[ckan-discuss] Why is the bibbase package deleted?

[Rufus Pollock]

On 25 September 2010 20:45, Richard Cyganiak <richard at cyganiak.de> wrote:
> On 25 Sep 2010, at 13:04, Rufus Pollock wrote:
>>
>> On 25 September 2010 12:22, Ed Summers <ehs at pobox.com> wrote:
>>>
>>> Thanks for the responses Rufus and Richard. I noticed that my little
>>> script that updates the graph data started getting a 403 for the
>>> (new?) bibbase package?
>>>
>>>   http://ckan.net/api/rest/package/bibbase
>>
>> If this package has been deleted you'll get a 403.
>
> Well now this is somewhat frustrating.
>
> This was a valid package containing high-quality metadata.
>
> Why has this package been deleted? By whom? Could someone with the required
> privileges (what privileges are required and who has them?) please check
> this?

The package state has been set back to 'active' and the package is
available again.

Looking at the diffs in the last commit someone changed the state from
'active' to the empty state ('null'):

<http://ckan.net/revision/diff/bibbase?diff=702bb0a3-03b7-49ac-87ad-e489c414962f&oldid=5447842d-b6ed-41d9-9cfd-8bb73b85c409>

In some sense this is a bug in that we should restrict the state to
'active' or 'deleted' (and therefore reject requests via api or
otherwise that attempt to set it to something different).

> Is this maybe another instance of this one?
> http://knowledgeforge.net/ckan/trac/ticket/460

Yes, it is another instance of this. I have also commented on that
ticket to make one slight correction: package owners not just
sysadmins are entitled to change the state of packages (so they can
e.g. delete the package).

>> This is a known 'feature'. The logic is that someone with higher
>> privileges (e.g. sysadmin) could look at this package. There has already
>> been suggestion that from user perspective a 404 would make more sense.
>
> It appears that the web front-end was implemented using the same logic: When
> I go to http://ckan.net/package/bibbase, then it redirects me to my user
> page (I am already logged in). I suppose the idea is that a user might
> possibly have another set of credentials with different permissions?

Yes.

> No, this behaviour is not helpful.

I agree that if you are already logged in redirecting you isn't
useful. But what about if you aren't logged in? Should you get a 404
(since package deleted) or redirected to login page as you may have
permission to edit?

Rufus

> Richard
>
>
>
>>
>>> I am skipping over it for the time being, but I was wondering if there
>>> might be something wrong, or if it is in some embargoed space. I can't
>>> seem to find it in CKAN proper, but it is in the lodcloud group:
>>>
>>>   http://ckan.net/api/rest/group/lodcloud
>>
>> If it is indeed deleted then this is a 'bug' in that it should not
>> show up there.
>>
>>> Oh, and thanks for the sunburst example. I've been fishing around for
>>> examples of how to manipulate the visualization with external
>>> controls.
>>
>> Not at all. I'm also playing around a lot with hierarchical data but
>> that is off-topic!
>>
>> Rufus
>>
>>> //Ed
>>>
>>> On Sat, Sep 25, 2010 at 6:47 AM, Rufus Pollock <rufus.pollock at okfn.org>
>>> wrote:
>>>>
>>>> On 23 September 2010 06:18, Ed Summers <ehs at pobox.com> wrote:
>>>>>
>>>>> It was awesome to see that the new LOD Cloud leverages data in CKAN. I
>>>>> really like the idea of driving more people to documenting Linked Data
>>>>> packages on CKAN in order to get their content into the LOD Cloud.
>>>>>
>>>>> I have been doing a bit of playing around with the Protovis JavaScript
>>>>> library recently to visualize owl:sameAs links in the Billion Triple
>>>>> Challenge dataset [1]. It's pretty easy to use, at least in the
>>>>
>>>> That's fantastic. I've been *really* impressed by protovis and that's
>>>> a really nice usage of their force directed layout!
>>>>
>>>> (I've been using protovis for where does my money go sutff e.g.:
>>>> <http://www.wheredoesmymoneygo.org/wdmmg-js/src/sunburst.html>)
>>>>
>>>>> simplistic way I have been playing around with it. I thought it would
>>>>> be fun to use the same approach to graphing the LOD Cloud using the
>>>>> CKAN API. So I went and did it [2]. The fact that it only took an hour
>>>>> or so was all due to the way you have leveraged CKAN, and CKAN's
>>>>> lovely REST API.
>>>>
>>>> Thanks :)
>>>>
>>>>> While processing the data I think I noticed 50 packages in the
>>>>> lodcloud group which referenced other CKAN packages
>>>>> (package['extras']['links:*'] in the json) that were not also in the
>>>>> lodcloud group. I didn't actually dig in to see if they are definitely
>>>>> on CKAN, but at least a few of them already were. I haven't been
>>>>> following the earlier discussion too closely, to know if this is
>>>>> already a known issue. I'd be willing to help add the packages to the
>>>>> lodcloud group if you are looking for an extra set of hands.
>>>>
>>>> Richard's answered on this so I won't add anything (plus he's the
>>>> expert!).
>>>>
>>>> [...]
>>>>
>>>> Regards,
>>>>
>>>> rufus
>>>>
>>>
>>
>>
>>
>> --
>> Open Knowledge Foundation
>> Promoting Open Knowledge in a Digital Age
>> http://www.okfn.org/ - http://blog.okfn.org/
>>
>> _______________________________________________
>> ckan-discuss mailing list
>> ckan-discuss at lists.okfn.org
>> http://lists.okfn.org/mailman/listinfo/ckan-discuss
>
>



-- 
Open Knowledge Foundation
Promoting Open Knowledge in a Digital Age
http://www.okfn.org/ - http://blog.okfn.org/