These are chat archives for frictionlessdata/chat

22nd
Mar 2016
Daniel Fowler
@danfowler
Mar 22 2016 07:36 UTC
Thanks @augusto-herrmann :)
Daniel Fowler
@danfowler
Mar 22 2016 07:56 UTC
@rgrp We now have Rich Field Types in JTS (http://dataprotocols.org/json-table-schema/#rich-field-types). Maybe I’m not fully understanding, but, based on the example, how do I know which property of the Country RDF class (e.g. “name”, “geo”, “url") is being mapped to a given column in the CSV? cc: @amercader
Paul Walsh
@pwalsh
Mar 22 2016 10:13 UTC
@danfowler about above, you don’t. the RDF class just declares it is a thing of this type.
Daniel Fowler
@danfowler
Mar 22 2016 13:59 UTC
@pwalsh OK. I guess I was wondering what the use case is
EarlButterworth
@EarlButterworth
Mar 22 2016 21:07 UTC
Hello. I work in a government agency supporting human and social services. In our agency, like in several others, the preparation and publishing of open data is the responsibility of the data custodians. We're a Windows based organisation, and we don't have python skills.
My question is ; How do you train a program officer to understand and define a JSON table schema? I understand the benefits of what is being proposed, but the solution isn't useable by the average data custodian.
EarlButterworth
@EarlButterworth
Mar 22 2016 21:22 UTC
If I recommended Frictionless Data Packages to our data custodians the coefficient of friction would increase significantly, as it is a "pure" technology response. Where is the JSON schema generator, and the end user manual? Where is the CKAN UX extension which is data custodian friendly?
EarlButterworth
@EarlButterworth
Mar 22 2016 21:53 UTC
I also have a concern that every data package has a file with the name datapackage.json will result in confusion when dealing with multiple data packages. If a consumer of open data downloads and unpacks several data packages into the one folder, how do they manage all the different schema definition files? Would including a document object identifier, or an alternative global object identifier into the JSON schema and for the data package reduce the confusion partially?