[okfn-discuss] New name for ckan.net: call for proposals

Sun Nov 21 20:12:23 UTC 2010

Jessy Kate wrote:
> hmm, well with all the descriptions so far i'm actually still not really
> clear what "it" is :).
> 
> maybe it's worth describing why someone would use it? (and why they would
> use it over other data repositories?)
> 
> ie, would i used it to store arbitrary datasets? or would i use it to link
> to other datasets? could it be compared to, say, infochimps? they don't
> necessarily host all the data listed there. the unifying thing is that it's
> a place to index and store information about what other data is out there
> (and, of course, to optionally charge for it).
> 
> is this like infochimps for free and open data?
> 

FWIW, the original requirements for CKAN are recorded here:
http://desire.appropriatesoftware.net/products/30/

"Knowledge and data will increasingly be provided in packaged form to be 
re-used combined in a manner similar to software. Need a registry to 
keep track of what open knowledge projects and packages exist."

The "Scope of the System" shows that CKAN was originally intended as a 
central register of metadata.  The main use case is recorded here:
http://desire.appropriatesoftware.net/stories/84/

"Create, read, update, or delete records of open knowledge projects or 
packages"

So the "archive network" part was always a bit of a misnomer. Obviously 
enough, the issue with archiving the data (which Perl scripts mostly 
don't suffer from) is that archiving data sets can require very large 
storage capacity. So it makes sense to have the data decentralised.

Of course the metadata can also be decentralised (that's how the 
Location Information Infrastucture side of data.gov.uk will work) and as 
Rufus points out Google does a pretty good job of indexing stuff. 
Nevertheless, there appears to be a value in maintaining an inventory, 
particularly for large organisations, and especially for large public 
organisations like central governments.

J.

> jessy
> 