[OSGeodata] Geodata through CVS or SVN...

[Chris Holmes]

I've been thinking along these lines quite a lot as well.  Soon I'll get 
to the blog post about it, but basically I think WFS-T may be the 
answer.  It's available as GML, but stored in a real backend format like 
PostGIS.  It already does some GML validation, and GeoServer (which 
already does WFS-T) has a validation engine that lets you define 
arbitrary rules to validate against.

The thing missing in WFS is the concept of a diff and a rollback.  I'm 
hoping to work on this soon (though I have no funding so it waits for 
non-contracted time), to add a version table to WFS.  This would be 
great for WFS in general, since a smart client could just suck down the 
dataset, and easily keep it up to date by just querying the version 
table to see what changed since it last checked.

I've thought a bit about the svn route, somehow backing WFS versioning 
on svn, but didn't come up with any good solution.  Other nice thing 
about storing in a spatial database is you can connect a WMS to it and 
easily visualize the data too.  I'm hoping to find time to work on it in 
the fall, and if we get it going it should be easy enough for a group 
like yours to set up (at least that'd be the goal).

best regards,

Chris


Sampson, David wrote:
> At the Ottawa GRASS Users Group "OGUG" 
> (_http://cemml.carleton.ca:8080/OGUG_) we're wanting to host and share 
> community data gathered by, well, the recreational geographer or 
> enthusiastic geomatics tech. Of course once the OSGEO repository is 
> under way there are opportunities to share.
> 
> I want to find out different ways people have made collaborative 
> workflows for growing data sets.
> 
> I am not a programer but have been exposed to CVS and SVN (my preference 
> right now) for managing source code. Many data types are binary (eg 
> SHP). SVN can deal with Binary but I was thinking something a little 
> more open like GML. Also since GML is just XML text then it can be 
> treated like source code (theory)
> 
> Then I see some scenarios.
> 
> 1. Someone can access SVN directory to get latest version.
> 2. Someone can access the file, add/edit it and then upload a new 
> version. This assumes people won't mess it all up, but we're a small 
> community.
> 
> 3. Someone can acess the file, add new data to a new GML file, upload 
> the new GML and a script will append the new GML to the old GML for a 
> new version (I'm not a sys or server admin).
> 
> I'm interested in what else has been tried or thought of instead of 
> passing around a bunch of files.  I thought maybe an application 
> througha website to a postgis DB, but that's too much for our group to 
> dedicate time to.
> 
> So how can we think of geodata as source code?  What am I missing in my 
> thinking?
> 
> I guess a GML validation might be good.
> 
> How would this system might work for a binary?
> 
> Any ideas?
> 
> Cheers
> 
> 
> !DSPAM:1003,44aeb32092882223018498!

-- 
Chris Holmes
The Open Planning Project
http://topp.openplans.org
-------------- next part --------------
A non-text attachment was scrubbed...
Name: cholmes.vcf
Type: text/x-vcard
Size: 269 bytes
Desc: not available
Url : http://lists.osgeo.org/pipermail/geodata/attachments/20060707/a63c4e40/cholmes.vcf