[ckan-discuss] Adding multiple OpenlyLocal URLs/datasets to CKAN?

[Glen Barnes]

Ah right - I assumed from reading below that your database aggregated data from multiple database as opposed to its own unique dataset. 

Maybe there does need to be some kind of mark as to whether it is derivative, a new dataset, or a source dataset. I'm not sure how best to handle it but 10 companies all supplying various versions of the same dataset could get messy in the catalogue. 

Glen 
 
On 23/07/2010, at 10:38 PM, CountCulture wrote:

> Not sure that's quite true. As far as I'm aware there is no database of local authority pension funds and no database of police forces that maps to councils (at least Home Office and Min of Justice weren't able to provide one and didn't seem to think it existed). 
> 
> Similarly there is often no database of council membership as such, only a list of memberships, which may only exist on paper or as hand-written HTML on a member's page. 
> 
> Of course you could say that all these are abstract datasets, from which the databases are derived and where we can link to the canonical source, though that's difficult if it's been assembled by hand or from a spreadsheet you were emailed that you've had to heavily edit and supplement with other info. 
> 
> I guess the point is it's probably useful to list if it's currently the only known source of the information as data, whether or not it is the owner of the data, which in any case is a tricky and debatable issue. 
> Chris
> 
> -----------
> OpenlyLocal :: Making Local Government More Transparent 
> http://openlylocal.com
> Blog: http://countculture.wordpress.com
> Twitter: http://twitter.com/CountCulture
> 
> On 23 Jul 2010, at 11:17, Glen Barnes <glen at opengovt.org.nz> wrote:
> 
>> This brings up an interesting point around what goes into the database. Technically OpenlyLocal is not a dataset source but a derivative. Do you put derivatives into the database or only link to the canonical source of information? For New Zealand what I wanted to do was only list the source datasets but on each dataset page have the ability to link to derivative works and re-processed data. My thinking is that you can only trust the source data to be 'right' and 'owned' by the government. We know where to feedback issues if this is the case. For any third party data you have to work out first who caused the problem.
>> 
>> Anyone have any thoughts around this?
>> 
>> Glen
>> On 18/07/2010, at 12:17 AM, Chris Taggart wrote:
>> 
>>> I was thinking not of one for each police force, but for the collection of all police forces e.g. http://OpenlyLocal.com/police_forces (accesible as XML and JSON too)
>>> 
>>> On Sat, Jul 17, 2010 at 1:13 PM, Rufus Pollock <rufus.pollock at okfn.org> wrote:
>>> cc'ing ckan-discuss as this question is of general interest!
>>> 
>>> On 17 July 2010 12:58, Chris Taggart <countculture at gmail.com> wrote:
>>> > Someone who I followed just asked on twitter for list of police forces
>>> > available as data. I pointed him to the OpenlyLocal list, but it occurred to
>>> > me it's not obvious that we have all these datasets (also police authorities
>>> > and local auth pension funds)
>>> > What do you think about me adding separate entries to CKAN? I'm easy either
>>> > way, and don't want it to appear spammy, but may make it easier for some
>>> > people.
>>> 
>>> This is a standard question that has come quite a lot :) (and comes up
>>> with data.gov.uk too). The simple answer at the moment is that I don't
>>> think it is worth creating individual datasets/packages for *each*
>>> police force/pension fund on CKAN.
>>> 
>>> Rather I'd suggest creating a single dataset/package on CKAN and if
>>> possible have a single file for download that has all the relevant
>>> data -- you can always add notes explaining how to get individual data
>>> items via e.g. the API. (You may also want to add Resource item for
>>> the api using the api/... format we are standardizing on [1]).
>>> 
>>> [1]: <http://wiki.okfn.org/ckan/doc/package#FormatStrings>
>>> 
>>> If you don't have a single file for the whole lot but single
>>> files/urls for given police forces I'd suggest, for the time being,
>>> having a single Resource item linking to one specific example with a
>>> description string making this clear. Again you can then explain in
>>> the notes how to get all of them. This avoids having to create 500
>>> resources (you can do that if you wish but IMO it is not very useful
>>> -- much better to have a single example and move towards having a tar
>>> ball or url for the whole lot).
>>> 
>>> Rufus
>>> 
>>> > --
>>> > -------------------------------------------------------
>>> > OpenlyLocal :: Making Local Government More Transparent
>>> > http://openlylocal.com
>>> > Blog: http://countculture.wordpress.com
>>> > Twitter: http://twitter.com/CountCulture
>>> >
>>> 
>>> 
>>> 
>>> --
>>> Open Knowledge Foundation
>>> Promoting Open Knowledge in a Digital Age
>>> http://www.okfn.org/ - http://blog.okfn.org/
>>> 
>>> 
>>> 
>>> -- 
>>> -------------------------------------------------------
>>> OpenlyLocal :: Making Local Government More Transparent 
>>> http://openlylocal.com
>>> Blog: http://countculture.wordpress.com
>>> Twitter: http://twitter.com/CountCulture
>>> _______________________________________________
>>> ckan-discuss mailing list
>>> ckan-discuss at lists.okfn.org
>>> http://lists.okfn.org/mailman/listinfo/ckan-discuss
>> 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.okfn.org/pipermail/ckan-discuss/attachments/20100724/9d472ec3/attachment.htm>