- From: Rob Atkinson <rob@metalinkage.com.au>
- Date: Tue, 05 Sep 2017 22:48:52 +0000
- To: kcoyle@kcoyle.net, public-dxwg-wg@w3.org
- Message-ID: <CACfF9Lw9U79_VVF2cvCfnbXwRMcXa7vivbXTkFyaj1zF-qUvXg@mail.gmail.com>
Having been through all the use cases and requirements and looking for commonality, i would suggest that profiles and DCAT are not orthogonal concerns. Given the nature of profiles, and particularly the nature of negotiation by profile and the impossibility of capturing all fine-grained semantics inside DCAT vocabulary itself, perhaps the single most important functional improvement for DCAT is the specification of profiles within DCAT. This means (i expect) canonical ObjectProperties to specify the profile of DCAT being used in the description, the interoperability profiles supported by the dataset and the specific profiles supported by each distribution for both access method and response. IMHO Any large information infrastructure that supports heterogeneous data is broken without the ability to declare and then link to descriptions of data characteristics in this convenient manner. Rob On Wed, 6 Sep 2017 at 01:37 Karen Coyle <kcoyle@kcoyle.net> wrote: > Jaro - > > The filtering works on the html version[1] for me. I selected the > "profile" option and have MANY comments to make on the requirements, but > I wonder if we all shouldn't look first at DCAT, as that is a priority. > > kc > [1] https://w3c.github.io/dxwg/ucr/ > > On 9/4/17 12:37 AM, Jaroslav Pullmann wrote: > > > > Hello Karen, > > > > Rob has merged in the code, you might now experiment with the > prototype. > > I'd suggest to download/clone the UCR page and open in a browser > locally. > > It uses the tags attached to use cases and resolves their related > requirements. > > > > Talk to you later, best regards > > Jaro > > > > > > > > On Saturday, September 2, 2017 19:18 CEST, Karen Coyle < > kcoyle@kcoyle.net> wrote: > > > >> Jaro - yes, that sounds useful, if it can be done. We had noted that the > >> requirements themselves were not tagged so we did not have an easy way > >> to gather them by deliverable. Could you demonstrate what you describe > >> below for the group? > >> > >> Thanks, > >> kc > >> > >> On 9/1/17 2:49 AM, Jaroslav Pullmann wrote: > >>> > >>> Dear Karen, dear all, > >>> > >>>> determine the requirements for the three deliverables > >>> > >>> couldn't we use the tags to filter accordingly? > >>> > >>> There are 3 deliverable tags (content_negotiation, dcat, profile) > >>> with every use case tagged with at least one of them. > >>> > >>> The spec reader might create a view filtered by deliverable and > further > >>> tags. > >>> Only matching use cases and their related requirements were visible: > >>> Tag > UC > Req. > >>> > >>> Does this simple approach fit the task or are there exceptions, where > e.g. > >>> requirements are relevant for a deliverable while none of the related > >>> UCs does? > >>> > >>> Best regards > >>> Jaroslav > >>> > >>> > >>>> Some possible organizing principles are: > >>>> > >>>> - which requirements are for which deliverables? (Yes, some > requirements > >>>> may be valid for more than one deliverable.) > >>>> - what are the logical categories that requirements fall into within > >>>> each deliverable? > >>>> - what is the priority for each requirement? (e.g. absolutely > required; > >>>> required if possible; nice to have but not required) > >>>> > >>>> Following are some examples from previous working groups that the team > >>>> is aware of. If you know of others please reply with them. > >>>> > >>>> - https://www.w3.org/TR/dwbp/#challenges (DWBP challenges) > >>>> - https://www.w3.org/TR/dwbp/#requirements (DWBP requirements) > >>>> - https://www.w3.org/TR/shacl-ucr/#requirements (SHACL requirements) > >>>> > >>>> Please give some time to this and share an ideas with the group. > >>>> > >>>> kc for the team > >>>> > >>> > >> > >> -- > >> Karen Coyle > >> kcoyle@kcoyle.net http://kcoyle.net > >> m: 1-510-435-8234 (Signal) > >> skype: kcoylenet/+1-510-984-3600 <+1%20510-984-3600> > >> > > > > > > > > > > -- > Karen Coyle > kcoyle@kcoyle.net http://kcoyle.net > m: 1-510-435-8234 (Signal) > skype: kcoylenet/+1-510-984-3600 <+1%20510-984-3600> > >
Received on Tuesday, 5 September 2017 22:49:42 UTC