PR18 - Add new property to Dataset to express the validity of the data



Add new property to Dataset to express the validity of the data.

Proposed solution

Add optional property dct:valid to Dataset to specific validity before updating data.






Tue, 24/03/2015 - 19:59

Proposed resolution: add dct:valid to Dataset

Wed, 08/04/2015 - 07:35

What is the definition of 'validity'? Also dct:valid is a date so it expresses the timestamp at which the dataset was valid.

Also what is the objective of this property.  Is it a catalog management property that is intended to support a machine ingestion process indicating it has passed the ingestion quality control then I do not see immediate added value.

If it indicates that the metadata has been validated and check according to the DCAT-AP specification, then we should clearly specify to which extend. Mandatory properties only? Recommended properties? Also the ranges (note that many of the Member States use different vocabularies in order to better address their need)?

So prior accepting this property I propose we define what validity means.

Thu, 09/04/2015 - 13:50

I agree, we need the use case explained by Brecht, as it's not clear.

Thu, 09/04/2015 - 14:31

The issue was originally raised by Sonia Castro of I have asked her to clarify.

Thu, 09/04/2015 - 20:04

We have a use case at the Publications Office about the "Who is Who" datasets.

Those datasets describe the directory and all the end of job and nomination during a period of time. Each dataset covers a period of time which must be indicated in the dataset description. If someone needs to retrieve all the nominations in the european institutions for a year, he should select a complete collection of datasets covering the complete period.

I would also think of statistical datasets which cover a period of time.



Fri, 10/04/2015 - 17:44

Does validity relate to quality? If so the W3C DWBP WG Data Quality Vocabulary (work in progress) may be a relevant reference.

Fri, 10/04/2015 - 20:01

Deirdre, we don't know what the intention was of the people who requested this. We're trying to get more information. Until we get this, maybe we should not try to speculate?

Thu, 16/04/2015 - 11:15

Here is some additional information provided by Jean Delahousse:


The title of the demand is not right, the subject is not the "Validaty of the dataset", but the "temporal coverage of the dataset".  For some datasets is it very important to give an information about the temporal coverage of the data, especially if there is a set of datasets, each covering a consecutive temporal period. An example is a directory of governmental institutions with the history of the changes during a period of time, like the EU WhoIsWho. For each dataset, published on a monthly period, there is a need to describe the temporal period covered by the Dataset.  We propose to add as an optional property dct:temporal to Dataset. As dct:spatial we already use, it is a subproperty of dct:coverage.

Fri, 17/04/2015 - 10:07

The title of the issue is correct: this was what the submission asked for. The issue of temporal coverage is different, and already covered: dct:temporal was already an optional property for Dataset in the previous version of the profile.

Mon, 27/04/2015 - 16:50

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