– DRAFT –
DCAT team 2018-06-07

<trackbot> Sorry, but no Tracker is associated with this channel.

<SimonCox> ?

<trackbot> Sorry, but no Tracker is associated with this channel.

<trackbot> Sorry, but no Tracker is associated with this channel.

<riccardoAlbertoni> hi simon, buona sera a te, here it is morning ;)

confirm agenda

SimonCox: Background to agenda..

PWinstanley: Analysis of national DCAT-Ap is about 90% done..

minutes of last meeting

<SimonCox> https://‌www.w3.org/‌2018/‌05/‌31-dxwgdcat-minutes

<SimonCox> +1

+0 (absent)

<riccardoAlbertoni> +1

<PWinstanley> +1

Analysis of national DCAT-AP profiles

<SimonCox> PWinstanley: Analysis of national DCAT-Ap is about 90% done..

approved minutes of last meeting

approved minutes of last meeting

PWinstanley: Describes what he has done before passing it across to Andrea etc
… In doing this, finding some of typos, case differences etc
… suggests that we could improve what exists editorially

SimonCox: If we had proper guidelines previously then that problem would be much less
… acts as evidence for guidance doument

SimonCox: This should feed into analysis for dcat revision but also profile guidelines

Quality-related information

<SimonCox> https://‌rawgit.com/‌w3c/‌dxwg/‌dcatQualityRelatedInformationIssue57Riccardo/‌dcat/‌index.html#quality-information

<SimonCox> https://‌github.com/‌w3c/‌dxwg/‌pull/‌245

<riccardoAlbertoni> https://‌github.com/‌w3c/‌dxwg/‌issues/‌57

<SimonCox> https://‌github.com/‌w3c/‌dxwg/‌milestone/‌12

riccardoAlbertoni: Pulled together a new milestone for quality related information
… after discussion, made proposal (in github)
… (describes proposal)

SimonCox: Essentially, suggestion is to use existing DQV annotations and predicates as per example in wiki
… if quality information for a dataset is available then it can be used in the catalogue
… Where would this info reside?

riccardoAlbertoni: It could be collected in the dataset, unless we need DCAT to be selfcontained
… connecting the dataset with the dataset with the quality statement is possible but there are challenges. But there are concerns/other issues

<alejandra> Not following the audio, so sorry if you've already mentioned this, but STAT-DCAT-AP uses DQV and associates the quality to the dataset

riccardoAlbertoni: Proposal as it stands could progress with an new issue raised to cover the outstanding questions

<alejandra> this is an example from the STAT-DCAT-AP document

<alejandra> <http://‌qualifications.org/‌QualityCertificate1> a oa:Annotation. <http://‌nationalportal.org/‌data/‌employment> a dcat:Dataset; dqv:hasQualityAnnotation <http://‌qualifications.org/‌QualityCertificate1>. <http://‌qualifications.org/‌QualityCertificate1> a dqv:QualityCertificate; oa:hasBody <http://‌qualifications.org/‌QualityCertificate1/‌body>.

SimonCox: This proposal is very in line with the RDF approach, but in the real world there will be non-RDF implementations. Its that case that I was asking about
… probably doesn't need to be answered in one place - we could lay out other examples

s/askling/asking/

<SimonCox> I think we should at least say something about possible implementation patterns

riccardoAlbertoni: Examples are useful bit clearly are not limiting

<SimonCox> e.g. in-line in blank nodes, or quality statements in the the same catalog, or perhaps in a separate quality service!

SimonCox: Obviously makes sense to reuse DQV (probably for PROV as well)
… but we also need to consider cases e.g. where quality statements come from third parties

riccardoAlbertoni: Useful to go beyond the linked data perspective

<PWinstanley> Platform stability (Average Availability; Mean time to failure; Mean time to repair; Mean time between failure; etc) Platform response (Response time; Throughput; Capacity; ) Scalability (static - add more kit; dynamic; predictive;) Latency Help Desk availability (none; 9-5; 24/7)

PWinstanley: Given we are thinking of dataservices - we maybe need to think about these (above) dimensions?

riccardoAlbertoni: These kind of quality parameters can be supported by DQV (at least that's the intention)

<PWinstanley> DaveBrowning: I think this is a good approach

<SimonCox> DaveBrowning: riccardoAlbertoni approach looks good in terms of expressivity

<SimonCox> ... as a publisher want to be able to make statement that dataset meets quality policy, metrics

<SimonCox> ... case of third-party assessment is interesting

<SimonCox> ... probably would be in a third place, is not 'part of' the dataset, but is third party annotation

<SimonCox> ... split between consumers and producers

<SimonCox> riccardoAlbertoni: different patterns are possible in architecural design, portal with UI to annotate data? or portal taking inputs from 3rd-party service - use of DQV is neutral

<Zakim> SimonCox, you wanted to note that dqv:QualityMeasurement/dqv:isMeasurementOf should support assement of services as well as datasets

<SimonCox> PWinstanley: yes, need more examples now we are moving into service description!

SimonCox: Suggests we add some implementation models
… while reinforcing the use of DQV. Potentially a few paragraphs that suggest some plausible patterns

<SimonCox> suggest adding some <informative> commentary on implementation options (_not_ recommendations!)

<SimonCox> Good work riccardoAlbertoni - very clear

<general agreement>

Action: riccardoAlbertoni create issue to write some commentary on possible implementation patterns

<trackbot> Sorry, but no Tracker is associated with this channel.

Qualified relations

<SimonCox> https://‌github.com/‌w3c/‌dxwg/‌wiki/‌Qualified-relations

riccardoAlbertoni: Leave current issue/proposal pending any last feedback, aim to approve in a week or so

SimonCox: Motivation to ensure that people on the same page etc
… provide some examples. Then describes requirements/use cases that could utilise this pattern
… provides 3 example solution patterns. Still a work in progress
… Some feedback from Andrea, mostly incorporated
… perhaps no direct action from this work todate, but suggestion is to use this material to provide solutions to the specific requirements

<riccardoAlbertoni> https://‌www.w3.org/‌TR/‌prov-dc/

riccardoAlbertoni: Makes sense to me. There is some relevant work - has that been taken into account

SimonCox: No, not yet

<SimonCox> thanks for reminding me of prov-dc work