Rasdaman as a Research Project? [was: Re: Should OSGeo accept "benevolent dictator" projects into OSGeo? ]

classic Classic list List threaded Threaded
1 message Options
Reply | Threaded
Open this post in threaded view

Rasdaman as a Research Project? [was: Re: Should OSGeo accept "benevolent dictator" projects into OSGeo? ]

Bruce Bannerman-3
Hello Peter,

Rasdaman is certainly an impressive collection of software, however I'm concerned at this description of Rasdaman as a Research Project.

If we go down this route we are severely limiting the Project's potential and uptake.

I have been discussing Rasdaman with a number of people over the last year or so and am seeing a number of large organisations monitoring activities and development within the community project to assess its potential for collaboration and for use within their organisations.

These organisations will not be interested in deploying operational software that has been sourced from a project that badges itself as a Research Project.

This is why the Rasdaman Community Project needs wider representation in in the PSC, so that it can achieve its true potential.

I suggest that we pause for a while and reflect on the discussions over the last few weeks prior to taking further action.

Bruce Bannerman
OSGeo Mentor to the Rasdaman Project

From: Discuss <[hidden email]> on behalf of Peter Baumann <[hidden email]>
Organization: Jacobs University Bremen
Date: Thursday, 5 May 2016 at 20:24
To: Cameron Shorter <[hidden email]>, Even Rouault <[hidden email]>, Incubator <[hidden email]>
Cc: Discuss OSGeo <[hidden email]>
Subject: Re: [OSGeo-Discuss] [Incubator] Should OSGeo accept "benevolent dictator" projects into OSGeo?

Hi Cameron,

I tried very much to make the situation transparent. Maybe the notion of Principal Investigator helps here (cf Wikipedia - although biased towards medical science):

A principal investigator (PI) is the holder of an independent grant administered by a university and the lead researcher for the grant project, usually in the sciences, such as a laboratory study or a clinical trial. The phrase is also often used as a synonym for "head of the laboratory" or "research group leader." While the expression is common in the sciences, it is used widely for the person or persons who make final decisions and supervise funding and expenditures on a given research project.

I am the PI of rasdaman, and that will not change, also not indirectly through wordsmithing as proposed.

OSGeo is entering new domains with rasdaman, which is: scientific research projects. Like some other communities, these have existed long before OSGeo, and have their own ethics, procedures, and rules. It is unlikely that science will change and give up freedom of research based on its principles well accepted by the whole community. If OSGeo intends to change these in general then maybe starting with rasdaman as an isolated item in a vast universe is not the optimal point.

OSGeo may find out that its very special (although obviously not unambiguously codified) views constrain it to particular ecosystems. But I am not imposing nor judging. Just trying to explain.


On 05/04/2016 09:18 PM, Cameron Shorter wrote:
Hi Peter,
Could you please answer Even and Johan's question.

I'm happy to use another term for the governance model.
"Does one person have ultimate control over the project? Or does ultimate control lie with a committee, possibly with a tie breaker vote designated to one person or one role (eg chair)?"

Warm regards, Cameron

On 5/05/2016 3:29 am, Even Rouault wrote:
Le mercredi 04 mai 2016 18:34:27, Peter Baumann a écrit :
HI Cameron,

first, as this word has been used too often now, the current model has
nothing at all to do with dictatorship. What is the suggested opposite,
BTW - "dictatorship of majorities"? ;-)
Actually reading http://www.rasdaman.org/wiki/Governance it seems the sentence
that cause trouble is "Should such consent exceptionally not be reached then
Peter Baumann has a casting vote." Does that mean that in case there's a tie
in voting (which cannot happen with a 3 member PSC as currently), Peter breaks
the tie ? If so, that seems acceptable to me (should probably be rephrased in
a more neutral way to say to designate the chair of the PSC rather than a
named individual).

I actually see that Johan Van de Wauw asked the same question but this hasn't
been answered clearly.

Perhaps http://www.rasdaman.org/wiki/Governance could gain in clarity by
defining precise voting rules (which majority, delays, etc...) As an example of
simple rules (not necessarily to follow them, but to show the plain language
used): https://trac.osgeo.org/gdal/wiki/rfc1_pmc /
http://mapserver.org/development/rfc/ms-rfc-1.html /
http://docs.geoserver.org/latest/en/developer/policies/psc.html ).

If it would at least be called a "technocracy", that I could accept:
rasdaman has always been driven by purely scientific elaboration _and_
consensus orientation and respect. Genius rules, regardless where it comes
from - this is at the heart of our scientific progress.

It is the fundamental freedom of science that is at stake here.

I guess that OSGeo needs to decide whether it can accept a model based on
scientific ethics ...or not.


On 05/04/2016 02:01 PM, Cameron Shorter wrote:
Hi Peter,
Are you open to considering relinquishing rasdaman's current "benevolent
dictator" governance model?

Many (most?) OSGeo projects that I'm aware of are managed similarly to
your description below.
There is usually a sage or two amongst the community, typically someone
who founded the project. The sage(s)  have more experience with the
project, and their opinion holds greater weight amongst the community.
This informal relationship continues even with a formal Project Steering

As you would understand, building a successful Open Source community
involves a significant amount of mutual respect, and mutual recognition
of team members. Community members typically show respect by giving
extra weight to the opinion of founders, and founders often show respect
and trust of their community by sharing project governance.

If you are a good open source leader, and it appears you must be, there
is little risk you will loose your current influence on the project. Its
also unlikely there will be an unresolvable difference between yourself
and the community. But if there is, and the project forks, whether you
are head of the official PSC or the new rouge PSC will have little
impact on the final result.

So please do consider adopting a shared PSC governance model.

If you do wish to go ahead with a "benevolent dictator" model, I agree
with Andrea's that we should put the question to OSGeo Charter members
to vote, as it would be a new direction for OSGeo.

Warm regards, Cameron

On 3/05/2016 5:46 pm, Peter Baumann wrote:
interesting discussion, with valuable thoughts!

True, micro management is not the case in rasdaman - on the contrary, we
are most happy about helping hands, and are constantly thinking about
opportunities for process improvements. Personally, I am so much
overloaded that I enjoy handing over tasks, and yes: with appropriate
responsibility; in practice that means that we openly discuss pros and
cons with myself being "primus inter pares" (first among equals). I
have not received any complaint over the years that anybody would not
get heard appropriately. Regularly I just need to lean back
(metaphorically) and await the outcome of the discussion of the
experienced developers, and add my nodding to the group consensus.

We regularly try to involve the community in such design and
implementation discussions (and I am urging devers to do that), but
feedback invariably was minimal. Which I see as a sign of trust when
looking at the download figures at www.rasdaman.org.

It may be worth noting that we have installed mechanisms for openly
commenting and voting on patches; ever clicked on the Review URL in the
Patch Manager?

Actually, it is more about deciding not by election, but by
qualification. Concepts and code of rasdaman are extraordinarily
complex; large and experienced companies like Oracle, Teradata, and
ESRI have tried to copy rasdaman, and failed. Therefore, it
unfortunately takes patience for a newcomer to immerse to a degree that
allows making suggestions that are fully backed by the team. That said,
we do not attach maturity labels to coders ;-), rather the technical
merit of each individual contribution is weighted carefully.

Another constraint, of course, are project considerations- there is a
contract behind where ESA, the European Commission, or whoever-else
expects fulfilment.

Bottom line, the atmosphere in rasdaman is highly cooperative and
consensus-based, I just reserve jumping in as a last resort. Someone has
questioned the term used in this discussion as not quite adequate; I
like the diplomacy aspect raised.


On 05/03/2016 01:54 AM, Julien-Samuel Lacroix wrote:
I found this nice description of the benevolent dictator governance:

It's a nice read, but I want to highlight this part:
In many ways, the role of the benevolent dictator is less about
dictatorship and more about diplomacy. The key is to ensure
that, as the project expands, the right people are given influence
over it and the community rallies behind the vision of the project
Another good one from (linked from the above):

they let things work themselves out through discussion and
experimentation whenever possible. They participate in those
discussions themselves, but as regular developers, often deferring to
an area maintainer who has more expertise. Only when it is clear that
no consensus can be reached, and that most of the group wants someone
to guide the decision so that development can move on, does she put
her foot down and say "This is the way it's going to be."

  From my (really) naive point of view, the "benevolent dictatorship" is

do-ocracy were the committers get the right, or influence, to lead
parts of the projects and where the "dictator" is acountable of its
decision to the community. The key ingredients are the same as other
governance : - Be easy to contribute patches and features
- Be open on the direction of the project
- Be forkable

If someone wants to contribute a new feature, they ask the mailing-list
and the committer responsible for this part of the software, not the
"dictator", will approve or suggest changes. The approach is less
formal than with a PSC, but still works the same.

This is of course an ideal scenario, but can be as open as a PSC, I
think, as long as the project as a good "forkability".

Back to the incubation discussion, Rasdaman seems to have multiple
committers and 2 main organisation behind it. What I would like to ask
is, what's the "bus number". Is there a second (or third) in command
that could ultimately take care of the project after the dictator's
"end-of-term"? From my point of view, a PSC of 3, 2 being from the
same company, is a small PSC and will probably lack a bit of variety
in opinions. Is there any other key contributors that the "dictator"
refers to when trying to get inputs and defer technical decisions?


On 16-05-01 07:29 AM, Jody Garnett wrote:
This is kind of a larger topic than just the incubation committee, but
no I do not believe we should. It is a defining characteristic of our
foundation to not place many restrictions on our projects - but demand
that the projects be inclusive and open to collaboration.

I do not believe that the "benevolent dictator" fits this ideal.

I also do not think we need to stress the PSC approach as the one true
way, smaller projects that only wish to have committers vote on
decisions (rather than form a PSC) is perfectly acceptable - provided
there is a provision for new committers to be added into the mix.

We also have an outstanding request from our president to make the
foundation more inclusive. With this in mind we are a lot less
demanding on our community projects - which provides a way for
projects that do not meet some of our ideal criteria to be part of
the foundation.

On 1 May 2016 at 00:44, Cameron Shorter <[hidden email]

[hidden email]> wrote:
      OSGeo discuss, OSGeo incubation, OSGeo board,
            I'm hoping the greater OSGeo community will consider and comment
      on this question:
            Should OSGeo accept a "benevolent dictator" [1] governance model
      for incubating projects?
            -0 from me, Cameron Shorter.
      * As part of incubation, Peter Baumann, from Rasdaman has
      requested a "benevolent dictatorship" governance model [2].
      While "benevolent dictatorships" often lead to successful
      projects, all prior OSGeo incubated projects have selected
      "equal vote by PSC members". Someone with better legal training
      than me might find "benevolent dictatorships" to be
      unconstitutional according to OSGeo bylaws. [3]
            [1] Eric Raymond's "Homesteading the Noosphere":
      6.html [2] http://www.rasdaman.org/wiki/Governance
            On 1/05/2016 3:56 pm, Peter Baumann wrote:
            I understand where you are coming from, and your
      characterization is definitely correct. While our process is
      and always has been absolutely open to discussion so as to
      obtain the scientifically and technically best solution this
      "benevolent dictatorship" has brought rasdaman to where it
      stands now - it is designed by innovation, not by committee.
      Just to get me right, our model is certainly not the right one
      for every endeavour. Here it is the most appropriate, and hence
      we will keep it.
            As you observe, this model is not contradicting OS as such, and
      many projects run it. So ultimately it lies in the hand of OSGeo
      to decide whether they accept the existing plurality of
      approaches (in this case manifest with rasdaman).
            On 04/30/2016 10:47 PM, Cameron Shorter wrote:
      Bruce, Peter,
      I've read through the incubation process documentation, and can
      only see one thing which I think breaks our OSGeo principles.
            The Governance model includes a statement:
      "In all issues, the PSC strives to achieve unanimous consent
      based on a free, independent exchange of facts and opinions.
      Should such consent exceptionally not be reached then Peter
      Baumann has a casting vote."
            This is describing a "benevolent dictator" model, which has
      proved to be an effective model for many open source projects.

      See Eric Raymond's "Homesteading the Noosphere":

      However, it is not in line with existing OSGeo Incubated
      projects, which have documented a "vote by PSC" as the defining
      governance process. In practice, the PSC community debate
      alternatives, and if needed, respectfully revert to reasoned
      advice provided by the "benevolent dictator".
            Peter, are you open to changing the governance model to a "vote
      by PSC"?
      I'd be comfortable with a "vote by PSC, with PSC chair being
      given 1.5 votes to break any deadlocks. I'd also be ok with PSC
      chair defaulting to Peter (as founder), until such time as
      Peter resigns from the role."
            Warm regards, Cameron
      Cameron Shorter,
      Software and Data Solutions Manager
      Suite 112, Jones Bay Wharf,
      26 - 32 Pirrama Rd, Pyrmont NSW 2009
            P<a href="tel:%2B61%202%209009%205000" value="+61290095000" target="_blank">+61 2 9009 5000 <tel:%2B61%202%209009%205000>,

<http://www.lisasoft.com>,  F<a href="tel:%2B61%202%209009%205099" value="+61290095099" target="_blank">+61 2 9009 5099

      Incubator mailing list
      [hidden email] [hidden email]

Incubator mailing list
[hidden email]

Dr. Peter Baumann
 - Professor of Computer Science, Jacobs University Bremen
   mail: [hidden email]
   tel: <a href="tel:%2B49-421-200-3178" value="+494212003178" target="_blank">+49-421-200-3178, fax: <a href="tel:%2B49-421-200-493178" value="+49421200493178" target="_blank">+49-421-200-493178
 - Executive Director, rasdaman GmbH Bremen (HRB 26793)
   www.rasdaman.com, mail: [hidden email]
   tel: 0800-rasdaman, fax: 0800-rasdafax, mobile: <a href="tel:%2B49-173-5837882" value="+491735837882" target="_blank">+49-173-5837882
"Si forte in alienas manus oberraverit hec peregrina epistola incertis ventis dimissa, sed Deo commendata, precamur ut ei reddatur cui soli destinata, nec preripiat quisquam non sibi parata." (mail disclaimer, AD 1083)

Discuss mailing list
[hidden email]