[Board] a long story starting with SOAP and ending in Switzerland
jo at frot.org
Tue Jul 10 11:10:59 PDT 2007
This is a long post-conference braindump which i would really
appreciate it if people read, as i need help to break it down into
actionable component parts...
Okay. I am sitting here in the Porto airport lounge writing a
SOAP opera. This is after having returned from the EC-GIS
INSPIRE/Envirodata workshop. During which Athina got me into a
couple of really interesting chats, which got my head spinning
about the future of OSGeo here in Europe, and beyond...
The JRC is, roughly, the technocratic arm of the European Commission.
This means they coordinate, run, fund and participate in many
different data/systems activities. For me their context has always been
they are running the INSPIRE consulting and implementing process
(the European SDI initiative, due to become binding May 2008).
But they also do a lot of other e-government, info society things.
They are responsible for putting together a "reference implementation"
for INSPIRE, and for proving the concepts behind its Implementing Rules.
In-house, they use at least Geoserver, Mapserver (and thus Geotools
and GDAL), GeoNetwork and gvSIG already. There is a pencilled-in
invitation now for the European directors of OSGeo to visit them
in early October, talk about the projects, the foundation, etc.
This is a SOAP opera because of the Implementing Rules on Network
Services, which are not published in draft yet but are being
openly discussed. These Rules describe the set of interfaces
which a system, or system of systems needs to have in order to
be judged "compliant" with the European SDI Directive.
By "Network Services" is meant W*S, CSW, etc - all and any
internet interfaces geodata sharing and processing systems.
These Rules are put together by Drafting Teams, which are composed
of "volunteers" whose time is given by all kinds of organisations.
The process is not observable from the outside until the release
of the first draft. Then they can be responded to by SDICs / LMOs,
which are like community interest groups. We (OSGeo, the FOSS GIS
community at large) have a voice in this thanks to Markus' foresight
in registering a FOSS SDIC, early on in the INSPIRE process.
Now, the Network Services drafting team have decided to *mandate*
SOAP for transfer and UDDI for registry stuff. Now as far as i am
concerned these are corporate-intranet standards which had hype bubbles
several years ago and are not much used on the wider internet.
But, this is a different story for a different email.
The JRC seem terrifically pragmatic. The presentation that impressed
me about most in Porto was the one from ConTerra about the exercises
that the JRC funded them to carry out in testing the usefulness of
the more complex catalog standards (CSW2, ebRIM), finding the design /
engineering flaws and being candid about them, and getting the
standards process fixed, to some extent, through code-based reification.
They are implementation-stresstesting OGC specifications, neutrally.
Not being an OGCista, i hadn't heard about this SOAP thing - that
SOAP interfaces to OGC W*S is are not defined or not well-defined,
and there is quite some talk now of retrofitting this to the specs.
But will it work? Well, the only way to find out is to try it.
So later in the year the JRC will start a procurement process,
to get different projects to implement SOAP interfaces to network
services as part of the reification process. "Will it work?" because,
if it doesn't, that's going to look really lame in European law.
They talk to gvSIG and GeoNetwork a lot socially, already.
But their projects are more often on the scale where you need at
least one space agency and one multinational infraco in your
consortium even to get noticed. Software projects are too small.
But OSGeo taken as a collective, could be big enough to talk to.
This is Europe. Which means things that get funded with European
money, can't directly create value for non-European entities.
And the concern is this:
The projects sign over their code copyright to the Foundation.
This equates to "intellectual property rights", about which
people worry too much. To fund work on a project which has
copyright held by a US based corporation - even if that is
a OSI style copyright license - is to create intellectual
property for a US organisation, and thus unpalatable.
I think back to the GeoTools CLA discussion that went on so long.
What bogged that down more than anything (from an outsider POV)
is that the core GeoTools committers are from all over the world
and the CLA made statements oriented to US law, and in some cases
made statements counter to e.g. the napoleonic law that operates
in Mittel Europe. And US-based lawyers are surprised now at the
transnational-style CLA which the GeoTools community offered.
An "OSGeo Europe" could be an option. But I don't think this
would address the perception-of-IP problem. Assertion of copyright
couldn't be split into different territorial entities without
causing a confusing mess. The local chapters, even with
independent legal status and accounts, are more about end users.
Free software has no territory!
When i got back, i braindumped to a friend who suggested this:
- Emigrate the Global OSGeo legal entity to Switzerland.
Nowhere says "neutrality" like Switzerland. There are a lot of
collaboration treaties etc with Europe. If an OSGeo Europe
was needed as a paper shell, that would be possible, but it
shouldn't be. To have the Foundation locally papered as a 501(c)3
in the US would still be terrifically useful.
There is a really good GIS scene in Switzerland.
More information about the Board