[OSGeo-Discuss] OSGeo/LocationTech relationship

[Jody Garnett]

> > I may as well link to my more recent talk (https://vimeo.com/142989259)
> as
>
> Interesting talk Jody, thank you !
>

Glad you enjoyed it, afraid I was a bit punch drunk after a long conference.


> One thing it wasn't clear to me (I might have dreamt it):
> did you say that LocationTech only accept non-copylefted projects
> in the foundation ? I think it came out by the very end
> of the talk, in response to a question from Luca Delucchi (~30:00)
>

A license conversation :) Now I know we are on an open source discussion
list :)

It is a bit more restrictive than that, there is a set list of approved
licenses (similar to how Apache Foundation, Free Software Foundation and
others operate). The Eclipse Foundation is actually more pragmatic in this
then many of the software foundations - they are not limited to their own
licenses for example.

Let me see how I do from memory (please correct me if I am wrong):

- They had some IBM license which was replaced with the Eclipse Public
License (I kind of view it as LGPL but updated with language around patents
and trademarks...)
- Eclipse Distribution License (exactly like BSD except the organization is
"Eclipse Foundation" is the organization - took me ages to figure that it
was just a BSD style license).
- Permissive licenses (Apache, BSD, MIT)
- LGPL is a borderline case - it is a "business friendly" license but it is
not on the approved list. We wrote a couple projects into the LocationTech
charter (JTS, GeoTools, GDAL) and asked the eclipse board for permission.
It appears that this process can be repeated on a case-by-case basis.
- the board has been asked about GPL and AGPL and both of those have been
rejected as restrictive licenses.
- open data license are a work in progress, we had to ask the board for
approval distribute the EPSG data base (which is a very odd open data
license)

One thing that was a shock to me during uDig integration was the focus on
making sure the build chain was all free. I had to remove some UML diagrams
that were produced with ObjectAid (that while free to download and use was
not open source).

Even though the above sounds cut and dried, all these foundations have a
mandate to promote open source. Some like the Apache Foundation and Eclipse
Foundation work very well together (anything produced by Apache is
automatically approved for use by an Eclipse Project). Even ones that are
are approaching the game from different vantage points have that common
ground - the Eclipse Foundation and Free Software Foundation where very
helpful untangling some license questions for an OSGeo project during
incubation.

I find the language used about licenses (permissive / restrictive,
free/libre, commercial friendly) tend to be less important that how the
licenses are used. Specifically what a projects intentions are when they
choose a license. One interesting take is to dual license software (no I do
not mean the open source / commercial sense - I view that as a kind of rude
way to monitize community effort). As an example uDig switched its license
to dual BSD/EPL in order to collaborate with two diverse groups of
developers. EPL allows easy communication with eclipse RCP developers (all
of which have had the EPL license approved for use in their organization).
BSD (as a universal donor) allows code to be shared with GeoTools and
friends.

One thing I want to say is the above is a pragmatic choice - as
professionals we rarely get get choose what license we work with (as we are
doing work on behalf of customers). For the bulk of my career I have been a
consultant and clients demanded LGPL (or similar). Have had the good
fortune to work on GeoServer (which is GPL). Sadly GPL is not meetings its
original intension of making the source code available to users, since many
applications are being run on servers/cloud .. so the GPL only provides
source code to the system administrator running the software. AGPL seeks to
correct this with mixed success.

While LocationTech software is restricted in what it can depend on,
LocationTech events are welcome to all. I mentioned PostGIS, GeoServer and
QGIS being the focus of student workshops. At our previous LocationTech
event Refractions Research showcased a new geocoder they had written for
the BC Government. We were all very excited that the work was being made
open source ... and released under a AGPL license.

By comparison - OSGeo accepts anything that is approved by the Open
Software Initiative (this diversity is a key strength). However OSGeo has
its rarely explored limits at the boundary between open source software
that requires a closed source platform. but we get very uncomfortable
though when the software requires a commercial component be installed. It
was difficult when the Java programs joined, although Java is now open
source some of the extensions for Java are still proprietary.

I tried browsing the locationtech.org website but found no mention of this
> limitation.
>

I thin it is in the charter, most things are in the charter:
https://www.locationtech.org/charter

The Intellectual Property Policy of the Eclipse Foundation will apply to
all activities in this group. The group will follow the Eclipse
Foundation's IP due diligence process in order to provide clean open source
software released under licenses approved by the group and the Eclipse
Foundation Board of Directors. Approved licenses for this group include
EPL, MIT, BSD, and Apache 2.0. This list may be amended from time to time
by the group and the Eclipse Foundation Board of Directors. The EPL license
is the recommended license for projects in this group.




> It's confusing, because you early mentioned that projects can be
> in both foundations while if that's confirmed projects like PostGIS,
> GRASS or QGIS (to name a few) could _not_ be.
>

That is correct - joining any group is based on agreeing with their polices.

My point was OSGeo is not exclusive, GRASS and QGIS are welcome to join the
Free Software Foundation for example.


> It is interesting that they have dedicated IP stuff, would come to
> think the actual goal is to help companies ride the "Open Source" tide
> (still big, and still growing) w/out risk of getting wet...
>

Going to guess you mean "staff".  Yes it is nice, but it is not a free pass
- they check way more things than I expected.

The other side of the coin is that it is a huge responsibility to publish
software as open source (and something very easy to screw up).

The ability to point to a public IP review can open a lot of doors for a
project. Even if you do not have staff to review your project ... those
reviews are still happening (you are just not being told about it).

I have told this story and will tell it again. GeoTools gets one or two
external IP reviews a year. Usually we hear about them after the fact ("oh
I tried to use geotools but legal would not let me", "why what was the
matter?", "Oh I cannot tell you as that would be legal advice", ".... um
... I am sorry to hear that"). We have had a couple willing to talk to us
recently (IBM China and LocationTech) so hopefully I will have less of
these awkward conversations.
--
Jody

>
> --strk;
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/discuss/attachments/20151113/2de36243/attachment-0002.html>