One thing I've been thinking about a lot is how to encourage more user collaboration in geospatial information. I've given some recent talks on this, and have found a lot of resonance, so want to start to actually build some of the ideas in to GeoServer (mostly as extensions, so we're not polluting the core with all kinds of crazy users).
The central idea is to get at the things SDI initiatives want - people sharing data. Most of the SDI initiatives I've seen have started with Metadata (get lots of catalogs), then focused on Users (get people filling out metadata and searching for it), and finally end up on Data (it will come after we get lots of metadata and users).
I've viewed GeoServer's approach as the opposite - start with data, getting real data on the web. The next step is to focus on users, and after that lots of good metadata will fall out.
How? Through users doing what they do on an infrastructure that supports them. Think youtube - with most viewed, highest rated, most favorited. Imagine GeoServer letting users create maps of sets of layers, with different styles and uploaded data. And imagine them rating and commenting on everyone else's. This produces valuable metadata about which data is actually used by others.
So today I started an RnD page on the improvements I see needed to get to the basics. There's all kinds of fun things to derive once we get past the basics, but I want to start relatively easy. Let me know if you have feedback, see http://geoserver.org/display/GEOS/User+Collaboration
This is focused on the GeoServer story, but probably more than half the story will be compelling GeoExt tools to make it all happen. But it'll be easier to build those with the server infrastructure. Elsewhere I'll try to write up the whole picture.
thanks for listening.
Chris
--
Chris Holmes
OpenGeo - http://opengeo.org
Expert service straight from the developers.
The central idea is to get at the things SDI initiatives want - people
sharing data. Most of the SDI initiatives I've seen have started with
Metadata (get lots of catalogs), then focused on Users (get people
filling out metadata and searching for it), and finally end up on Data
(it will come after we get lots of metadata and users).
I've viewed GeoServer's approach as the opposite - start with data,
getting real data on the web. The next step is to focus on users, and
after that lots of good metadata will fall out.
I fully concur, after 20 years doing SDI, I've become convinced that
metadata catalogs have become a "displacement activity" and we should
derive metadata during the process of putting data online. Most other
metadata records (in the SDI sphere) should be seen as an admission of
guilt - I havent made this available even though you (the taxpayer)
paid me to collect it.
How? Through users doing what they do on an infrastructure that
supports them. Think youtube - with most viewed, highest rated, most
favorited. Imagine GeoServer letting users create maps of sets of
layers, with different styles and uploaded data. And imagine them
rating and commenting on everyone else's. This produces valuable
metadata about which data is actually used by others.
I think this creates valuable metadata about _some_ aspects of the
data, but not all. FYI I'm currently leading research projects
looking at ways to make richer structured metadata about data content
visible. The "value" metadata discussed here will be vital for
finding data in a large (well-populated) SDI environment. I'd just
love to see the data out there at all!
thanks for listening.
thanks for sharing..
Rob Atkinson
How? Through users doing what they do on an infrastructure that
supports them. Think youtube - with most viewed, highest rated, most
favorited. Imagine GeoServer letting users create maps of sets of
layers, with different styles and uploaded data. And imagine them
rating and commenting on everyone else’s. This produces valuable
metadata about which data is actually used by others.
I think this creates valuable metadata about some aspects of the
data, but not all. FYI I’m currently leading research projects
looking at ways to make richer structured metadata about data content
visible. The “value” metadata discussed here will be vital for
finding data in a large (well-populated) SDI environment. I’d just
love to see the data out there at all!
For sure. I tend to discount traditional metadata a bit, in order to emphasize that there’s lots of kinds of valuable data about data, and lots of it we can derive and encourage. But I’d love to see wiki type editing and just nicer more intuitive editors of more structured metadata, along with much nicer tools to collect metadata when data is being gathered. And of course if you can get a top down requirement for people to make good metadata - attach it to funding, etc., then that’s great too. The more metadata the better, but lack of good metadata should not prevent anyone from putting real data out there.
Chris
thanks for listening.
thanks for sharing…
Rob Atkinson
Open Source Business Conference (OSBC), March 24-25, 2009, San Francisco, CA
-OSBC tackles the biggest issue in open source: Open Sourcing the Enterprise
-Strategies to boost innovation and cut costs with open source participation
-Receive a $600 discount off the registration fee with the source code: SFAD
http://p.sf.net/sfu/XcvMzF8H
Geoserver-devel mailing list
Geoserver-devel@anonymised.comsts.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/geoserver-devel
For sure. I tend to discount traditional metadata a bit, in order to
emphasize that there's lots of kinds of valuable data about data, and lots
of it we can derive and encourage. But I'd love to see wiki type editing
and just nicer more intuitive editors of more structured metadata, along
with much nicer tools to collect metadata when data is being gathered. And
of course if you can get a top down requirement for people to make good
metadata - attach it to funding, etc., then that's great too. The more
metadata the better, but lack of good metadata should not prevent anyone
from putting real data out there.
Would totally agree with this. "Good" is very subjective, so get the
data out while you work out what is good in different contexts!
Rob
Chris
>
> thanks for listening.
>
thanks for sharing..
Rob Atkinson
------------------------------------------------------------------------------
Open Source Business Conference (OSBC), March 24-25, 2009, San Francisco,
CA
-OSBC tackles the biggest issue in open source: Open Sourcing the
Enterprise
-Strategies to boost innovation and cut costs with open source
participation
-Receive a $600 discount off the registration fee with the source code:
SFAD
http://p.sf.net/sfu/XcvMzF8H
_______________________________________________
Geoserver-devel mailing list
Geoserver-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/geoserver-devel
wrote:
Interesting ideas.
Consider whether to only allow users to add data (such as uploading to YouTube), or whether people can interact in the data (YouTube members can comment on videos by others) or interact on the map (letting people modify objects which already exist).
Yeah, commenting and rating are clearly on my roadmap for this. As for interacting with the map, I'd like to at least see users be able to open their data to others for modification.
Should there be a full modification history, so people can see the changes, undo changes, and bookmark a specific version? For that matter, what is a version -- the entire globe at one point in time, one specific dataset, or the datasets which one person has open? One can create a map by choosing parts of existing data, so can such combinations be saved as an object which others can interact with?
Yeah, full modification is something we're interested in, and have already built a pretty decent backend, see http://geoserver.org/display/GEOS/Versioning+WFS
Though a crazy distributed modification history would be awesome, the way I'd start (and am starting), is to tie modification history to an individual data layer. A user has ownership over that, and can choose to open it to others. So it'd be like open source software, with a central repository that is the definitive version. Eventually would support branches, but there would still be the notion of the dataset. So in composing a map you'd pick the data layers, and if you wanted to change those you'd have to get your contributions back to the original, or else go through a major act of 'forking', establishing your own definitive version.
But no matter what I'd like to see a user able to upload data and then encourage others to edit it, and versioning changes should help a lot for people to feel better about opening up their data.
best regards,
Chris
--
Chris Holmes
OpenGeo - http://opengeo.org
Expert service straight from the developers.
On Tue, Feb 17, 2009 at 2:23 PM, Chris Holmes <cholmes@anonymised.com> wrote:
wrote:
Interesting ideas.
Consider whether to only allow users to add data (such as uploading to
YouTube), or whether people can interact in the data (YouTube members
can comment on videos by others) or interact on the map (letting people
modify objects which already exist).
Yeah, commenting and rating are clearly on my roadmap for this. As for
interacting with the map, I'd like to at least see users be able to open
their data to others for modification.
Interestingly one of my MGIS students pitched this as a course project
over the weekend - I'm encouraging her to think about it as a
dissertation project (it's well beyond the scope of my course). Her
current plan is based on local/micro brew pubs that people could add
to and rank through a map interface. I could see it working with a
WFS(-V) and some javascript + openlayers.
Ian