Re: Missing DCAT property

Martin,

I think a great way to make progress here is to document various scenarios that we'd like to describe in dcat.

Here some fictional examples:

“There's an XML file called itf2007.xml inside a Zip file itf-raw.zip. The Zip file is downloadable from http://example.gov/data/itf-raw.zip. There is a web page with documentation and a link to the zip file at http://example.gov/content/proj/itf/download.html.”

Or:

“There's a RESTful web service to the ITF database. The homepage of the database is http://example.gov/content/proj/itf/index.html. The API documentation is at http://itf.example.gov/doc/api. All API calls start with http://itf.example.com/api?…”

Or:

“The ITF database is available for download as a number of Excel spreadsheets, one per month. They are available as http://example.gov/data/YYYY/itf-MMMM.xls”, where YYYY is the year (e.g., 2007) and MMMM is the month (e.g., “january”). There's a web page with a table that lists all the downloadable files at http://example.gov/content/proj/itf/download.html.”

Once we have a collection of such scenarios, we can figure out and document how to express them all in dcat. This could also form the basis of a “cookbook” for using dcat to model real-world data access scenarios; and should perhaps become part of the use cases document too.

Maybe you could make a start by writing up the scenarios that motivated you to add the access type concepts? Perhaps start with an email to the list; I imagine that the editors can find a place where to put this on the wiki.

All the best,
Richard


On 8 Dec 2011, at 08:12, Martin Alvarez-Espinar wrote:

> Hello Phil,
> 
> Good work on the Working Draft [1]! (also Fadi and John, of course). I
> totally agree with this early version, but I would like to point out
> the need of an additional property for the dcat:Distribution class. It
> would be used to indicate if the access to data refereed by
> dcat:accessURL is either 'direct' or 'indirect'. We discussed it time
> ago, but we didn't modify the draft on the wiki.
> 
> Some examples:
> -> Direct: accessURL points to a WebService, RSS, XLS, or XML, which
> offers the distribution directly.
> -> Indirect: accessURL points to a REST WebService or API
> documentation (how to use it, parameters, etc) | an XML zipped
> 
> We have solved this issue using the property dcterms:type and a couple
> of concepts (indirect-access, direct-access) to set type of each
> distribution.
> 
> [] a dcat:Distribution ;
>   dcat:accessURL "http://.../file.xml"^^xsd:anyURI ;
>   dcterms:type <http://purl.org/ctic/dcat#accessMode-direct> ;
>   ...
> 
> [] a dcat:Distribution ;
>   dcat:accessURL "http://.../file.zip"^^xsd:anyURI ;
>   dcterms:type <http://purl.org/ctic/dcat#accessMode-indirect> ;
>   ...
> 
> I would like to give a hand enriching the draft if you need it. Maybe
> we should provide more implementation examples for each property
> (resources, taxonomies, etc.).
> 
> Best regards,
> 
> Martin
> 
> [1] https://www.w3.org/2011/gld/group/WD-DCAT-20111218.html
> 
> -- 
> Martin Alvarez Espinar
> W3C Spain Office Manager        tel.:+34 984390616
> http://www.w3c.es/Personal/Martin   mlvarez@w3.org
>

Received on Thursday, 8 December 2011 16:21:47 UTC