[datahub-discuss] Datahubio outdated datasets

Mon Aug 18 20:50:44 UTC 2014

Thanks for outlining the tradeoffs, Ivan.

In the case of http://datahub.io/dataset/riese, I would certainly hope it is dropped from the lodcloud group.
That would “deprecate” it quite a bit.

Adding a tag “discontinued” would be a lightweight option to establish the distinction you identify.
Since all of the lodcloud group datasets were also tagged “lod”, then we’d still have the “discontinued” “lod” datasets with a combination of two tags.
I don’t think that the “tag space size” and capitalization concerns outweigh this option.
“discontinued” I hope is self-explanatory enough to not need a description.

I’m still confused between the “organizations replaced groups” features of CKAN, so I’d recommend avoiding those options.
Also, a group/organization involves more overhead than the “dead” datasets deserve, I would think.

-Tim

On Aug 18, 2014, at 3:56 PM, Ivan Ermilov <earthquakesan at gmail.com> wrote:

> Tim,
> 
> I thought about this. In that case it should be clearly possibly to see the border between the historical metadata (which point to non-existent datasets) and the actual metadata. The three possible ways at the moment are:
> 1. Special tag 
>    + easy to add (everyone can add tags to his/her datasets)
>    - tag space is quite big http://datahub.io/api/3/action/tag_list
>    - tag descriptions are somewhat lacking http://datahub.io/api/3/action/tag_show?id=access-commercial (is it possible to have some more metadata for tags?)
>    - "Historical" and "historical" are two different tags (i.e. case-sensitive) - users will have several tags in the end
> 
> 2. Special group
>    + no groups at the moment http://datahub.io/api/3/action/group_list 
>    + group can have a description http://demo.ckan.org/api/3/action/group_show?id=data-explorer
>   + user can only choose from existing ones it seems (can't test it, because no groups exist on datahubio)
>    - I couldn't add a group to my own dataset
>    - was it deprecated in favor of "organizations"?
> 
> 3. Key/value pair (custom field)
>   - No description for the field
>   - Case-sensitive and need typing from dataset maintainer.
>   - I created a key/value pair for my dataset, now can't delete it. Looks like bug =\
> 
> Therefore, I would go for creating a special group (organization?) for the historical datasets, if we want to keep them.
> 
> Kind regards,
> Ivan.
> 
> 
> 2014-08-18 18:06 GMT+02:00 Timothy Lebo <lebot at rpi.edu>:
> Richard, Ivan,
> 
> Is it appropriate to *delete* these?
> Perhaps just tag and remove from groups as appropriate?
> It would be nice to have some historical perspective on outdated LOD datasets.
> 
> Regards,
> Tim
> 
> On Aug 18, 2014, at 11:59 AM, Richard Cyganiak <richard at cyganiak.de> wrote:
> 
>> Ivan,
>> 
>> I can help with deleting those datasets.
>> 
>> Could you send a list of the dataset URLs to the mailing list, along with a statement of justification for each (e.g., “project page states that the project is discontinued”)?
>> 
>> Best,
>> Richard
>> 
>> 
>> On 18 Aug 2014, at 15:58, Ivan Ermilov <earthquakesan at gmail.com> wrote:
>> 
>>> Hi everyone,
>>> 
>>> I'm working on LODStats portal and checking the dataset availability and metadata consistency. Some datasets are not available due to server errors (basically hosting for dumps is down), they may appear online again. But some of them are really dead, for instance this one:
>>> http://datahub.io/dataset/riese (if you try to navigate to the resource, you will see that the project is discontinued). 
>>> 
>>> My question is if it's possible to delete this dataset and some other in the future (I will make a list of fixes and can report it to the mailing list). If deletion is not an option, than there should be the way to fix it (maybe someone will volunteer to contact maintainers and then host RDF).
>>> 
>>> Kind regards,
>>> Ivan Ermilov.
>>> _______________________________________________
>>> datahub-discuss mailing list
>>> datahub-discuss at lists.okfn.org
>>> https://lists.okfn.org/mailman/listinfo/datahub-discuss
>> 
>> _______________________________________________
>> datahub-discuss mailing list
>> datahub-discuss at lists.okfn.org
>> https://lists.okfn.org/mailman/listinfo/datahub-discuss
>> 
> 
> Timothy Lebo
> lebot at rpi.edu
> https://impactstory.org/TimothyLebo
> 
> 
> 
> 

Timothy Lebo
lebot at rpi.edu
https://impactstory.org/TimothyLebo

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.okfn.org/pipermail/datahub-discuss/attachments/20140818/deafd7d1/attachment-0003.html>