Question on the usage of MimeType and MediaType for "zipped/container" distributions

17/06/2016

In the current German OGD 2.0 standardisation phase

https://joinup.ec.europa.eu/discussion/format

we have to be interoperable with existing CKAN installations and with DCAT-AP.

 

Currently I cannot find official information about how to encode a ZIP distribution with DCAT-AP. 

Strictly following the information given in the standard, I would put "application/zip" and the checksum of the zip-file.

 

Is this the right way to do or is there a way to handle inner_mimetypes?

How would you map the usefull mimetype_inner in CKAN to DCAT-AP 1.1.?

Component

Documentation

Category

support

Comments

Wed, 22/06/2016 - 15:39

There is no way currently in the base DCAT Recommendation to distinguish 'outer'-'inner' formats; only the format/mediaType of the file is supported, and DCAT-AP did not extend DCAT on this aspect.

I can see that the issue may be even more complex than inner-outer formats. What happens if the distribution is a ZIP file that contains several files with different file formats (e.g. a ZIP file that contains an xls file, a ods file, a cvs file) or even a ZIP file that contains several ZIP files that etc. etc.)

This may be something for further discussion at W3C if they were to consider revising the base Recommendation. Also in other ways, the modeling of Distributions in DCAT may be (too) simplistic.

Currently, the correct way to do this is indeed to indicate media type application/zip with no way to tell what is inside the ZIP file until you open it.

Fri, 24/06/2016 - 20:37

This issue (and others) can be discussed in a workshop in Amsterdam 30 Nov - 1 Dec 2016:

The outcome may be a new W3C Working Group chartered to extend DCAT and determine how human and machine-readable metadata profiles are defined and made discoverable.

Thu, 28/07/2016 - 13:55

A possibility would be to encourage the use of media type suffixes.

So a zip file with csv inside would be indicated by application/csv+zip

 

See https://tools.ietf.org/html/rfc6839 section 3.6.

Thu, 28/07/2016 - 15:44

Thanks Matthias, that is a useful reference.

Tue, 06/09/2016 - 16:11

The content of this field is kept private and will not be shown publicly.