[datacatalogs] RFC: A proposal for Simple Data Catalogs Interoperatibility based on DCAT

Tue Nov 12 17:27:01 UTC 2013

Hi all,

With the recent raise in Open Data Portals around the world there has
been a growing need to share metadata between them. Sites like
PublicData.eu aggregate metadata from several other catalogs running
on different software and with different metadata profiles.
Maintainers from instances willing to incorporate their metadata into
other catalogs need a clear and simple approach to publish their
metadata so it can be easily shared.

Note the emphasis on *simple*. On our experience resource and time
constraints mean that often publishers can not be asked to develop an
advanced API on top of their existing catalog, even if this means that
client applications have a more limited access to it.

We propose a simple approach based on dumps of metadata (similar to
Project Open Data's data.json file approach) plus optional parameters
for paginating and filtering on large instances. We believe that this
provides an easy way for getting metadata online and is simple enough
so it can be expanded and improved in the future.

We've been working on a CKAN implementation for this proposal that has
already been successfully tested on the Publicdata.eu staging site
with test data kindly provided by our friends at the Spanish and
Slovenian data portals.

If this sounds like something you might be interested on, please have
a look at the draft document describing the proposal, including
context, dataset serializations and request description (Note
specially the "Issues to discuss" sections):

https://docs.google.com/a/okfn.org/document/d/1JFgk-U7so4V0aecut53JS2GPtI2tgFMShUvLoINyKQY/edit#

We hope people will find this useful and look forward to any feedback,
feel free to discuss on this mailing list or adding a comment on the
document.

Best,

Adrià