Feedback F23 PIN (issue 8, TM): II.2.7.2 and II.2.7.3 a time has been added to the Start and End date

[valexande]

II.2.7.2 and II.2.7.3 a time has been added to the Start and End date

[valexande]

We have removed the time from this property

[csnyulas]

Reopening this issue, as the solution implemented is not as straightforward as it appears to be.

The reason for concatenating the time part 'T00:00:00' after the date, in the first place, was to make the generated value compatible with EPO v3.1.0, where both the start (epo:hasBeginning) and end (epo:hasEnd) of a period is of type xsd:dateTime (and not xsd:date). Perhaps we could have done better by attaching 'T23:59:59' to the end date of the period, instead of 'T00:00:00'.

The problem that @muricna identified is that we can't just add an arbitrary time to the given dates, in order to define a proper moment of the period's start and end, if this is not specified in the input data.

The problem (and te reason for reopening the issue) is that we cannot accurately represent the input according to ePO 3.1.0. We will either:

1. need to add extra information (for the time part, to be able to type it xsd:dateTime)
2. receive SHACL shape violations, if we type the given value as ^^xsd:dateTime (as a date value of form "YYYY-MM-DD" is not a valid xsd:dateTime)
3. receive SHACL shape violations, if we type the given value as ^^xsd:date (as xsd:date is not a subtype of xsd:dateTime)

In our initial implementation we chose option 1. In the fix we chose option 3. As said, neither is perfect.

Questions: Which solution would be the best to go with? If 1, are there more appropriate values to concatenate than the ones that we had originally? Are there other, better, solutions than these 3?

[muricna]

A solution has to be found whereby when a time is provided in the xml it is reflected in the rdf. When there is no time in the xml no time should be found in the rdf.

[csnyulas]

Linking this to #269 which provides additional information on the problem.