Currently, GeoNetwork indexes the metadata of all Layers in GeoNode and allows all of them to be searched by any user. The current tickets for fine grained Layer permissions call for Layers that are unviewable for a given user to be omitted in search results. AFAIK, GeoNetwork has it's own model of users and permissions which is neither pluggable, nor exposed by any (intentional) API aside from some web forms in the GeoNetwork UI -- so getting GeoNetwork to do this for us might be ugly. We could also post-filter search results in GeoNode -- though I think this is fairly suboptimal too since it complicates batching results and we would have to add a layer over CSW queries as well to correctly support it. Doing either is totally possible, but it seemed worth checking whether we really care about this enough to dig into this now. Would we be happy (for the time being) with search results that list metadata for unviewable layers and indicate that the full data is not available to you? Other ideas on how to accomplish this? - Luke
On 07/19/2010 12:30 PM, Luke Tucker wrote: > Currently, GeoNetwork indexes the metadata of all Layers in GeoNode > and allows all of them to be searched by any user. The current > tickets for fine grained Layer permissions call for Layers that are > unviewable for a given user to be omitted in search results. AFAIK, > GeoNetwork has it's own model of users and permissions which is > neither pluggable, nor exposed by any (intentional) API aside from > some web forms in the GeoNetwork UI -- so getting GeoNetwork to do > this for us might be ugly. We could also post-filter search results > in GeoNode -- though I think this is fairly suboptimal too since it > complicates batching results and we would have to add a layer over CSW > queries as well to correctly support it. > > Doing either is totally possible, but it seemed worth checking whether > we really care about this enough to dig into this now. Would we be > happy (for the time being) with search results that list metadata for > unviewable layers and indicate that the full data is not available to > you? Other ideas on how to accomplish this? > > - Luke Showing metadata for unviewable layers (with some visual annotation that they are un-viewable) is fine. GeoNetwork is going to be a public service on a GeoNode site so papering over things in the Django UI won't actually prevent moderately savvy users from getting to the metadata records anyway. -- David Winslow OpenGeo - http://opengeo.org/
+1 I think this is a great item for the roadmap though. On Mon, Jul 19, 2010 at 12:51 PM, David Winslow <dwinslow@opengeo.org>wrote: > On 07/19/2010 12:30 PM, Luke Tucker wrote: > > Currently, GeoNetwork indexes the metadata of all Layers in GeoNode > > and allows all of them to be searched by any user. The current > > tickets for fine grained Layer permissions call for Layers that are > > unviewable for a given user to be omitted in search results. AFAIK, > > GeoNetwork has it's own model of users and permissions which is > > neither pluggable, nor exposed by any (intentional) API aside from > > some web forms in the GeoNetwork UI -- so getting GeoNetwork to do > > this for us might be ugly. We could also post-filter search results > > in GeoNode -- though I think this is fairly suboptimal too since it > > complicates batching results and we would have to add a layer over CSW > > queries as well to correctly support it. > > > > Doing either is totally possible, but it seemed worth checking whether > > we really care about this enough to dig into this now. Would we be > > happy (for the time being) with search results that list metadata for > > unviewable layers and indicate that the full data is not available to > > you? Other ideas on how to accomplish this? > > > > - Luke > > Showing metadata for unviewable layers (with some visual annotation that > they are un-viewable) is fine. GeoNetwork is going to be a public > service on a GeoNode site so papering over things in the Django UI won't > actually prevent moderately savvy users from getting to the metadata > records anyway. > > -- > David Winslow > OpenGeo - http://opengeo.org/ > -- Sebastian Benthall OpenGeo - http://opengeo.org