Re: suggestions on dcat design (was Re: [dcat] Tomorrow's dcat Agenda)

Hi Li,

Thanks for the comments! Mine are inline.

On 23 Apr 2010, at 10:28, Li Ding wrote:
> I have some comments about the DCAT vocabulary.
> 1. modular vocabulary with minimal core.
> 2. choice of term. I'm worrying that some terms in dcat are not well
> defined and may cause confusion.  e.g.
>   dcat:dataQuality - the range could be really wild,

Yes, dcat:dataQuality (and many other terms) needs documentation that  
provides clear guidelines on what kinds of values should be used. This  
is on the to-do list.

>   dcat:keyword  - why not use dct:subject  or skos concepts

Several of the surveyed catalogs use both keywords and a taxonomy of  
themes/topics. With dct:subject, it would not be clear wether the  
value is a keyword or a topic from the taxonomy. Hence we added two  
subproperties of dct:subject:

- dcat:keyword for keywords/tags,
- dcat:theme for a link to topic from the catalog's theme taxonomy

Again, we need to make this clear in the documentation.

> 3. best practices for actual usage.
>    Your demo [1] is interesting but I bet a normal web user could get
> confused on that.

The point of the demo is that we have integrated catalog *data*. The  
UI is just the off-the-shelf RDF visualization that comes along with  
the D2R Server software. The UI enables navigation of the produced RDF  
data, but it is not intended as something that normal web users should  
ever use or even look at.

> more details about my comments can be found at  [2][3].

The design principles in [2] are helpful, it's a useful checklist that  
we should keep in mind for the dcat terms.

Best,
Richard



>
> [1] http://lab.linkeddata.deri.ie/govcat/directory/CategoryScheme
> [2] http://tw.rpi.edu/weblog/2010/04/23/three-principles-for-building-government-dataset-catalog-vocabulary/
> [3] http://data-gov.tw.rpi.edu/wiki/TWC_Data-gov_Vocabulary_Proposal
>
> best,
> Li

Received on Monday, 26 April 2010 16:20:13 UTC