I think that the scope of the bundle was originally quite easily defined (a rest api for scb) but that it is now to broad for being a single platform artifact. I suggest I finish the improvements around backup I'm working on and the open an issue for splitting up the bundle.
Reto On Tue, Jun 21, 2011 at 12:21 PM, Hasan Hasan <[email protected]> wrote: > when the bundle was created, there was no intention to offer gui over the > web. > over the time people added stuff which now generates a naming problem. > In future we should be careful in choosing name or in placing stuff into a > bundle. > Before releasing zz, changing bundles' name is tolerable to certain > extents. > In this case, I agree to change the name, but I would suggest something > like > > platform.tcaccess.web or platform.graphaccess.web > It indicates a platform service providing access to (and management of) > triple collections (graphs) over the web (jax rs) > > Although tcaccess is more precise than graphaccess, the term tc is not > clear > to most zz-newbies. I am ok with both, but tend to choose graphaccess (as > we > also use for graphnodeprovider, graphprovider) > > cheers > hasan > > On Tue, Jun 21, 2011 at 11:49 AM, Reto Bachmann-Gmuer < > [email protected]> wrote: > > > I'm looking at the rdf.web bundle. I though it would be nice to make the > > backup-functionality available from the menus. > > > > The problem is that the menu functionality is platform specific so it > > shouldn't be in an rdf.*-bundle. This is possibly the reason why the > > sparql-endpoint isn't available via menus. The graph-management page > > however > > is menu-accessible, and a few ssp violate the principle of having no > > rdf->platform dependencies as well. > > > > Shall I rename the artifact to platform.rdf.web or are there other > > suggestions? > > > > Cheers, > > Reto > > > > On Tue, Jun 21, 2011 at 8:53 AM, Hasan Hasan <[email protected]> wrote: > > > > > agree, for sure it is better if we can keep information whether a > > > serialized > > > graph was coming from an mgraph or a graph. > > > and this can be an improvement issue to be solved later after the > > release. > > > if one has time to do it before the release I wouldn't stop him from > > doing > > > it ;) > > > > > > cheers > > > hasan > > > > > > On Tue, Jun 21, 2011 at 8:43 AM, Reto Bachmann-Gmuer <[email protected] > > > >wrote: > > > > > > > In the triplecollections.nt file in a backup-zip all triple > collections > > > are > > > > described to be of type <http://clerezza.org/2009/11/backup#Graph>. > > > > Shouldn't different types be used for graphs and mgraphs? I don't see > > how > > > > restoring an instance is possible from such a backup zip without > > knowing > > > > when to create a graph and whn to create an mgraph. > > > > > > > > Cheers, > > > > Reto > > > > > > > > > >
