an option for people looking to form a group or reform the processes of an existing group an example of a framework that could be followed and hopefully to iteratively refine that framework as it is used by more groups.

Reply to "Is TechCom a SIG?"

Questions on membership

KSmith (WMF) (talkcontribs)

Regarding "Membership can include anyone who is a demonstrated stakeholder, and the groups should be between 5-15 people." I have 2 questions:

Does the first half mean that any demonstrated stakeholder would automatically be able to be a member?

Is the second half intended to limit the size, or is it merely a recommendation?

Quiddity (WMF) (talkcontribs)

For the first half, that is probably up to each group to define more specifically for themselves, based on their expected scope of activities and meeting schedules. E.g. the support of a quorum of existing members could be required, or it could be self-selection based. We expect that some might be higher-intensity very small groups, and some might be more casual and larger groups.

For the second half, that is intended to limit the size, and was based on the external research (in particular "To allow rapid progress, Working Groups are intended to be small (typically fewer than 15 people) and composed of experts in the area defined by the charter."), plus standard advice about group size (the best short summary I can quickly find is "Thus, the best number of people for one project might be five, while the best size for another group might be 12. As a general rule, groups that have more than 18 to 22 people or less than five become more challenging to manage.")

KSmith (WMF) (talkcontribs)

Thanks Quiddity. That all makes sense. It might help to clarify in the text the degree to which those are permissive, prescriptive, recommendery, etc.

Greg (WMF) (talkcontribs)

I want to make this edit but I'm not comfortable being the one changing it from "anyone who is a demonstrated stakeholder" to "experts in the area defined by the scope". I *like* the W3C's definition and want that to be the case, with the obvious statement that the SIG can't be overly (beyond some basics regarding anonymized meeting notes or so) private/secret in it's working.

CPettet (WMF) (talkcontribs)

I can appreciate the inclusiveness of the W3C's definition. I respectfully disagree that we should adopt it outright. I don't think expertise is the benchmark for productive involvement. Stakeholder is at once a mildly defensive term to make it clear to all involved what the others interest(s) and perspective is regarding the topic, and a way to communicate an understanding of shared fate. Being knowledgeable/expert alone for me does not pass that test. I'm very hesitant to agree this is the useful measure for issues I have seen interested in this following this model.

My proposal is to keep stakeholder and add a glossary and outlining generally the embodiment of stakeholderhood. This will allow for common metrics across the wildly different technical and social special interests I hope.

Greg (WMF) (talkcontribs)

The newly added definition for stakeholder alleviates my concerns, I believe. Enough that I'm willing to propose a current group I'm leading to be a SIG. My concern mostly had to do with "where's the line drawn" and "sometimes everyone could argue they are a stakeholder", but this definition strikes the right balance.

KSmith (WMF) (talkcontribs)

A specific group I'm working with strictly controls membership. It sounds like that would not disqualify them from being a SIG (in contrast to my initial interpretation).

However, another group I'm involved with will probably be uber-inclusive, allowing membership by anyone interested (stakeholder or not), and without a specific upper limit. It sounds like they would be disqualified from being a SIG (which feels wrong to me).

BDavis (WMF) (talkcontribs)

Not all organized groups fit the SIG model certainly. This is not a silver bullet. The model that is being proposed attempts to optimize for constructing groups focused on a particular task. The more narrow the scope of that task the better. Group size limits are intended to avoid deadlock and decision avoidance. Groups may be a better org model fit for your second example group.

Reply to "Questions on membership"

Naming things and Definitions

Quiddity (WMF) (talkcontribs)

We researched external groups who used these models:

  • "Special Interest Group" often means: "discussion-focused group, no member limit or restrictions, should have annual meetings or more, permanent existence".[1][2][3][4]
  • "Working group" are typically defined as some combination of: "deliverable/task-focused group, should be 15 or less people who are key stakeholders, should have monthly meetings or more, should be disbanded once task complete".[5][6][7][8]

However, these definitions are not consistently or rigorously applied, especially for smaller groups, e.g. the Kubernetes and OpenStack SIGs are a combination. There are also many conflicting/ambiguous usages of "Working Groups".

We've decided to not over-complicate/over-optimize things initially, and just start with SIGs as task-oriented small groups.


  1. "SIGs provide an open public forum to discuss topics of interest" -
  2. "The primary goal of an Interest Group is to bring together people who wish to evaluate potential Web technologies and policies. An Interest Group is a forum for the exchange of ideas." -
  3. "A Special Interest Group (SIG) will provide you with additional ways and means to interact with other AEA members with similar backgrounds and further promote the professional status of Enterprise Architects in your specific industry or professional segment. They: (1) provide a forum for individuals and organizations that share a common industry, interest, or focus. (2) share specific project management ideas, challenges, and concerns, (3) promote activities that include: networking, establishing standards and practices within a focus area, developing technical papers" -
  4. wikipedia:en:Special Interest Group
  5. "Working Groups (WGs) may be formed around a particular work item and exist only as long as it takes to complete the task(s)." -
  6. "Working Groups typically produce deliverables (e.g., Recommendation Track technical reports, software, test suites, and reviews of the deliverables of other groups)." -
  7. "Workgroups are the means by which AEA members collaborate to achieve some outcome." -
  8. wikipedia:en:Working group
Reply to "Naming things and Definitions"
There are no older topics