Data on the Web Best Practices Working Group Teleconference

<trackbot> Date: 30 October 2015

o worries

no

<ericstephan> ericstephan present+

<scribe> scribe: deirdrelee

<SumitPurohit> There is noise.difficult to hear.

<riccardoAlbertoni> very noisy line today .. deirdrelee please speak louder . .

PROPOSED: Approve minutes from meeting of the 9th October

http://www.w3.org/2015/10/09-dwbp-minutes

<ericstephan> 0 (didn't attend)

<riccardoAlbertoni> +1

<ericstephan> +1 sorry didn't read

<fradulov> +1

RESOLUTION: Approve minutes from meeting of the 9th October http://www.w3.org/2015/10/09-dwbp-minutes

<jerdeb> 0 (didn't attend)

<SumitPurohit> 0 (did not attend)

PROPOSED: Approve minutes from meeting of the 23rd October http://www.w3.org/2013/meeting/dwbp/2015-10-23

+1

<SumitPurohit> 0 (did not attend)

<fradulov> +1

<ericstephan> 0 ( ok for sure didn't attend)

<jerdeb> 0 (didn't attend)

<riccardoAlbertoni> +1

three ppl only online today that were online last week

RESOLUTION: Approve minutes from meeting of the 23rd October http://www.w3.org/2013/meeting/dwbp/2015-10-23

Data Usage Vocabulary

<ericstephan> https://www.w3.org/2013/dwbp/wiki/New_thoughts_on_citations

<ericstephan> http://w3c.github.io/dwbp/images/vocab-overview.png

ericstephan: the first link is the outline I'm going to talk about
... the second link is the overview of the vocab that we talked about in Sao Paulo
... what struck me about our discussion is that the citation concept is too lightweight
... is it adding any value apart from acting as a placeholder

<ericstephan> https://www.w3.org/2013/dwbp/wiki/New_thoughts_on_citations

<ericstephan> Different Perspectives

ericstephan: berna, sumit and I have been reviewing our models
... the different perspectives we were thinking about is 'i have a datasets and I want other people to be able to reference it, for example in an article/journal'
... the other perspective of a citation is that I have a journal article and I want to point to a dataset that I use within the article

<ericstephan> https://www.w3.org/2013/dwbp/wiki/New_thoughts_on_citations Questions

ericstephan: we started talking about these perspectives early on in th evocab discussion, but then we moved to simple how to cite a dataset on the web
... the new thoughts and questions we asked ourselves were is the citation just a piece of inforamtion, like a blog, or is it a machine-readable concept that can be reused
... for someone who is published for a while, there is differences in how for example a name is expressed. there is work to model authors in a common form, so that they're seen as agents, similar to other concepts in the DUV
... because citations care about identifiers such as DOI, should we care about that, should we carry that over into the DUV, so that the citation contains enough information to ensure accessiblity/interoperability
... we were also thinging about how we cite things on the Web, and how about retractions?
... if somebody makes a statement, but then wants to retract it, how is this retracted carried forward into places where the data is cited?
... these are some of our ideas around our thinking on citations

<ericstephan> Activities

ericstephan: will jump into activities now
... we've been looking at some of the citation ontologies out there. There are so many pieces to some of these ontologies!!

<ericstephan> Standards (Not necessarily web) Activity Examples

ericstephan: there are other activities, like the Force11 group, Phil mentioned efforts around cross-referencing. They're very detailed, but there's lots of good work that we may be able to reuse in a meaningful way
... are we going about this the wrong way? not only looking at what ppl in the Semantic Web community are doing in terms of ontologies, but maybe we should look wider at how difference communities / standard bodies are recommending how to use citations
... this helped think about different elements

<ericstephan> Additional Examples

ericstephan: especially how journals are cited in the publication world
... in the Examples section, there is some guidance on rules to cite data. I'm not suggesting we should incorporate all these examples, but we could try to utilise some of these patterns

<ericstephan> Path forward on building the model

ericstephan: make citations machine-readable in some way
... we just have to acknowledge there is not one standard way of citing data on the Web. At the meeting, we had a meaningful exchange with Antoine, etc. about reusing vocabularies
... when we were trying to reuse vocabs, we were looking to bring in allthe concepts and relationships. But what we think we'll do is to bring in the guidelines from communities on how to cite data
... and where appropriate bring in these ideas, but we will build the DUV classes from the rest of the DUV model, but incorporating these guidelines
... what is sometimes confusing for ppl is to think of citations as being directly tied to dataset
... however sometimes the citation might be used in a differenct sense, for example, as part of feedback
... another example is if in the DQV if a user wants to use a citaion
... the last part of the model is the DUV component, maybe the citation is just a user's guide for data
... what we're trying to build out is that citation is more robust and malliable and people can build these out and can use citations as they need to, in difference scenarios, e.g. quality, whatever

deirdrelee: could you explain how this thought process is reflected in the actual concepts used in the model?

http://w3c.github.io/dwbp/images/vocab-overview.png

scribe: what do the concepts represent for you?

ericstephan: what we have now is a really simplified model: cito:CitationAct, foaf:Document and duv:Citation
... what i'm wondering about is, from the class diagram, do we have to actually inherit from cito?
... another thing is in the picture we have prov:Agent. Discussing authors that might be part of a citation, should we be more explicit in that we need a more detailed description of these authors?
... if all we have is first name / surname, do we want to make this discrete so that it's standardised
... some of the standards that we've seen, have a similar duv:Citation class. We'll be looking to these for more specific properties
... we will offer a more streamlined approach, trying to align with the citation community, as opposed to a certain ontology

deirdrelee: how do you see duv being used as opposed to continuing to use their own ontology?

ericstephan: there's a force11 meeting in spring, will get feedback from them
... but will be in touch with them asap around the citation model. I've been talking to some contacts there

<ericstephan> Agreed Phil

riccardoAlbertoni: very intersting discussion
... i haven't understood if you will restructure the schema based onthis research or is this the final schema on which we can comment

ericstephan: what we need to is translate the thoughts from the wiki page into the model, so that the group can comment on it
... we should have a picture by next week and also details on the concept/properties

<PeterWinstanley> Hi

<PeterWinstanley> apologies for being late

DQV

riccardoAlbertoni: not fair to talk about issue-205 when antoine isn't here, would like to have a technical discussion about this
... what I would like to talk about is if Christoph can remain as editor as he has left the group
... I would like to ask to the chairs
... what is the policy here?

deirdrelee: maybe ask in an email to phil, as this is a w3c policy issue

<riccardoAlbertoni> http://w3c.github.io/dwbp/vocab-dqg.html

riccardoAlbertoni: if you look at the picture there is the DQV that is DCAT centric, which comes from many discussions via calls and f2f
... one of the consequence of this discussion is that i tried to restrict the DAQ ontology (Jeremy & co's work) to refer only to DCAT dataset and distribution
... so we can have a quality measure that comes from DCAT, and we can have an inverse property that is not yet in the model,but is generated from computed on

<jerdeb> q

riccardoAlbertoni: here i would like to restrict to dcat dataset and distribution. Is this something we want to keep? or it's not a big problem

or should the DQV be open to non dcat datasets

<riccardoAlbertoni> i cant't hear can you type

<riccardoAlbertoni> yes this is the idea..

jerdeb: if you restrict to only dcat datasets, you are restricting to a small group of datasets that are out on the web
... also i ask if we could use dqv for triple stores, then only pointing to dcat restricts this

riccardoAlbertoni: this was the idea actually, restricting to dcat. in daq you are addressing a larget set of resources
... the question here is if we want to restrict to dcat. this could help the vocab to be more dcat centric. I have no issues with this and could be a distinction between dqv and daq

<riccardoAlbertoni> sorry i can't hear..

<jerdeb> (sorry I'm eating).. but i think that we (as dqv) should not be limited to just a group of datasets

deirdrelee: not sure i agree with 'restricting' usage of dqv

<jerdeb> i agree with deirdre

riccardoAlbertoni: it seems like it's not a good idea to restrict to dqv. it's not a big problem, but in the f2f we discussed a lot about our vocabs being dcat centric
... we can stay general
... and we can maybe restrict after next working draft

<riccardoAlbertoni> ok..

deirdrelee: don't see why it has to be either or - dcat-centric or restrictive - why not both?

I mean dcat-centric and non-restrictive

riccardoAlbertoni: one more issue to discuss

issue-200?

<trackbot> issue-200 -- Can we align the quality dimension hints in DQV with the ones in ISO 25012? -- open

<trackbot> http://www.w3.org/2013/dwbp/track/issues/200

riccardoAlbertoni: we could replace the metrics with the ISO ones suggested by Nandana, this will advance the discussion anyway
... we can decide if these are a good list of dimensions to use. At least we will then have a clearer section in DQV
... let's close the issue and open an action

deirdrelee: sounds good

<jerdeb> q

jerdeb: i think we should also consider the paper from ? as it covers dimensions that may not be covered in ISO standards

riccardoAlbertoni: idea is to add iso dimensions and then add the dimensions from the paper to fill in any gaps
... so jerdeb's proposal is included in our plan

<riccardoAlbertoni> zaveri

<riccardoAlbertoni> thanks

<riccardoAlbertoni> bye

<SumitPurohit> thanks

<jerdeb> bye

<ericstephan> thanks!

<fradulov> bye all

<phila_irc_only> Thanks Dee for chairing and scribing!