Re: xsd:date in DCAT

Hi,

Thanks for the feedback on DCAT!
I want to let you know that I have changed the text in DCAT Spec as suggested by Phil
see for example: https://dvcs.w3.org/hg/gld/raw-file/default/dcat/index.html#Property:dataset_release_date

Best regards,
Fadi
On 5 Apr 2013, at 08:33, Stasinos Konstantopoulos <konstant@iit.demokritos.gr> wrote:

> Dear Phil, GLDers,
> 
> kind reminder of some thoughts I had posted earlier, regarding ways to
> compromise between current (mal)practice and accurate ways to express "I
> don't know exactly when".
> 
> http://lists.w3.org/Archives/Public/public-gld-wg/2012Jan/0094.html
> 
> best,
> s
> 
> 
> On Thu Apr  4 10:16:49 2013 Phil Archer said:
> 
>> Having noted this afternoon's agenda item on ADMS I'm working on
>> that document right now which is causing me to look at DCAT more
>> carefully than I have of late - which is my excuse for just noticing
>> something I should have seen before. I ask that the WG treats this
>> as a last call comment.
>> 
>> In the text related to the use of dcterms:issued [1] we say:
>> 
>> "rdfs:Literal typed as xsd:date. The date is encoded as a literal in
>> "YYYY-MM-DD" form (ISO 8601 Date and Time Formats). If the specific
>> day or month are not known, then 01 should be specified."
>> 
>> I remember raising this at our previous f2f last year as I find it
>> objectionable that we actively encourage sloppy practice and
>> inaccurate data.
>> 
>> If I know that something was issued in March 2013 then I can write
>> 
>> dcterms:issued "2013-03"^^xsd:gYearMonth
>> 
>> That conveys exactly what I mean - that the thing was issued at some
>> point between 2013-03-01T00:00:00 and 2013-03-31T23:59:59. But I
>> don't know when. It is accurate, if not precise.
>> 
>> But DCAT says we shouldn't do this. We should render it as
>> 2013-03-01 which means that the thing was issued sometime in the 24
>> hour period known as 1st March. That may be wrong by as much as 30
>> days and gives an entirely bogus impression of accuracy.
>> 
>> I suspect that the reason for this is that catalogues habitually
>> don't understand xsd:gYearMonth. If that's the case then that's
>> application-specific and a profile may wish to make it clear that
>> dates must be xsd:date only, even if that means it will create
>> inaccuracies where none exist in the original data, but IMHO this
>> sloppiness should not be included in the DCAT spec. Therefore I
>> suggest that the text for this property says:
>> 
>> "rdfs:Literal using the relevant ISO 8601 Date and Time compliant
>> string and typed using the appropriate XML Schema datatype
>> [[XMLSCHEMA-2]]"
>> 
>> 
>> 
>> [1] http://www.w3.org/TR/vocab-dcat/#Property:catalog_release_date
>> 
>> 
>> 
>> -- 
>> 
>> 
>> Phil Archer
>> W3C eGovernment
>> http://www.w3.org/egov/
>> 
>> http://philarcher.org
>> +44 (0)7887 767755
>> @philarcher1
> 

Received on Thursday, 30 May 2013 13:28:07 UTC