This proposal is intended to provoke productive discussion, surrounding our current governance structure, by highlighting some of the deficiencies that currently exist. While not exhaustive, it attempts to explain why the current governance structure is insufficient for the success and growth of the community, by comparing and contrasting our existing governance model with that of other organisations at a high level. It also suggests how our governance structure might be changed to address those deficiencies.
It is the author's hope that all recipients of this proposal will take the time to reflect on and carefully consider the points made here before responding. This proposal is primarily directed at the OGB, as representatives of our current governing structure. However, all recipients are encouraged to respond. The inspiration for this proposal is a direct result of recent events which revealed that governance of the community is at the heart of issues facing the community today.
The OpenSolaris community has existed as a self-governing entity since Friday February 10th, 2006
[1]. Since that time, individual parts of the community (and thus, the community as a whole) are continuing to make progress in many areas, including: technical, communication, and growth
[2]. The community has grown slowly, but surely, into something that we can continue to be proud of. The Advocacy (User Group), Desktop, DTrace, and ZFS community groups are just a few examples of that growth and progress.
However, the majority of this progress is a result of Sun's indirect leadership
[3], involvement, and the contributions of many individuals within the community. Many of those individuals are paid by Sun to work on Solaris, OpenSolaris, and related community projects. It is important to note the distinction of “paid by”; as many individuals are not employees of Sun (contractors) or were not employed by Sun at the beginning but currently are. By observation, it is apparent that none of this progress would have been possible without Sun's initiative to provide the source code that served as the nucleus around which the OpenSolaris project formed, and without their ongoing, significant financial support (which the author estimates to be in the range of millions of dollars).
Clearly, governance is one of the most important aspects of the community. However, governance alone is not sufficient to achieve sustained growth and success in a completely self-governing body, such as the one we currently have. The leadership hierarchy must be clear, and seen as inspirational
[4], creative, shrewd, and fair.
Upon reflection, it should become apparent that leadership and guidance is a necessary part of governance. To help us better understand our current governance model, it is helpful to compare and contrast our own governance model with that of others. Narrowing our focus, from the many governance models widely known, results in several which we will briefly examine. Commercially related projects include: Mac OS X
[5], PostgreSQL
[6], MySQL
[7], and Ubuntu
[8] (created and supported by Canonical
[9]). Other projects are those such as Fedora
[10], which are essentially alpha or beta representations of commercial products
[11]. Finally, we have Apache
[12] and OpenBSD
[13]; which are organised around completely open source
[14] products.
All these projects or products share several common characteristics. However, some characteristics are common and clearly visible: sustained growth and success. Each project or product has a parent entity that continues to build a community providing sustained growth and success, whether they are primarily proprietary in nature [5], have taken a hybrid approach between open source and proprietary add-ons [7], or have the primary focus of the project completely as open source [6, 8, 10, 12, 13]. They may also be experimenting with pay-for-contribution models.
In each case, clear leadership within well-defined areas of expertise is evident. There is a direct correlation between the quality of the leadership and the sustained growth and success of each project or product. The results are evident in a successfully delivered and widely-adopted end-product within their respective target markets.
For a moment then, let us consider the leadership that is integral to these projects and products. In Apple's
[15] case, few would dispute that Steve Jobs is clearly the primary source of leadership, and has been directly responsible for their current success
[16]. Likewise, the OpenBSD project is well known for its leader, Theo De Raadt, who, while sometimes outspoken
[17], has relentlessly driven the project towards an admirable level of fervour and success
[18] as defined by its stated goals
[19].
Investigating further, we see that each one of these projects has structured its leadership or governance differently. The Fedora project chose a mix of appointed and elected members with RedHat given final veto power (with the intention that it is to be used infrequently)
[20]. The OpenBSD project seems to have chosen its leader as a result of a meritocratic
[21] view (Theo is responsible for the project's existence and also responsible for a significant portion of engineering and other efforts) [19]. Finally, the Ubuntu community also chose to follow a meritocratic governance model
[22] that includes teams (similar to our community groups), a technical board (similar to our ARC), a community council (similar to our OGB), and a clear leader (otherwise known as SABDFL
[23] to which we currently have no equal).
With the current governance structure of our community, the author does not see a way for our community to achieve the required sustained growth and success. The current governing board, as originally intended, does not have the ability to provide the clear leadership that our community requires. Recent discussions on our mailing lists have made it quite clear
[24] that community groups are responsible for the day-to-day leadership and activities within our community
[25].
However, the constitution does not specify which community groups are responsible for which activities and leadership roles. As a result, recent discussions [
26,
27] have lead to limited consensus about whom has authority over these activities that are seen as representing the entire community
[28].
It is apparent, as a result of this, that our community is unable to decide what the vision, purpose, and goals of our community really are. This is surprising, given they were defined at one point
[29]. Clearly, our community has not decided who represents our leadership in these areas.
This leadership issue has resulted in the failure of our community to achieve the level of sustained growth and success required for our continued existence. It also illustrates the need for clear, inspired leadership within our community. Clear, inspired leadership naturally provides the sort of vision [4] that we require to determine a clear direction, and a set of goals; while ensuring consensus about who controls different areas of the community.
The real issue behind our current troubles is not primarily technical or logistical (as the author erroneously previously believed) in nature; it is not about naming, trademarks, or branding; it is about the failure of community groups to take up the responsibilities, that the OGB, empowered by our constitution, has delegated to them [24] (which may be because they were not informed of this delegation; adequately or at all).
The primary failure, as the author sees it, is that we currently have no efficient, effective way to ensure that decisions will be made when no clear consensus exists
[30]. The OGB is currently unable to guide our community in many situations, as they are not empowered to do so, and voting on every issue is likely to end in deadlock either due to the apathy of eligible voters
[31] or a vocal minority that prevents consensus from being achieved.
Consequently, the author believes that our governance structure must be revised to prevent the current retardation in the growth and success of our community. To achieve this, a specific individual must be designated to provide the clear, inspired leadership role that our community needs, along with empowering the OGB, to ensure that an overall direction and vision for our community can be established.
We need to remember that Sun deserves special consideration as the founding member of this community, the primary (and only as far as the author is aware) financial supporter, and trademark holder. They deserve, nay, must, have a key role in directing this community, especially in the areas of product development and marketing.
The author believes that Sun must be allowed to fulfil this role as a key leader for many reasons. Sun is accountable to their shareholders for their significant financial support they provide to the OpenSolaris project; whether that is directly or through a foundation is immaterial. The potential for our sustained growth and success, and to a certain extent, Sun's, is directly tied to the community. Failure or Success by either Sun or of the community will affect both.
In conclusion, to resolve these deficiencies, it is the author's belief that our existing governance structure must be revised to ensure that clear, inspired leadership is provided to our community (as a whole). Communities surrounding the Linux kernel
[32], Ubuntu, Python
[33], OpenBSD, Apple, et al. have shown us how significant, sustained growth and success can be achieved when a specific individual helps provide the clear, inspired leadership every community needs. Our community would not even exist today if were not for the decision of leadership (notably, Jonathan Schwartz)
[34] at Sun to provide the source code that provides the reason for our existence.
The only remedy available, in the author's view, is to ensure that four changes are made by amending our constitution:
1) The OGB is empowered to make more decisions for the community.
2) An individual is chosen by our community to work with the OGB. They will provide clear, inspired leadership and vision. It must be made known that this position is one that is likely to be full-time and require their complete focus. Any individual that is part of our community should be eligible for this position regardless of whom they are or are not employed by.
3) That Sun is permitted, as the principal stakeholder in our community, to play a key role in product development and marketing of the OpenSolaris trademark (which they own) given their clear experience, accountability to their shareholders, and success in this area. This role must be given a greater degree of authority than what is currently granted by the constitution.
4) That the role of product development and marketing, as outlined in our constitution, should be shared with Sun in a well-defined manner with qualified members of the community.
Without Sun's continued support, the vision and clarity of inspired leadership that a qualified individual can provide, and the further empowerment of the OGB; the community will not achieve the sustained growth and success that it requires for its continued existence
[35].
--
Shawn Walker
P.S. The author appreciates and acknowledges the help of other community members (you know who you are!) who persevered ( :-) ) through earlier drafts of this document. This proposal would not have been possible without your assistance.
P.P.S. This proposal was originally posted on the ogb-discuss mailing list. You can read the responses to it
here. I encourage everyone to respond on the ogb-discuss mailing list (or here of course).
