[ckan-discuss] Purging revisions?

Rufus Pollock rufus.pollock at okfn.org
Wed Apr 7 19:23:47 BST 2010

On 7 April 2010 18:16, David Read <david.read at okfn.org> wrote:
> Jonathan,
> Thanks for this. Yes, we're hesitant about allowing general users
> delete powers - not just that they may delete perfectly good packages,
> but we'd have to write some UI to view 'deleted' packages and
> 'undelete' them. We have the compromise that administrators simply
> view them all have this additional field in the package edit form.

Just to be clear: admins already have the power to purge revisions
(and hence packages -- just purge all revisions related to that
package). We could also add ability to purge packages directly but
personally, in my experience with spam so far just easier to purge the
relevant revision (usually there is only one revision for the spam
package -- the revision creating it ...)

> I'm keen to for general users to nominate articles for deletion as you
> suggest. Perhaps the easiest way to do this is to ask users to email
> the admins for the instance? I'll create a ticket for this:
> http://knowledgeforge.net/ckan/trac/ticket/283 so feel free to
> continue the discussion here or on the ticket.

Can I *strongly* suggest we just use the existing
perfectly-good-system for flagging stuff called "tags" :)

I suggest we have agree in the community a standard set of "meta" tags
for this kind of stuff. E.g. i'm already using the "duplicate" tag for
marking duplicates (I also add in notes the link to duplicate package
but that's optional). So I suggest we:

a) Create "reserved" tag prefix "meta"
b) Create following specific tags (suggestions/comments welcome):

  * meta.duplicate - duplicate of another package. If possible
indicate in notes or an extras field (to be decided) what it is
duplicate of
  * meta.spam

Editors can then just visit http://ckan.net/tag/read/meta.spam and
work through list of packages there.

I also have ideas for a "todo" set of tags but that should be reserved
for another mail ...


More information about the ckan-discuss mailing list