[Incubator] platforms discussion

Jody Garnett jody.garnett at gmail.com
Thu Sep 23 09:37:46 PDT 2021 

Bruce:

We presently follow the OSI approved list, much easier to follow then to
manage our own criteria.

However this discussion appears to be around governance and control; and I
think we are very clear on these topics. Even during the Rasdaman
discussion we were very clear that the benevolent dictator model did not
meet our goals for open governance. It was frustrating this the discussion
took so long: benevolent dictator is of course a valid model, it is just
not one that matches our ideals.

OSGeo does support (and sometimes require) CLAs providing
asymmetrical rights on code bases such as GeoTools and GeoServer (
https://www.osgeo.org/about/licenses/). This increased permissions for the
PSC has allowed these committees to donate fixes to more permissive
projects such as GeoServer (GPL with EPL exception) --> GeoTools (LGPL); or
GeoTools (LGPL) to JTS (BSD). The use of CLA to establish
asymmetrical rights has a community building use in this respect; however
the same tool is used to enable some of the harmful (to the open community)
practices being remarked on.

It is a shame folks like Elasticsearch and MapBox abandoning open source,
as I really appreciate it as a way to produce excellent software and manage
risk across a wide range of stakeholders. In many cases the justifications
feel self imposed. If an organization is not seeing enough contributions to
justify continuing a project as open source, I can point to an organization
that did not setup equal governance to promote contributions between equals.

Personally I like free-software, and open-source software held by a vendor
neutral organization (such as OSGeo). Everything else seems a bit risky.

OSGeo does not provide such a hard line, but by stressing projects need to
be willing for others to take part in governance (no benevolent dictator
model) I believe the foundation is clear in what kind of project we
support. There is a reason projects such as MapBoxGL or Cesium did not
apply to be part of OSGeo after all.
--
Jody Garnett


On Wed, 22 Sept 2021 at 17:31, Bruce Bannerman <
bruce.bannerman.osgeo at gmail.com> wrote:

> Hi Jody,
>
> These thoughts extend on your ‘there can be only one’ comment below.
>
> I think that we in the Incubation Committee and our potential future
> incubating projects would benefit from our better defining what type of
> open source project we will support.
>
> Greg Troxel contributed valuable insight into a discussion that we were
> having in the OSGeo Standards list on the proposed OGC/OSGeo MOU. See [1]
> for context.
>
> With apologies to Greg for my paraphrasing, I understand Greg’s arguments
> to strongly differentiate between open source projects that are subject to
> proprietary relicensing and those that are not.
>
> Greg contributed the blog link at [2] as part of the discussion. This is
> an insightful read.
>
> If we had had a clear policy on the type of open source project that we
> are willing to support during the unfortunate situation with the Rasdaman
> Incubation vote, we could have avoided a lot of the angst and
> miscommunication that occurred on both sides.
>
> So before we progress too much further on the ‘Platforms’ discussion, I
> think that it would be beneficial to resolve the OSGeo Incubation supported
> open source model issue definitively.  I believe that this will make it
> much easier to progress the platforms discussion, and future incubations.
>
> While I have concerns over the interpretation of what a Contributions
> Licence is, I believe that the links at [1] and [2] are a good starting
> point to get this sorted out.
>
> Kind regards,
>
> Bruce
>
> [1] https://lists.osgeo.org/pipermail/standards/2021-September/001274.html
>
> [2] https://sfconservancy.org/blog/2020/jan/06/copyleft-equality/
>
>
> On 17 Sep 2021, at 01:59, Jody Garnett <jody.garnett at gmail.com> wrote:
>
> 
> It is kind of like the “highlander-principle” if the phrase “there can be
> only one” applies … you are not open source.
>
> Jody
>
> On Wed, Sep 15, 2021 at 6:38 PM Jody Garnett <jody.garnett at gmail.com>
> wrote:
>
>> There is a useful definition of a framework (quite technical) here
>> https://en.wikipedia.org/wiki/Software_framework
>>
>> Software frameworks have these distinguishing features that separate them
>>> from libraries or normal user applications:
>>> https://en.wikipedia.org/wiki/Software_framework
>>>
>>>    - inversion of control - In a framework, unlike in libraries or
>>>    normal user applications, the overall program's flow of control is not
>>>    dictated by the caller, but by the framework.[1]
>>>
>>>
>>>    - default behavior - A framework has a default behavior. This
>>>    default behavior must actually be some useful behavior and not a series of
>>>    no-ops.
>>>
>>>
>>>    - extensibility - A framework can be extended by the user usually by
>>>    selective overriding or specialized by user code providing specific
>>>    functionality.
>>>
>>>
>>>    - non-modifiable framework code - The framework code, in general, is
>>>    not allowed to be modified. Users can extend the framework, but not modify
>>>    its code.
>>>
>>> While all of that is technically true it is perhaps a bit too detailed
>> for our purpose.
>>
>> The core distinction is earlier in the thread:
>> - Does your open source code support a single website? Or is is setup for
>> use by others?
>> - Are you building a community around services? This is a user community
>> ...
>> - Are you building a community around software? This is still a user
>> community ...
>> - Are you building a community around software where the software source
>> code is available to look at? This is still a user community ... looking at
>> you Elasticsearch
>> - Are you building a community around software with shared
>> responsibility and risk (enabled by a license to view *and change* source
>> code)? This is a free or open-source community (depending on which license
>> chosen by the group)
>>
>> --
>> Jody Garnett
>>
>>
>> On Fri, 3 Sept 2021 at 18:32, Bruce Bannerman <
>> bruce.bannerman.osgeo at gmail.com> wrote:
>>
>>> Jody,
>>>
>>> I don’t expect anything constructive from the AGM. People won’t have
>>> time to reflect.
>>>
>>> …now what is a ‘Framework’?
>>>
>>> I suggest that we define what we support and take it from there. We can
>>> always adjust, if required.
>>>
>>> Kind regards,
>>>
>>> Bruce
>>>
>>> On 4 Sep 2021, at 05:27, Jody Garnett <jody.garnett at gmail.com> wrote:
>>>
>>> 
>>> Thanks Bruce,
>>>
>>> I did not get anything useful from the board; perhaps it is a subject
>>> for the AGM.
>>>
>>> For now the way forward seems to be to recast the platform as a
>>> framework and ensure the resulting software stack and be picked up and run
>>> independently (with a quickstart or similar).
>>> --
>>> Jody Garnett
>>>
>>>
>>> On Wed, 14 Jul 2021 at 21:57, Bruce Bannerman <
>>> bruce.bannerman.osgeo at gmail.com> wrote:
>>>
>>>> Thanks Jody,
>>>>
>>>> Personally:
>>>>
>>>>
>>>>    - I don’t mind a situation where an open source project attempts to
>>>>    develop and maintain software that is intended to integrate a number of
>>>>    software components into a working product which could perhaps be called a
>>>>    platform.
>>>>
>>>>
>>>>
>>>>    - I can see many situations where the ‘platform’ might deploy both
>>>>    components and the platform’s specific customisations concurrently.
>>>>
>>>>
>>>>
>>>>    - However, I would not support a situation where that product (or
>>>>    platform) can only be implemented once. I’d prefer that it can be
>>>>    implemented many times by different organisations.
>>>>
>>>>
>>>>
>>>>    - While such a ‘platform’ project would look after its own
>>>>    "integration related software", I’d see that the individual components
>>>>    would be subject to their own open source project community’s governance
>>>>    practices.
>>>>
>>>>
>>>>
>>>>    - This could get quite messy, when the integration related software
>>>>    is actually a customisation of an existing software component with its own
>>>>    open source community already in existence. This would require careful and
>>>>    close collaboration between both communities…
>>>>
>>>>
>>>> That will do for now, let’s see what others think.
>>>>
>>>> Kind regards,
>>>>
>>>> Bruce
>>>>
>>>>
>>>>
>>>>
>>>> On 15 Jul 2021, at 01:21, Jody Garnett <jody.garnett at gmail.com> wrote:
>>>>
>>>> It is indeed overloaded, and no I cannot clarify as the applicants that
>>>> are coming in are slightly different from each other.
>>>>
>>>> Turn-key portals such as
>>>> https://www.osgeo.org/choose-a-project/information-technology/portal/
>>>> these showcase a range of projects. Some like geomoose are presented as
>>>> frameworks, others like GC2/Vidi are presented as a platform.
>>>>
>>>> I would be cautious about an open source project that just supports a
>>>> single website (like http://github.com/mapstory), but perhaps that is
>>>> my own bias? There is an advantage to users of a platform being able to
>>>> review the code responsible for the service they are using. But this
>>>> represents new ground for OSGeo, hence the discussion.
>>>>
>>>> I also brought this discussion to the osgeo board list; so we do not
>>>> need to decide on our own.
>>>> --
>>>> Jody Garnett
>>>>
>>>>
>>>> On Tue, 13 Jul 2021 at 18:56, Bruce Bannerman <
>>>> bruce.bannerman.osgeo at gmail.com> wrote:
>>>>
>>>>> Hi Jody,
>>>>>
>>>>> The concept of a platform is quite overloaded and means different
>>>>> things to different people.
>>>>>
>>>>> Can you please clarify what you mean by ‘platform’?
>>>>>
>>>>> Kind regards,
>>>>>
>>>>> Bruce
>>>>>
>>>>>
>>>>> > On 12 Jul 2021, at 18:30, Jody Garnett <jody.garnett at gmail.com>
>>>>> wrote:
>>>>> >
>>>>> > Discussion topic for incubation committee:
>>>>> >
>>>>> > We are getting applications from platforms seeking to join OSGeo.
>>>>> >
>>>>> > What do you think?
>>>>> > --
>>>>> > Jody Garnett
>>>>> > _______________________________________________
>>>>> > Incubator mailing list
>>>>> > Incubator at lists.osgeo.org
>>>>> > https://lists.osgeo.org/mailman/listinfo/incubator
>>>>>
>>>>>
>>>> --
> --
> Jody Garnett
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/incubator/attachments/20210923/2d274e94/attachment-0001.html>

More information about the Incubator mailing list