Data on the Web Best Practices Working Group Teleconference

<trackbot> Date: 24 September 2015

Wo hoo! Good to have you back zakim bot :-)

<Makx> so I understand the problem was that the start time given on the agenda (12:00 Berlin) was in error. I see that 8:00 Sao Paulo is 13:00 CEST.

<Makx> Some of us were here an hour ago.

:-( A thousand apologies Makx

<Caroline_> Hello!

<Makx> I am connected to WebEx but hear no sound

<Makx> yes, sound is on

<deirdrelee> agenda: https://www.w3.org/2013/dwbp/wiki/Sao_Paulo#Agenda

Welcome

deirdrelee: Thanks Nic BR for hosting us and looking after us so well
... note the poster!
... We're at quite a mature stage now. We need to get into the details of the docs
... we get through a lot at F2F meetings
... I'm sure these days will be the same.
... We're quite punctual. Set up within 20 mins

Vagner_Br: Welcome everyone. Pleasure to have Phil and Dee at the conference yesterday
... Thanks Brazil team for coming as well

Data Quality Vocabulary

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

<deirdrelee> agenda: https://www.w3.org/2013/dwbp/wiki/Sao_Paulo#Friday_25th_September

deirdrelee: Editors have prepared questions for us all

<deirdrelee> dqv agenda: https://www.w3.org/2013/dwbp/wiki/F2F_S%C3%A3o_Paulo_-_Agenda_Proposal_DQV

deirdrelee: Turns over the Antoine and Riccardo

<RiccardoAlbertoni> sure ..

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

antoine: Offers a brief review of the doc at http://w3c.github.io/dwbp/vocab-dqg.html

<scribe> scribe: phila

<scribe> scribeNick: phila

phila: I really like the intro text

antoine: A reminder that we're not defining what quality is - rather, we're offering a core framework that can be extended to create their own metadat about the quality of datasets
... Idea is to make it easy to compare quality assessments and enhance the interop of these systems
... usual sections on conformance and namespaces
... Some vocab review, presents the main parts of the vocab
... RiccardoAlbertoni created this section - very usefeul
... the Datasets class is the subject of any description. Below and around that are the different aspects of quality that we've identified so far
... Top class is the class for dcat:Datastet or dcat:Distribution
... on the left is the part about quality measures
... not so much inspired as copied from DAQ
... the middle part - there are a couple of classes about annotations
... this is an area for discussion later
... on the right is the part about conformance
... we'll have discussions later about conforming to certain standards
... and we're not sure how to indicate that an SLA is available
... so these are the core elements of hte model
... The dotted line are about representing provenance
... this has been discussed on the mailing list. We re-use Prov of course but there is some discussion about containment

phila: Clarified the different divisons

antoine: There's an open nquestion on what role Prov should play, how does that relate to quality
... one aspect of prov is the prov of the quality metadata
... this is important but that's at a different meta level

BernadetteLoscio: Is the quality annotation done by the publisher?

antoine: It can be, or the data re-users
... this will be a topic for future work
... which will go hand in hand with DUV

BernadetteLoscio: Yes, if it's given by the consumer then it's related to the usage, where we have methods for feedback

antoine: Yes.
... That's the a good transition
... You can see that one of the first issues for DQV is the relationship/overlap with DUV
... that's scheduled later in the agenda

WagnerMeiraJr: It seems that you're considering quantitative measure.
... I've seen this done - is that a parallel path?

antoine: This is a good point. Right now, the point about measures is quantitative.
... This comes from the DAQ vocabulary that we've re-used.
... I'm not clear how to add a qualitative measure. That could be seen as an annotation unless there's a clear example of how to make it more structured
... A problem that we face is that we don't have a lot of examples
... so please if you know if any, please provide them

WagnerMeiraJr: One example - in info retrieval. If you have a live experiemnt with users evaluating text, you give them a set of possible responses (like sentiment analysis)
... and you're measuring how the dataset matches expectations
... and benchmarks
... I can collect some examples and send them to the group

antoine: So you mean things like 5 star scales

WagnerMeiraJr: Yes.

antoine: If you could write that up it would be very helpful
... Is a scale quantitative or qualitative? - that's a question
... Probbaly not a good idea to enter deep discussion on this as we've not considered it. So examples would be very helpful so that we can see where they fit in the model.

action meira to collect examples of qualitative feedback and send them to the group, including 5 star scales

<trackbot> Error finding 'meira'. You can review and register nicknames at <http://www.w3.org/2013/dwbp/track/users>.

action wagner to collect examples of qualitative feedback and send them to the group, including 5 star scales

<trackbot> Created ACTION-200 - Collect examples of qualitative feedback and send them to the group, including 5 star scales [on Wagner Meira Jr. - due 2015-10-01].

laufer: Raises issue of metadata about different distributions of the same dataset

antoine: I'm tempted to ask you to raise that later as it's very specific. I don't want to skip it, but from the perspective of the discussion process I'd like to continue the agenda

<RiccardoAlbertoni> i am taking a note

antoine: What I want to go on to is not so exciting... section 1 is the standard text about defining the various classes and properties in the model
... It's currently organised by classes, and then the properties that can typically be applied to this class.
... This is a little different from some vocabs where all the classes are grouped together and then the properties.
... If you think that's not a good way to proceed then please say so.
... otherwise I won't dive into all these tables.
... they just reflect the detailed discussions that we've had
... The tables just reflect what the RDFS/OWL definitions will be
... Section 6 is the example stuff I was mentioning before. We have a general framework so we think it's important to provide examples.
... It's not as complete as it could be so we'll keep on asking for more examples
... There's only so much that we can create ourselves.
... It's good if we can put in real info from real use cases.
... In 6.3 and 4 we see the difficlty about Prov that I mentioned earlier. There's the prov of the annotation and the prov of the dataset
... this will overlap with the BP doc probably.
... We want to include an example of a certificate (e.g. from the ODI)
... and finally something about quality of a linkset. Things published separately - e.g. aligning with SKOS concepts and we want to say something about that.
... We don't want to tell people what quality is but we want others to share their notion of quality
... so we're providing the framework for that.

<Zakim> phila, you wanted to make a suggestion about examples

phila: Suggests using the next publication as a trigger to get new examples

antoine: Can we note an action on the editors to add such a note to the doc

<scribe> ACTION: antoine to add note to DQV document seeking examples from external reviewers [recorded in http://www.w3.org/2015/09/24-dwbp-minutes.html#action01]

<trackbot> Created ACTION-201 - Add note to dqv document seeking examples from external reviewers [on Antoine Isaac - due 2015-10-01].

antoine: So back to section 7 - hints for dimensions and metrics
... this refers to the use cases and elements that we have
... need to come back to this with Riccardo
... Main areas that we want t explore are the ones raised by various contributuions to the WG.
... Ideally we want examples relevant to all the elements in section 7
... The last section of the doc is the one that lists the requirements that we elicited previously. It's a motivation section.

deirdrelee: Is it worth going back to the UCR at this stage?
... might that give us more examples do you think?

antoine: What do you have in mind?

<nandana> http://www.w3.org/TR/dwbp-ucr/#requirements-for-quality-and-granularity-description-vocabulary

deirdrelee: Maybe update the UCR based on the examples that come in?

<antoine> https://www.w3.org/2013/dwbp/wiki/Quality_Aspects_In_Use_Cases

antoine: In principle it makes sense but many of the use cases are very general. In the previous phase of work, we analysed each use case to see what they had about quality
... Many of the UCs are too general to get some example from there without going to back to each UC owner
... Only hesitation is the amount of work involved
... maybe that's a discussion for later?

deirdrelee: OK, let's go ahead with the next point on your agenda

antoine: Any comments on the doc overall?

phila: I like it

<laufer> +1

<nandana> +1

deirdrelee: I think the doc looks very strong. There's a logical flow to it, I think it's def in the right direction and it will be very useful

antoine: Then we shouodl get into the very specific issues

<RiccardoAlbertoni> ok

antoine hands over to Riccardo to go through the issues

RiccardoAlbertoni: So we can start with the first issue

issue-184?

<trackbot> issue-184 -- Is an dqv:ServiceLevelAgreement a kind of certificate, or a standard? -- open

<trackbot> http://www.w3.org/2013/dwbp/track/issues/184

<RiccardoAlbertoni> issue-184?

<trackbot> issue-184 -- Is an dqv:ServiceLevelAgreement a kind of certificate, or a standard? -- open

<trackbot> http://www.w3.org/2013/dwbp/track/issues/184

RiccardoAlbertoni: It's current defined as a kind of standard. One part of the discussion is whether this is right or is it a certificate
... Some have said that an SLA is neither. It's a collection of promises
... So it seems to me that an SLA is quite a complex thing.
... There have been 2 proposals. 1 - keep it as a kind of document
... or model it using entities
... Christophe suggested modelling it as a doc, Antoine suggetsed modelling it along the ODRL model

phila: ASks for clarification of poss use of ODRL

<RiccardoAlbertoni> there is a lot of echo ..

phila: they express things like licensing statements

<nandana> +q to comment why ODRL might be useful

phila: It can fit pretty much any sort of agreement between parties, so we could see an SLA as an instance of that

BernadetteLoscio: maybe I missed something, it;s not clear for me, why do you need this SLA info. Will you use it to calculate sometehing?

RiccardoAlbertoni: The idea is that the SLA tells you how reliable the service is
... so it's related to hte quality

BernadetteLoscio: OK... but for the dataset...

Giancarlo_Guizzardi: There is a lot of work on this. An SLA is a type of social contract. These can be understood as aggregations of commitments and claims.
... This would be an interesting way to look at this. For example, partial satisfaction of a SLA might be met
... I can include some refs to work in this area.

<Zakim> nandana, you wanted to comment why ODRL might be useful

nandana: My first comment matches Bernadette, are there use cases that motivate its inclusion?
... I think ODRL can express commitments clearly

<deirdrelee> action Giancarlo_Guizzardi to share examples around service level agreement activity

<trackbot> Error finding 'Giancarlo_Guizzardi'. You can review and register nicknames at <http://www.w3.org/2013/dwbp/track/users>.

<deirdrelee> action Giancaro to share examples around service level agreement activity

<trackbot> Error finding 'Giancaro'. You can review and register nicknames at <http://www.w3.org/2013/dwbp/track/users>.

<deirdrelee> action Giancarlo to share examples around service level agreement activity

<trackbot> Created ACTION-202 - Share examples around service level agreement activity [on Giancarlo Guizzardi - due 2015-10-01].

http://www.w3.org/TR/dwbp-ucr/#UC-OKFNTransport and http://www.w3.org/TR/dwbp-ucr/#UC-RDESC both ask for SLAs

antoine: My approach to this is there are aspects of quality that everyone agrees are important. Something like the fact that a dataset is refreshed every week

<nandana> another ontology that defines SLAs http://ontology.it/itsmo/v1/itsmo.html#term_SLA

<Makx> +q

antoine: So you can measure that, and you might express it in an ODI Certificate

<Giancarlo_Guizzardi> In the following Core Ontology of Services, we address the use of Service Offerings in terms of social commitments and claims. This might be relevant in this context: http://www.inf.ufes.br/~gguizzardi/UFO-S.pdf

antoine: We can represent this in the mode in several ways
... We don't want to close the door to one way or another. For some people, info that data is udated every week is all they need to know/.

laufer: taking the example of a Distribution. We have the dataset itself and we have the service of providing this data.
... it's confusing to know what we're talking about (what's the subject of the triple)

PeterWinstanley: SLAs require an agreement between parties. So if we're talking about open data there's only one side
... secondly, if you begin to have the level of detail required in ODRL - I can see a lot of organisations who would pass it to the legal department... for a long time
... Danger is getting to a nice vocab that is only useful in a lab

<Zakim> phila, you wanted to weild Occam's Razor

<deirdrelee> phila: agree with peter. wonders should we use the term SLA as there are two sides,as peter said

<deirdrelee> ... if a publisher wants to publish a pledge then FOAF could be sufficient

Makx: We might be mixing up things here. An SLA is a promise or a pledge, it doesn't say anything about the quality of the data
... If I say I'm going to update it every day and trhen I don't, that's confusing
... an SLA doesn't talk about quality

antoine: Lots of points to answer there.
... start with the last first
... Yes, an SLA is not a measure of quality, but it is useful info to potential users

<Makx> +1 to antoine

antoine: an annotation might say nothing about quality either
... it depends omn the provenance whether you can assess these things
... I don't think an SLA will solve all quality issues.
... To Peter on ODRL, the risk of adding sometehing that is too complex. I wouldn't suggest that we go through ODRL and include it here.
... If there is a community that thinks it's good to expfess these pledges, then people should be entitled to do this. It's finding the most flexible way to makr this happen.
... I think we're not on very stable ground and I'm wary of closing doors.

deirdrelee: I agree that the SLA is about the service, not the data, but could be describe the presence of an SLA as a quality metric. So the metric would be - is an SLA present?
... which speaks to the relaibility

<jerdeb> 100% agree with deirdre here

deirdrelee: SO we cojld put that in one of the examples without it being explicitly in the model.

antoine: In this case, i'd consider adding the SLA/Pledge as a sort of annotation rather than a metirc
... something in the same area as the certificates
... The question was whether an SLA is a certificate or a standard. Now it looks like we want to move it to a third place.

RiccardoAlbertoni: I see there are opther people who want to say womthing about this. But I wonder if we're at the point where we can make a proposal
... we are not sure if an SLA is a measuer of quality bnut we don't want to close the door to useful information

<BernadetteLoscio> we had a long discussion about SLA during the F2F at Santa Clara

<BernadetteLoscio> http://www.w3.org/2014/10/30-dwbp-irc#T18-11-17

RiccardoAlbertoni: I'd like to suggest we try and answer a specific question. We have annotations, standards, etc. We want to leave things flexible.

deirdrelee: I think you're raising an important issue wrt to timing. So I'll start timing things.

laufer: I think it's importsant that SLA is important but we need to be clear. We have the provenance about the metadata, and then about the data etc.
... WE have meta meta data

BernadetteLoscio: During the F2F at Santa Clara we discussed SLAs and thought it was probably the wrong term. It's good to look back at that discussion

Giancarlo_Guizzardi: I think it's useful to have a sort of commitment. When an entity makes a commitment, that's a potential reason for using a dataset.
... I'm still puzzled by the use of the word service.
... If there is no social contract then even an SLA commitment might be better.
... Sorry if ontologists have strange concepts sometimes. Whewn you make a pledge, you generally have in mind someone who will receive that commitment.
... Who has the claim here?

antoine: maybe the terminology is the problem.
... Would the term policy, as defined by ODRL, be applicable

<RiccardoAlbertoni> Do we have a link to the definition ?

<nandana> http://www.w3.org/ns/odrl/2/#term-Policy

<Makx> dct:conformsTo

<antoine> Makx++

antoine: I agree with Phil that it's not standard
... but I really like dcterms:conformsTo as a property

deirdrelee: So is that a proposal

<Makx> A basis for comparison; a reference point against which other things can be evaluated

<deirdrelee> RiccardoAlbertoni: thinks we should go for a proposal, and we will have an example showing how an SLA is modelled

<RiccardoAlbertoni> yes

BernadetteLoscio:

If you're going to describe this commitment, maybe you should consider the measure you're going to use to describe the dataset.

scribe: The commitment should reflect the quality description

<Zakim> nandana, you wanted to ask whether antoine is proposing to use ordl:Policy or to define dqv:Policy instead of dqv:SLA

nandana: At some point I think I understood that we don't want to use the ODRL policy, but maybe we want dqv:Policy which might be a sub class of dcterms:Standard

antoine: I think we should have just that - dqv:Policy rdfs:subClassOf dcterms:Standard and then an example

<BernadetteLoscio> yes!

antoine: I assume that what Berna has in mind is the dimensions?

<nandana> I think antoine said dqv:QualityPolicy

laufer: It's interesting to see the differnet users of he document about the quality. We have the publisher, the consumer, etc.
... maybe an intermediary
... Users can make statemetns about the dataset, the service etc. That makes the info complicated

<RiccardoAlbertoni> replace the class dvq:ServiceLevelAgreement with dqv:QualityPolicy (subclass of ODRL:policy ?!? ), and check against some "real" example if this works for the group

deirdrelee: Invites Antoine and Riccardo to wrap this up

<nandana> dqv:QualityPolicy (subclass of dcterms:Standard)?

<antoine> Suggested re-write: 1. replace current dqv:ServiceLevelAgreement with dqv:QualityPolicy, subclass of dcterms:Standard and odrl:Policy 2. add an example with an SLA as Quality Policy, trying to use the same dimensions as metrics and annotations

PROPOSED: 1. replace current dqv:ServiceLevelAgreement with dqv:QualityPolicy, subclass of dcterms:Standard and odrl:Policy 2. add an example with an SLA as Quality Policy, trying to use the same dimensions as metrics and annotations
... 1. replace current dqv:ServiceLevelAgreement with dqv:QualityPolicy, subclass of dcterms:Standard and odrl:Policy

<Makx> +1

PROPOSED: Add an example with an SLA as Quality Policy, trying to use the same dimensions as metrics and annotations

Giancarlo_Guizzardi: I was thinking about what luafer said
... We have these differnet relationships between entities around the dataset
... I've made this dataset according to something else, like a quality policy
... Perhapos there is a general pattern to unify the two vocabs

<nandana> +1, we should also investigate a bit about odrl:Policy semantics as the definition didn't say much

Giancarlo_Guizzardi: You have certain activities... I can use a certain dataset, committing not to do sometehing... differnet roles and activities

<antoine> ODRL has prohibitions as part of the Policy

Giancarlo_Guizzardi: Explores various relationships between different actors.
... The pattern will be the same in DQV and DUV

RiccardoAlbertoni: Not sure if I understand the proposal of Giancarlo

antoine: If people want to represent these things, then that's when they can go and look into ODRL
... ODRL includes way to exprfess constraints and prohibitions
... it might be a good thing to point to odrl:Policy

PROPOSED: Replace current dqv:ServiceLevelAgreement with dqv:QualityPolicy, subclass of dcterms:Standard and odrl:Policy

<deirdrelee> PROPOSED: 1. replace current dqv:ServiceLevelAgreement with dqv:QualityPolicy, subclass of dcterms:Standard and odrl:Policy

<nandana> +1

0 (I don't think dcterms:Standard is right but I defer to Makx)

<yaso> +1

<antoine> +1

<newtonca_> +0

<RiccardoAlbertoni> +1

<BernadetteLoscio> 0

<Gisele> 0

<PeterWinstanley_> 0

Splitting the proposal

<deirdrelee> PROPOSED: replace current dqv:ServiceLevelAgreement with dqv:QualityPolicy,

<RiccardoAlbertoni> let's split in two..

+1

<RiccardoAlbertoni> +1

<Giancarlo_Guizzardi> +1

<antoine> +1

<Gisele> +1

<nandana> +1

<PeterWinstanley_> +1

<laufer> +1

<jerdeb> 0 (i am still not sure about the concept fitting quality metadata - but will give my views after seeing an example)

RESOLUTION: replace current dqv:ServiceLevelAgreement with dqv:QualityPolicy

RESOLUTION: replace current dqv:ServiceLevelAgreement with dqv:QualityPolicy

<deirdrelee> PROPOSED: dqv:QualityPolicy will be subclass of dcterms:Standard and odrl:Policy

<antoine> +1

<deirdrelee> PROPOSED: dqv:QualityPolicy will be subclass of dcterms:Standard

<antoine> +1

<laufer> +1

<nandana> +1

+0 Only 0 not -1 because Makx thinks it's right

<RiccardoAlbertoni> +1

<adrianov> +1

<PeterWinstanley_> 0

<BernadetteLoscio> 0

<Giancarlo_Guizzardi> 0

<antoine> we can add a specific ISSUE about this to call for feedback

<PeterWinstanley_> policy is not a standard

<jerdeb> 0

+1 to antoine adding it as a specific issue

<deirdrelee> PROPOSED: 2. add an example with an SLA as Quality Policy, trying to use the same dimensions as metrics and annotations

<antoine> +1

<nandana> +1

+1

<Giancarlo_Guizzardi> +1

<deirdrelee> +1

<Seiji> +1

<jerdeb> +1

<RiccardoAlbertoni> +1

<Caroline_> +1

<PeterWinstanley_> +1

<laufer> +1

<BernadetteLoscio> +1

RESOLUTION: Add an example with an SLA as Quality Policy, trying to use the same dimensions as metrics and annotations

<scribe> ACTION: riccardo to add an example with an SLA as Quality Policy, trying to use the same dimensions as metrics and annotations [recorded in http://www.w3.org/2015/09/24-dwbp-minutes.html#action02]

<trackbot> Created ACTION-203 - Add an example with an sla as quality policy, trying to use the same dimensions as metrics and annotations [on Riccardo Albertoni - due 2015-10-01].

deirdrelee: Because we didn't resolve to make it a subclass of dcterms:Standard, I assumed we have not resolved the odrl:Policy issue as wll

issue-185?

<trackbot> issue-185 -- dqv:QualityAnnotation modeling issues -- open

<trackbot> http://www.w3.org/2013/dwbp/track/issues/185

RiccardoAlbertoni: That was raised by Antoine.

antoine: Actually I;m not si sure what the issue was except a general call for feedback
... We would recommend that the instances of this class ?? from Open Annotation
... this was a call for feedback. There has not been a lot of feedback,. so maybe that foodback will come when we look at DUV. For now though I'd say this issue can be closed.
... If everyone's OK I'd say we could resolve to close.

RiccardoAlbertoni: Just wondering whether motivation should refer to ao SKOS concept - do we need a spedcific taxonomy?

<nandana> oa:Annotation -> http://www.w3.org/ns/oa#d4e434

antoine: My feeling is that we should just show some examples but not represent all possible motivations
... One exception might be that the quality dimensions coujld be modelled as motivations - but I'm not sure

<nandana> oa:motivatedBy -> http://www.w3.org/ns/oa#d4e230

BernadetteLoscio: ... The issue is the modelling or the usage of hte annotation? It's not clear what is a quality annotation?

RiccardoAlbertoni: AIUI, the ontology's vision is that you have to indicate the moitivation for your annotation and that's usually expressed as a SKOS concept.
... it could be a post or a reply
... etc.
... We could have a basic taxonomy?

antoine: I realised I've not been super clear... the proposal is to have at least one concept defined in our namespace - quality assessment. We can do that easily enough
... and then if they want to define sub concepts of that, OK

<RiccardoAlbertoni> ok

antoine: I think we should just have this one concept

Giancarlo_Guizzardi_: A comment about the relationship between user feedback and quality annotation
... user feedback can be about anything
... maybe the user feedback in data usage is more general than one that makes any statement about quality

laufer: I don't want to try and define quality, but when we talk about annotation, we can say whether the annotation is quality info or not

deirdrelee: Would it make sense 165 and move on to 185?

<deirdrelee> close issue-185 and move to issue-165

Giancarlo_Guizzardi_: We could capture what kind of metric/dimension this annotation is about

<antoine> Giancarlo_Guizzardi++

<antoine> (thought that would be a different issue)

<antoine> PROPOSED: dqv:QualityAnnotation is a subclass of oa:Annotation. The instances of this class should have one oa:motivatedBy statement with a an instance of oa:Motivation (and skos:Concept), which reflects a quality assessment purpose. We define this instance as dqv:qualityAssessment.

phila: I like Giancarlo_Guizzardi_'s ideas but they sound like something for hte lab, not the real world.

BernadetteLoscio: Just a comment - we have the policy that says what is expected, the quality statement by the publisher and the assessment of the user
... so maybe it would be nice to define everything using the same dimensions

<deirdrelee> phila: if every instance of the class has a particular property, can't you add that inthedefinition of the class. could you use owl to say that they property oa:motivation exists? otherwise redundent triples

phila: If the same triples are always defined, do we need to state them. Can we do it without having to state those triples every time?

antoine: Yes, you define it as an OWL equivalent class with those features

PROPOSED: dqv:QualityAnnotation is a subclass of oa:Annotation. The instances of this class should have one oa:motivatedBy statement with a an instance of oa:Motivation (and skos:Concept), which reflects a quality assessment purpose. We define this instance as dqv:qualityAssessment

<antoine> +1

<Giancarlo_Guizzardi_> +1

<nandana> +1

<Caroline_> +1

<yaso> 0

<RiccardoAlbertoni> +1

<PeterWinstanley_> 0

<Makx> +1

<laufer> +1

+1

RESOLUTION: dqv:QualityAnnotation is a subclass of oa:Annotation. The instances of this class should have one oa:motivatedBy statement with a an instance of oa:Motivation (and skos:Concept), which reflects a quality assessment purpose. We define this instance as dqv:qualityAssessment

<deirdrelee> RESOLVED: dqv:QualityAnnotation is a subclass of oa:Annotation. The instances of this class should have one oa:motivatedBy statement with a an instance of oa:Motivation (and skos:Concept), which reflects a quality assessment purpose. We define this instance as dqv:qualityAssessment

RESOLUTION: dqv:QualityAnnotation is a subclass of oa:Annotation. The instances of this class should have one oa:motivatedBy statement with a an instance of oa:Motivation (and skos:Concept), which reflects a quality assessment purpose. We define this instance as dqv:qualityAssessment

== Short Break ==

<scribe> scribe: yaso

<Caroline_> 5min break

<BernadetteLoscio> dqv:qualityAssessement is defined on the model?

<RiccardoAlbertoni> Antoine, is there any issues you want to focus in particular .. ?

<antoine> I think we could try to NOT discuss the provenance issues, which are very technical (181). The voc management ones could be postponed, but I guess they will be naturally as they are at the end :)

<RiccardoAlbertoni> Ok

<antoine> Riccardo, I think we should focus on 165, 187, 190, 164, 153, and GeoDCAT-AP

<RiccardoAlbertoni> let's go for the 165 ? which I am afraid will bring lot of discussion .. what do you think .. and then 187 I think there is a kind of consensus about keep the cardinality between dimension and metric as in DAq so I suppose we can easily close it ..

<antoine> 165 is needed as per the connection to DUV

<RiccardoAlbertoni> ok

<antoine> 187 may not be so consensual. Actually I had understood that the consensus was rather on not keeping the cardinality from daQ, i.e. relax it.

<RiccardoAlbertoni> ok let's go for 165, 187, 190, 164, 153, and GeoDCAT-AP ..

antoine: this is issue 165

<deirdrelee> issue-165

<trackbot> issue-165 -- What is the relation between duv:Feedback and dqv:UserFeedback? -- open

<trackbot> http://www.w3.org/2013/dwbp/track/issues/165

http://www.w3.org/2013/dwbp/track/issues/165

antoine: user feedback is a kind of quality annotation

<nandana> DUV -> http://w3c.github.io/dwbp/vocab-du.html#Vocab_Overview

Giancarlo_Guizzardi: things like correction and suggestion
... you can say "I don't like this vocab" and etc

<BernadetteLoscio> +1

Giancarlo_Guizzardi: no all feedback is quality annotation

antoine: actually i completely understand what Giancarlo_Guizzardi said

<RiccardoAlbertoni> +1 to QualityUserFeedback

Giancarlo_Guizzardi: that was going to be my suggestion

laufer: my question is to Giancarlo_Guizzardi: if only sometimes feedback are quality information, how can we say that an annotation is a quality information if we don'tlate it to a dimension
... yes, but how say to the user that he can or not put this information in this document

<nandana> +q to ask whether it was intentionally left Feedback to cover more things than user feedback? (so instead of using dqv:UserQualityFeedback and use just dqv:QualityFeedback)

Giancarlo_Guizzardi: I think this has to do with the issue regarding user Feedback

phila: in practice, you might have a CKAN portal, somebody makes use of a dataset - it's hard to get people to fit at your data model
... if it's a machine that is going to classify your data then you can have a more complicated model
... but if otherwise, not
... we should keep a strong focus on how it will be in a real world, with real application

BernadetteLoscio: we were discussing if we should have specific types of feedback
... if we want to specify the types of feedback

<Zakim> nandana, you wanted to ask whether it was intentionally left Feedback to cover more things than user feedback? (so instead of using dqv:UserQualityFeedback and use just

BernadetteLoscio: if we decide that we are going to have just feedback, so it fits the data usage vocab

nandana: you said feedback specifically

<BernadetteLoscio> q_

<nandana> I was just referring to the difference Feedback and *User*Feedback

BernadetteLoscio: I think we can keep feedback general

<RiccardoAlbertoni> Proposal: rename DQV:UserFeedback with dqv:QualityUserFeedback making it as duv:Feedback

BernadetteLoscio: but tomorrow we are going to discuss if we are going to have this subclass, right?

<Giancarlo_Guizzardi> Feedback = UserFeedback (as far as I understand it). You are right that only one of them should be used

<RiccardoAlbertoni> Proposal: rename DQV:UserFeedback with dqv:QualityUserFeedback making it as duv:Feedback subclass

antoine: right now I don't feel that is incompatible with this decision
... I like the fact that is from users

<RiccardoAlbertoni> +1 to antoine

<Giancarlo_Guizzardi> +1 (terminologically speaking it seems like a good idea)

it will be interesting to keep the user on the label of the class

<Giancarlo_Guizzardi> in that case we would have UserFeedback that is specialized in QualityUserFeedback

<nandana> +1

deirdrelee: ok so for now we will keep the user on the class

<Giancarlo_Guizzardi> which in turn specializes QualityAnnotation

<RiccardoAlbertoni> Proposal: rename DQV:UserFeedback with dqv:QualityUserFeedback making it as duv:Feedback subclass

<antoine> +1

deirdrelee: and in terms of ricardo's proposal

<Seiji> +1

<jerdeb> +1

<RiccardoAlbertoni> +1

<deirdrelee> +1

Giancarlo_Guizzardi: I guess that if we agree with antoine, feedback should be UserFeedback

<RiccardoAlbertoni> yeah please

<Giancarlo_Guizzardi> rename DQV:UserFeedback with dqv:QualityUserFeedback making it as duv:UserFeedback subclass

<adrianov> +1

<laufer> +1

deirdrelee: in general we agree with this?

+1

<BernadetteLoscio> +1

<Giancarlo_Guizzardi> +1

<Caroline_> +1

close issue-165

<trackbot> Closed issue-165.

<deirdrelee> RESOLVED: rename DQV:UserFeedback with dqv:QualityUserFeedback making it as duv:Feedback subclass

antoine: we proposed to move to issue-187

<deirdrelee> issue-187

<trackbot> issue-187 -- Do we want to keep the same occurrence constraints as defined in DAQ (for example, that every metric should belong to exactly one dimension)? -- open

<trackbot> http://www.w3.org/2013/dwbp/track/issues/187

Do we want to keep the same occurrence constraints as defined in DAQ (for example, that every metric should belong to exactly one dimension)?

Giancarlo_Guizzardi: dimension would be a quality dimension, like height etc
... so availability would be a sort of dimension

antoine: right now it belongs to one dimension
... imagine that uptime could be a metric for availability

<BernadetteLoscio> antoine - I can't find dqv:qualityAssessement in the model

<Zakim> jerdeb, you wanted to clarify

jerdeb: a dimension is made of many metrics

<antoine> BernadetteLoscio - yes the idea is that we've resolved to add it

Giancarlo_Guizzardi: then I would change the description in the document
... it is a bit confusing

<antoine> yes the wording with "unit" is confusing

<nandana> DAQ diagram -> http://butterbur04.iai.uni-bonn.de/ontologies/daq/daq#_introduction

<BernadetteLoscio> antoine - ok! sorry :)

RiccardoAlbertoni: I was wondering if Giancarlo_Guizzardi can write the definition that he suggested

<Gisele> RiccardoAlbertoni: we are closig the issue and keeping the constraints

<antoine> -1

<Giancarlo_Guizzardi> A Metric is not a unit of measuring. An Observation (QualityMeasure) assigns a value in a given unit to a Metric

<Gisele> jerdeb: I suggest that we use these constraints and propose to provide guidelines

<Gisele> ... I believe we should keep than as guidelines but not formally constrain them

<antoine> jerdeb++ it's really great that you've consulted with colleagues!

<Gisele> ... we should still provide guidelines

<nandana> +1 to jerdeb

<phila> +1

<Gisele> deirdrelee: that would mean that the concepts we use would have no constraints but only guidelines

<Giancarlo_Guizzardi> I meant "Unit of Measurement" instead of "unit of measuring" (although DQV mentions "unit of measuring")

<Zakim> phila, you wanted to ask if a metric is a slice

<Gisele> phila: aguideline is a slice thourgh a datacube

<Gisele> ..i'trying to match what are saying with my knoeldge on data cube

<Gisele> jerdeb: we can have slices from multiple observation but im not sure

<Gisele> deirdrelee: proposal for issue 187

<deirdrelee> PROPOSED: Don't keep the constraints from DAQ but provide guidelines

<ericstephan> good morning phila :-)

<deirdrelee> +1

<BernadetteLoscio> Hi Eric! :)

<antoine> +1

<phila> +1 Noting that DAQ is moving in the same direction

<Seiji> +1

<nandana> +1

<Gisele> +1

<RiccardoAlbertoni> +1

<laufer> +1

<ericstephan> ericstephan present+

<adrianov> +1

<jerdeb> +1 (will remove constraints from daQ as well)

Giancarlo_Guizzardi: what do we mean by "provide guidelines"

<deirdrelee> RESOLVED: Don't keep the constraints from DAQ but provide guidelines

close issue-187

<trackbot> Closed issue-187.

issue-189

<trackbot> issue-189 -- Aether VoID extension uses a different from the pattern that DQV inherits from DAQ -- open

<trackbot> http://www.w3.org/2013/dwbp/track/issues/189

Giancarlo_Guizzardi: we could have a uniform treatment of this part of the model. So I was wondering if we are going to have a discussion on this, or it will be discussed my email.. It's a general question
... there's a lot of work here that we could use, there's a lot o discussions here

<Giancarlo_Guizzardi> Ok. Thanks

<deirdrelee> issue-189

<trackbot> issue-189 -- Aether VoID extension uses a different from the pattern that DQV inherits from DAQ -- open

<trackbot> http://www.w3.org/2013/dwbp/track/issues/189

antoine: in our patter for measure on the quality of datasets there will be a connexion between

<RiccardoAlbertoni> to phila http://jiemakel.github.io/aether/#/

antoine: the question is if we are comfortable with a more complex proposal and if we should try to reflect this in the document

<Giancarlo_Guizzardi> A known proposal in the subject actually comes from VU: http://www.semantic-web-journal.net/sites/default/files/swj177_7.pdf

<antoine> can I try to re-phrase the problem?

<antoine> <http://data.europeana.eu/void/Dataset> <http://ldf.fi/void-ext#averageLiteralLength> "62,014"^^<http://www.w3.org/2001/XMLSchema#decimal> ;

<Giancarlo_Guizzardi> Other relevant references are: http://www.qudt.org, https://www.oasis-open.org/committees/tc_home.php?wg_abbrev=quomos. We also did some work on that which could be relevant: http://www.inf.ufes.br/~gguizzardi/PID2733627.pdf and http://ceur-ws.org/Vol-1301/ontocomodise2014_9.pdf

antoine: this about it and decide if this is something that we are comfortable with

<phila> scribe: Giancarlo_Guizzardi

RicardoAltertoni mentions that we need some of this complexity in the proposal but he also thinks that Antoin is right in advocating a simple way of representing statistics

<nandana> +1 to RiccardoAlbertoni. we might start with Aether but end up again something similar to DAQ if we want to represent the all the information to that we represent now

<Zakim> phila, you wanted to make a suggestion

<RiccardoAlbertoni> +1 to jerdeb

<phila> phila: I asked whether it might be possible to treat Aether VoID as a quality meansure within DQV, or use a CONSTRUCT query to convert from one to the other. The answer was no.

laufer, DQV is a way to besides the quality of the data, it semantically describes the data

<deirdrelee> laufer: trying to understand issue. i think that dqv is a way of as well as describin quality data, we describe the semanticcs

<deirdrelee> ... void is description of the data,not the semantics of the informaiton. so void and dqv are two differnet things

<deirdrelee> ... it is more complicated to describe the semantics ofthe data

<RiccardoAlbertoni> +1 to laufer

laufer, these are two separate things. Describing the semantics of the quality of the data is more complicated but it is a more general model

<jerdeb> jerdeb: re transform Aether to daQ -> no this cannot happen because we do not know any quality information about the aether property, such as category and dimension which are important to daQ

<deirdrelee> ... dav is a simpler model, but they don't have semantics

riccardo, once you have described your property with DAQ, it is possible to serialized it Aether but not the other way around

<deirdrelee> RiccardoAlbertoni: it is difficult to erialise the results in the header as daq is more complex

RiccardoAlbertoni: the bridge can only be done in one direction

<deirdrelee> antoine: it's not that the void extension has not semantics. the semantics are in the properties

antoin, it's not that the VoiD extension has no semantics.

<deirdrelee> ... the semantics are not positions as a quality metric

<deirdrelee> ... we feel as a group not satisfied, that we're missing something

<deirdrelee> ... it must be possible to express these simple triples

<deirdrelee> laufer: agree we have semantics in the properties, but in dqv in the description there will be semantics, a more sophisticated way

antoine: we seem to be calling semantics different things

<phila> issue-189?

<trackbot> issue-189 -- Aether VoID extension uses a different from the pattern that DQV inherits from DAQ -- open

<trackbot> http://www.w3.org/2013/dwbp/track/issues/189

<deirdrelee> ... we don't really have the means to find out they're about quality, they're just properties. dqv provides a framework that facilitates interoperability

<deirdrelee> deirdrelee: make a specific proposal around issue-189

<deirdrelee> RiccardoAlbertoni: proposal, keep dqv as it is,provide guidance on how to convert daq to aether void

<deirdrelee> antoine: we should have this for market adoption

<deirdrelee> phila: i've been looking at aether, i'm happy to say we keep dqv as it is

<deirdrelee> ... they're not antagonistic, we can extend dqv with aether, there are lots of other things that you could say about the dataset, not necessary to include

<deirdrelee> antoine: really have doubts about it. i'm in a community where they'll look at vocab like aether and say this is quality

<phila> phila: That makes it sounds as if I don't like Aether voID - it looks very interesting. I just don't think it's necessarily something we should feel obliged to move towards/include as an example

<deirdrelee> ... agree not all statistics are relevant for quality, but would like us to be stronger about

<phila> PROPOSED: keep dqv as it is, provide guidance on how to convert daq to aether void

<deirdrelee> antoine: in my community this is important

<deirdrelee> PROPOSED: keep dqv as it is, provide guidance on how to convert daq to aether void

<jerdeb> -1 (because there are other ontologies like aether, such as lodstats)

<phila> +1

<laufer> -1

<antoine> +1

<ericstephan> +1

<deirdrelee> 0

<PeterWinstanley_> +1

<Seiji> +1

0

<Caroline_> +1

<RiccardoAlbertoni> +1

<Gisele> 0

<deirdrelee> anthoine: it doesn't have to be aether

<deirdrelee> PROPOSED: keep dqv as it is, provide guidance on how to convert daq to another quality statistics vocabulary

<laufer> +1

<Seiji> +1

<nandana> +1

<jerdeb> +1

+1

<adrianov> +1

<antoine> +1 to "work with"

<Caroline_> +1

<ericstephan> +1

<RiccardoAlbertoni> +1

<deirdrelee> PROPOSED: keep dqv as it is, provide guidance on how daq can work with another quality statistics vocabulary

+1

<phila> +1

<deirdrelee> +1

<Gisele> +1

<laufer> +1

<jerdeb> +1

<adrianov> +1

<Seiji> +1

<RiccardoAlbertoni> +1

<Caroline_> +1

<deirdrelee> RESOLVED: keep dqv as it is, provide guidance on how daq can work with another quality statistics vocabulary

<nandana> _1

<nandana> +1

<ericstephan> ++++++++++++++++++1

<deirdrelee> close issue-189

<phila> RESOLUTION: keep dqv as it is, provide guidance on how daq can work with another quality statistics vocabulary

<trackbot> Closed issue-189.

<deirdrelee> issue-164

<trackbot> issue-164 -- Are statistics about a dataset a kind of quality info we need to include in the data quality vocabulary? -- open

<trackbot> http://www.w3.org/2013/dwbp/track/issues/164

<RiccardoAlbertoni> yeah , let's close it

<phila> I think we just resolved that issue

<deirdrelee> antoine: proposal is we agree that some statistics may be relevant for expressing data quality

<deirdrelee> ... after resolution for issue-189, we should have examples

<phila> PROPOSE: Close Issue 164 as previous proposal covers it

<phila> PROPOSE: Close Issue 164 as previous resolution covers it

<deirdrelee> +1

<BernadetteLoscio> +1

<phila> +1

<Seiji> +1

<laufer> +1

<antoine> +1

<RiccardoAlbertoni> +1

<adrianov> +1

<ericstephan> +1

<newtoncalegari> +1

+1

<nandana> +1

<deirdrelee> RESOLVED: Close Issue 164 as previous resolution covers it

<Caroline_> https://www.w3.org/2013/dwbp/track/actions/153

<deirdrelee> action-153

<trackbot> action-153 -- Antoine Isaac to Look at completeness as one of the quality dimensions -- due 2015-04-20 -- OPEN

<trackbot> http://www.w3.org/2013/dwbp/track/actions/153

<deirdrelee> RiccardoAlbertoni: the discussion in the mailing list says that completeness is an example of one ofthe quality dimensions

<deirdrelee> jerdeb: agree that completeness should be a dimension. how can we measure this as linked data?

<deirdrelee> ... open world assumptions

<deirdrelee> RiccardoAlbertoni: we have to assume closed world to measure completeness

<deirdrelee> jerdeb: that's one of the main problems i'm having, difficult to measure

<deirdrelee> antoine: some more info on this aciton. like RiccardoAlbertoni said, completeness is important to measure, but didn't receive much feedback

<deirdrelee> ... asked for specific feedback on completeness but didn't get any

<deirdrelee> ... similar to statistics, used this as a proxy for completeness

<deirdrelee> ... if you are keen for completeness, please send examples

<deirdrelee> phila: one of the reasons i don't like using dcterms:conformsto is that ?

<deirdrelee> ... there are various things that you could point to something that defines what complete meants

<deirdrelee> RiccardoAlbertoni: there are some situation that the closed world assumptoin works fine, and we would like the opportunity to say something about it

<phila> phila: I was saying that one quality description could be that the dataset matches a spedcific profile, such as DCAT-AP, or maybe point to a SHACL description (http://w3c.github.io/data-shapes/shacl/)

<deirdrelee> phila: it could even be a description in a pdf. as long as you say it conforms to, e.g. to say that ckan dcat export

<deirdrelee> RiccardoAlbertoni: we are thinking about different kind of completeness. the one i'm talking about is the one in the linked data survey

<deirdrelee> ... being compliant to a certain profile

<phila> PROPOSED: That we include completeness as a quality metric. That can be defined in any way that puts boundaries around what the data should contain.

<nandana> +1 for including completeness dimension and asking for concrete metrics and examples.

<phila> PROPOSED: That we include completeness as a quality dimension. That can be defined in any way that puts boundaries around what the data should contain.

<phila> PROPOSED: That we include completeness as a quality dimension. That can be defined in any way that puts boundaries around what the data should contain, closes the world etc.

<BernadetteLoscio> +1

<Seiji> +1

+1

<antoine> +1

<nandana> +1

<RiccardoAlbertoni> +1

<deirdrelee> +1

<Gisele> +1

<phila> +1

<laufer> +1

<Caroline_> +1

<ericstephan> +1

<phila> RESOLVED: That we include completeness as a quality dimension. That can be defined in any way that puts boundaries around what the data should contain, closes the world etc.

<phila> RESOLUTION: That we include completeness as a quality dimension. That can be defined in any way that puts boundaries around what the data should contain, closes the world etc.

<jerdeb> +1 if we have specific examples for population completeness and schema completeness

<phila> close issue-153

<trackbot> Closed issue-153.

<RiccardoAlbertoni> yeah it was very useful ..

<antoine> Thanks everyone!!!

<phila> == Lunch ==

<BernadetteLoscio> Thanks!

<laufer> Thank antoine, riccardo

<phila> close action-153

<trackbot> Closed action-153.

<jerdeb> sorry but need to leave now. will join you again tomorrow morning (afternoon here in germany)

<deirdrelee> close-164

<deirdrelee> close issue-164

<trackbot> Closed issue-164.

<RiccardoAlbertoni> enjoy the lunch!

<deirdrelee> ok, let's get back...

<newtoncalegari> Let's go

<RiccardoAlbertoni> yeah.. I am on webex

<deirdrelee> https://www.w3.org/2013/dwbp/wiki/F2F_S%C3%A3o_Paulo_-_Agenda_Proposal_DWBP

<newtoncalegari> yaso: starting the meeting with BP agenda

<yaso> https://www.w3.org/2013/dwbp/wiki/Sao_Paulo#Agenda

<deirdrelee> scribe: newtoncalegari

<yaso> ISSUE-137

<trackbot> ISSUE-137 -- Review BP Preserve person's right to privacy -- open

<trackbot> http://www.w3.org/2013/dwbp/track/issues/137

yaso: issue-137 about privacy

BernadetteLoscio: during the last F2F we had a discussion about privacy

<deirdrelee> BP Editor's draft: http://w3c.github.io/dwbp/bp.html

BernadetteLoscio: and we raised some issues
... now we need to decide what we gonna do
... hadleybeeman was not sure if we will talk about privacy

<yaso> http://bernafarias.github.io/dwbp/bp.html#PreservePrivacy

BernadetteLoscio: so, there is this action to rewrite the BP. but we need to discuss what we need to do
... to rewrite, keep it as it is, or discontinue..
... the discussion in the end of last f2f was not clear if we should have a BP like this, about privacy
... what we should do with this section?

phila: yes, hadleybeeman make the point that it's not technical

<phila> It is clearly essential that individuals' privacy is respected. This means that Personally Identifiable Information needs to be handled according to policies and procedures that reflect the local jurisdictional context and it is therefore beyond the scope of this document to make specific recommendations on this topic.

phila: makx suggested that we put some text in the introduction
... we can potential link to some definition

<yaso> ericstephan,

phila: the scope is policy and law, and it's not in the technical scope

<phila> Makx's comment

ericstephan: phila is propposing to take this out of the BPs, and put it in another section
... it could be in a section called "Assumptions"

<phila> ericstephan: Perhaps we could have a section called 'Assumptions' - i.e. things that we recognise as being important but that are out of scope.

yaso: there are challenges in the Use Cases and this challenge is out of the scope

<deirdrelee> BP 20q+

phila: we can't in a Technical document say "you should follow the law"

<phila> Principles/assumptions - I;m OK wth either

(someone could help me with Peter speech?)

BernadetteLoscio: we can add some links appoint to this issue of Sensitive Data

yaso: is there any way to ask to other groups to deal with it?

phila: dealing with this issue is different of using dublin core, or vocabs

<phila> Peter was saying that in Enterprise Architectures often begin with a bunch or principles

deirdrelee: there is a technical element on saying to share, integrate, publish datasets

<Caroline> +1 to deirdrelee

<Caroline> to put in the BP document

laufer: the ideia of a paragraph is interesting to clarify people there is a law to deal with it
... we don't need to have a BP about it, the paragraph is enough

Caroline: BernadetteLoscio and I discussed about it
... if we don't keep it as a BP, it's important to say something about this topic
... say that this topic is broader than other BPs
... I Agree with Deirdre on point about this in the Document

Giancarlo_Guizzardi: I agree
... but it's very complicated
... this seems to be also related to one aspect of quality
... this is related with conformance

deirdrelee: if you classify a dataset, you can say some terms are commercial, others sensitivity
... maybe if we provide a classification for the dataset

BernadetteLoscio: it's quite similar of what laufer propposed
... if we have a proporty to describe of classify a dataset, maybe it could be a part of the dataset description

deirdrelee: but it could be not only human-readable

<antoine> http://www.w3.org/TR/vocab-dcat/#classifying-datasets ?

deirdrelee: need to be machine readable

yaso: I don't think we don't have to try to classify
... it's not only technical
... I'm in favor of keep as a note and maybe other group to talk about Privacy

phila: Norway classifies datasets

BernadetteLoscio: I think the idea of having an extra metadata to describe this kind of information could be nice

yaso: for instance, if facebook has an API and I get data from my friends. but suddenly facebook closes the API, and I can't get my data anymore
... is this data public? sensitive?
... for me it's hard to do that, to classify datasets in that way

Seiji: I think we can have an agreement among those ideas
... in the Draf there is not section called "Policy"

WagnerMeiraJr: I believe it's important to make a difference between what is Ethical and what is Illiegal
... each provider usually have different rules
... it's in our scope to discuss in a BP doc what is illegal or not (???)
... data is in some sense public
... for instance, the ashley madison web site
... they had a agreement about privacy
... but hackers have violated and have gotten the private data
... the second point is what is illegal?
... we may recommend that you verify license and terms

<Zakim> deirdrelee, you wanted to say gov already classify docs as open, confidential, secure, etc

deirdrelee: the data is on the Web, you should recommend a license

ericstephan: (??? sorry ericstephan -( )

laufer: the problem is not when you have the data, is when someone access the data
... you can have the some dataset with different licenses

<ericstephan> I was just mentioning that I agreed with WagnerMeiraJr and mentioned that this was complementary to the way people think of open and closed data

<deirdrelee> possible proposal: Remove BP 20 'Preserve people's right to privacy'. Instead add a note around data protection and linking to other related work

Caroline: I understand we maybe put as a note

BernadetteLoscio: in this case we need to remove the section, 2 BPs and challenge 'Sensitive Data'

<Makx> +1 to remove BP20. there is no universal 'right to privacy'

yaso: I don't know if we can remove the section

<Zakim> phila, you wanted to argue for BP21

<annette_g> these are two different questions

phila: I agree on removing the BP20
... but BP21 we cuold keep

<ericstephan> +1 to keeping BP21

<antoine> +1 Actually BP21 is much more on the technical side of things.

<RiccardoAlbertoni> +1 to phil

<yaso> +1 to keeping 21

<Makx> +1 to keep BP21

BernadetteLoscio: but you want to keep the Sesntive Data section, with the BP21 but without BP20?

phila: (you answered yes?)

<deirdrelee> newtoncalegari: should bp21 stay in the sensitive data section or it should be moved?

phila: it's up to the editors :-)

<phila> newtoncalegari: On keeping the BP on Provide data unavailability reference - does that mean keeping the section but with one BP?

<Makx> BP21 has nothing to do with sensitive data

<phila> phila: That's up to the editors

annette_g: for BP20, are we going to replace if it's removed?

<Makx> the only thing we can say about BP20 is that data providers should respect applicable laws

<yaso> +1 to Makx

phila: we are agreeing on taking out BP21 but write some notes about the topic

BernadetteLoscio: if we keep the section 'Sensitive Data', I think we can't have only the description without a BP

<annette_g> +1 to Bernadette

BernadetteLoscio: we can keep the BP21, but this BP seems to doesn't fit in the Sensitive Section
... and we need to look for a place for BP21

Caroline: I agree with BernadetteLoscio

<Seiji> +1 to what bernadette said

Caroline: maybe BP21 fits in Data Preservation section
... and the text of BP20 will be transformed in a note

yaso: we need to let people know Sensitve Data is a challenge, and I think we can keep the section

<Caroline> +1 to yaso that privacy is a challenge important to be mentioned

BernadetteLoscio: Ok. It can be a challenge, but no in the scope of this document, and I don't agree on having a section like others, but without any BP

deirdrelee: two counter suggestions

<Caroline> Caroline: the editors may make the change on Sensitive data and send to the group

annette_g: maybe we can rewrite the BP
... we need to consider those issues

PROPOSAL: removing BP20 - Preserve people's right to privacy

<phila> PROPOSED: That the Best practice on Preserve people's right to privacy be removed and replaced by a suitable note/paragraph

<annette_g> removing and replacing by a note are two different issues

<yaso> +1

<Makx> +1

<annette_g> 0

<Caroline> +1

<AdrianoCesar-InWeb> +1

<BernadetteLoscio> +1

<BernadetteLoscio> +1

+1

<Seiji> +1

<Gisele> +1

<ericstephan> +1

<deirdrelee> +1

<nandana> +1

annette_g: I think it should a BP

<Caroline> +q

annette_g: agree on removing BP20, but put another BP in the empty place (??? is that right, annette_g? )

phila: writing a draft proposal

<phila> Draft proposal - That the Best practice on Preserve people's right to privacy be removed

<annette_g> * :)*

<phila> Draft proposal that the section on sensitive data be reviewed in the broader scope of the doc

<phila> draft proposal - that privacy is an important issue and we shouold say something, even if it is only "think about this stuff"

<phila> PROPOSED: That the Best practice on Preserve People's Right to privacy be removed

<yaso> +1

<Makx> +1

<annette_g> +1

+1

<Caroline> +1

<RiccardoAlbertoni> +1

<Seiji> +1

<phila> +1

<ericstephan> +1

BernadetteLoscio: Editors will review the Senstive Data section

<phila> PROPOSED: That the section on sensitive data be reviewed in the broader scope of the document

<BernadetteLoscio> +1

<Makx> +1

+1

<laufer> +1

<phila> +1

<Gisele> +1

<deirdrelee> +1

<Seiji> +1

<Caroline> +1

<BernadetteLoscio> +1

<annette_g> +1

<AdrianoCesar-InWeb> +1

<phila> RESOLVED: That the Best practice on Preserve People's Right to privacy be removed

<yaso> newtoncalegari, Draft proposal - That the Best practice on Preserve people's right to privacy be removed

<phila> RESOLUTION: That the Best practice on Preserve People's Right to privacy be removed

<phila> RESOLVED: That the section on sensitive data be reviewed in the broader scope of the document

<phila> RESOLUTION: That the section on sensitive data be reviewed in the broader scope of the document

<phila> close action-164

<trackbot> Closed action-164.

<phila> close action-166

<trackbot> Closed action-166.

<phila> close issue-137

<trackbot> Closed issue-137.

<BernadetteLoscio> bye bye! thanks!

<annette_g> * closed instead of close? *

<yaso> http://www.w3.org/2013/dwbp/track/issues/166

<yaso> issue-166

<trackbot> issue-166 -- Should the data vocabularies section be removed? -- open

<trackbot> http://www.w3.org/2013/dwbp/track/issues/166

yaso: we're going to discuss about the issue-166

BernadetteLoscio: there is an ongoing discussion about this section
... on removing or not the vocabularies section

<annette_g> * oops, didn't see the whole interaction w/trackbot *

BernadetteLoscio: in this section we have more BPs related on publishing vocabularies, and only one on using vocabularies
... there is a discussion if we should have BP for publishing vocabularies or not

<RiccardoAlbertoni> bye nandana !

BernadetteLoscio: I discussed by email with antoine, we had some agreements, but we want know if the group agrees on keeping the section or not

laufer: when you say to remove the section, we need to remove the BPs of the section?

<antoine> this is a fair account, BernadetteLoscio

BernadetteLoscio: no. antoine and I have agreed on keep the section, but it's not about the vocabulary creation, but it's vocabulary publication

<antoine> also about re-use

<Seiji> +1 keep the section

BernadetteLoscio: not sure if BP14 should be there

laufer: I don't think BP14 is equal to the 'Re-use vocabularies'

<yaso> laufer has a point

laufer: for me vocabulary is a kind of dataset

<annette_g> examples?

laufer: we can't deal with vocabulary like a special kind of dataset

antoine: I disagree with laufer

<annette_g> +! that these are not datasets

<deirdrelee> +1

<scribe> scribe: Caroline

Giancarlo_Guizzardi: if you consider all the schemas it would be strange to make recommendations
... but if the focus is dataset
... the dataset may refers to one BP and define the terms
... it is a essencial aspect of quality of data to have these BPs
... we must say something about it

BernadetteLoscio: do you think we need to have BPs for the publication of this
... or we just keep the BP like "we should use vocabularies" and "we should use standards"

Giancarlo_Guizzardi: we should talk about the quality of relation between the datasets and what we use to annotation
... we should recommend that people use those, whatever they are

newtoncalegari: I agree on re-using vocabulary
... but it seems confusing on standarizing terms

phila: standarizing terms maybe not in a vocabulary, I personally woul keep it
... use them, re-use them and tell people what you have re-used is what we could keep

deirdrelee: if we are talking about vocabularies meaning standarized terms and also core vocabulaires or in sens of schemas

newtoncalegari: using standarized terms refering as vocabularies
... when you say it is like everyone using the same vocabulary could use FOAF, per example

<phila> newtoncalegari: Using standardised terms, referring to vocabs... if I use FOAF - should we always use the prefix 'foaf:'

BernadetteLoscio: it is not clear the difference vocabularies and standarized terms and also from core vocabulaires to other vocabularies

<phila> BernadetteLoscio: I don't know how to distinguish between core vocabs and others

<phila> BernadetteLoscio: We have BPs 15, 16, 17 and 19

BernadetteLoscio: we have 2 things: 1. BPs 15, 16, 17 and 19 are related to vocabularies publication

<phila> ... those are related to vocab publication.

BernadetteLoscio: 2. relation between 14 an 18

<phila> ... What is the relationship between 14 and 18

<phila> laufer: Is a code list a vocabulary?

<phila> scribe; Caroline

<ericstephan> If I cross my eyes I have no problems distinguishing between vocabularies and standardized terms. ;-)

<phila> scribe: Caroline

annette_g: I think vocabulary is not data
... if vocabularies were the same as data
... it is woth to try people how to be consistency when they use their own terms

<phila> +1 to annette_g

annette_g: if there is not a vocabulary you can try to be consistency

<Makx> There are two types of 'vocabularies'; a. predicate vocabularies (e.. Dublin Core terems) and b. value vocabularies (such as code lists, e.g. language codes)

ericstephan: the question about difficulties on semantic web that someone asked at Web.br
... I think one of the difficulties of adopting semantic web or existing vocabularies is ??
... I think having guidance on how people use vocabularies is helpful

Giancarlo_Guizzardi: I agree with phil
... use standarized terms makes sense
... and re-use vocabularies makes sense also
... BPs 14 and 18 may be miss interpreted

<Zakim> phila, you wanted to focus on Use Standardised Terms

Giancarlo_Guizzardi: we should be clear that we are talking about using vocabs to annotate metadada

phila: standarized terms could be written code lists
... maybe be not metadata

<antoine> for the record code lists are mentioned in BP 19

phila: to write it it is better to talk first about standarized terms and then code lists
... I would use BP 19
... the ones that are clearly about vocabs we could take them out

Giancarlo_Guizzardi: on BP 19 we should make it clear that we are not talking about vocabulary creation

<deirdrelee> ack Giancarlo_Guizzardi n

<phila> PROPOSED: That the BP on Use Standardized Terms be amended to talk about terms and code lists

<phila> draft prop - that

<phila> Best Practice 15: Document vocabularies, Share vocabularies in an open way , Vocabulary versioning be removed

laufer: we have a lot of communities that use some terms as standards for them

deirdrelee: to be part of a code list of vocabulary it doesn't have to be standarized

laufer: I don't think we have to use informal things
... but I don't know that our BPs are restricted to things that are only standarized

<phila> Community Standard is the usual term for something everyone uses that isn't a formal standard

<annette_g> I cannot hear anything

<phila> Examples include RSS, GTFS, robots.txt etc.

yaso: at netflix they classify films themselves

<annette_g> * just been about a minute*

yaso: they create a lot of relationships between the movies and what people write there
... they don't use W3C standards

<RiccardoAlbertoni> yes

<Makx> I think the term 'vocabulary' is really confusing!

yaso: are you saying that what netflix does would be that?

laufer: kind of

<annette_g> * I got it, local problem *

yaso: if I have a online news agency I could have standards there

Giancarlo_Guizzardi: the purpose of this BPs is to increase the interoperability
... to community to re-use vocabs

<Zakim> deirdrelee, you wanted to give example of informal codelist

Giancarlo_Guizzardi: we could capture that on these terms

deirdrelee: in Irland there are 4 or 5 vocabs valid for spacial references

BernadetteLoscio: we should define what we mean about standarized terms
... I propose we keep it and put code list
... considering also Giancarlo_Guizzardi suggestion's to show the consensus about a term

<Makx> And also define what we mean by 'vocabulary'

BernadetteLoscio: we can rewrite it considering these comments

<BernadetteLoscio> yes Makx ;)

<Giancarlo_Guizzardi> It think it is more about shared vocabularies than standard vocabularies, i.e., vocabularies that capture a consensus of the community the dataset refers to

<annette_g> * is someone speaking/ *

<RiccardoAlbertoni> could you write a proposal because I think I am a little lost ;)

<BernadetteLoscio> :)

<phila> Draft proposals:

<phila> - That Use Standardized Terms be amended to refer to code lists and other commonly used terms.

<phila> - That Document vocabularies , Share vocabularies in an open way, Vocabulary versioning be removed from the document.

<phila> - That Re-use vocabularies be retained

<phila> - That Choose the right formalization level be reviewed

phila: I have 4 proposals
... before we do all that, is that a consensus?

<laufer> +4

<annette_g> +4

+4

<yaso> +4 also (new kind of voting)

deirdrelee: from the external comments and feedback was anything about it?

<laufer> speed voting

BernadetteLoscio: no, only internal discussion for now

PeterWinstanley: I have to go! I see you tomorrow!

<annette_g> bye Peter!

<Giancarlo_Guizzardi> +4 (+1 (but we should make sure that we don't mean Use Standardized Terms in creating your vocabulary and we don't mean "use the right formalization level in creating...")

<Makx> -1

<Giancarlo_Guizzardi> and we don't mean re-use vocabularies in creating a vocabulary, etc...

RiccardoAlbertoni: I am ok with it
... but I would like to see something: "if you are defining your own vocab follow this"

yaso: are you saying that we should recommend if someone is defining a new vocab

<phila> The BP on re-using vocabs already points to the LD-BP document

RiccardoAlbertoni: we should at least adjust to follow the document that has been done in Linked Data Government group, per example

BernadetteLoscio: we can put this in the section introduction
... that there are other materials

annette_g: if you have to create a vocabulary, how are we mentioning something that already exists?
... if you don't find a existing vocab you could use something that already exists

RiccardoAlbertoni: if you have a data to publish and you have your own database and the schema is not mentioning what you are using
... you need to define more portion of a vocab
... as a publisher you have to make that undertandable
... that is why you suppose to publish your vocabulary
... and people can understand a specific attibute
... my suggestion is to put a link to Linked Data gov group because they already explain how to do data

<Zakim> deirdrelee, you wanted to sya that could be part of desc of bp18

RiccardoAlbertoni: if you don't want to go to linked data we can suggest that it could rely on something else

deirdrelee: I agree with RiccardoAlbertoni

<Makx> +1

deirdrelee: we can mention the vocabs to be seen

phila: there is already in the BPs
... it could be emphazied

<phila> RiccardoAlbertoni: Yes, it's there but we should emphasise that the doc talks about creating vocabs if they don't already exist.

Makx: BP 18 we use in a way that RDF uses vocabs
... in DCAT they talk in a different way
... sometimes we use vocabs differently
... my suggestion is that where we use vocabs as attibutes
... so people who are not familiar with linked data don't get confused
... I agree that on 14 we use the word standarized terms

laufer: when we are publishing data we have a BP to provide structural metadata
... a vocab like FOAF doesn't need to be explained
... if I will publish my own ontology the linked data WG can show how to do it
... we could put a link to this document
... I think we don't have to say how to do this

BernadetteLoscio: I think we can keep Phil's proposal and change the one to re-use the term vocab and say that the term vocab can be defined
... Makx do you agree?

<phila> draft 3 becomes... - That Re-use vocabularies be retained but the term vocabulary should be defined as a set of attributes

<phila> i.e. we get

<phila> Draft proposals:

<phila> - That Use Standardized Terms be amended to refer to code lists and other commonly used terms.

<phila> - That Document vocabularies, Share vocabularies in an open way, Vocabulary versioning be removed from the document.

<phila> - That Re-use vocabularies be retained but the term vocabulary should be defined as a set of attributes

<phila> - That Choose the right formalization level be reviewed

<yaso> Hi Sumit_Purohit :-)

<Sumit_Purohit> Hi Everyone.

<phila> PROPOSED: That Use Standardized Terms be amended to refer to code lists and other commonly used terms.

<yaso> +1

+1

<deirdrelee> +1

<antoine> +1

<phila> +1

<Seiji> +1

<annette_g> +1

<Makx> +1

<laufer> +1

<newtoncalegari> +1

<ericstephan> +1

<RiccardoAlbertoni> +1

<phila> RESOLVED: That Use Standardized Terms be amended to refer to code lists and other commonly used terms.

<AdrianoCesar-InWeb> +1+

<Sumit_Purohit> +1

<Giancarlo_Guizzardi> +1

<phila> PROPOSED: That Document vocabularies, Share vocabularies in an open way, Vocabulary versioning be removed from the document.

<deirdrelee> +1

<yaso> +1

<phila> +1

<Giancarlo_Guizzardi> +1

<laufer> +1

<Seiji> +1

+1

<annette_g> +1

<newtoncalegari> +1

<Makx> +1

<antoine> +1

<RiccardoAlbertoni> +1

+1

<phila> RESOLVED: That Document vocabularies, Share vocabularies in an open way, Vocabulary versioning be removed from the document.

<phila> PROPOSED: That Re-use vocabularies be retained but the term vocabulary should be defined as a set of attributes

<Makx> +1

<yaso> +1

<laufer> +1

+1

<phila> +1

<annette_g> +1

<Gisele> +1

<newtoncalegari> +1

<deirdrelee> +1

<antoine> 0

<Seiji> +1

<Sumit_Purohit> +1

antoine: I have been though this once
... if the group feels this should be there I am not opposing this

<Makx> Current text in intro says "According to W3C, vocabularies define the concepts and relationships (also referred to as “terms”) ..."

antoine: I am just warning it is not easy to do it

<Makx> Let's add (... "terms" or "attributes")

antoine: this section has 4 paragraphs trying to describe what vocab is

<phila> yaso: You're saying it will be difficult to re-write that BP. I wrote those 4 paragraphs introducing the section. It took several weeks - it's not easy

<Makx> +q

<ericstephan> +1 Makx

yaso: I was going to propose a extention of this definition

BernadetteLoscio: I don't know if we can define a vocab as a set of atributes

<RiccardoAlbertoni> +1 to Makx's proposal to add (... "terms" or "attributes") into "According to W3C, vocabularies define the concepts and relationships (also referred to as “terms”) ..."

BernadetteLoscio: Makx do you agree with the definition we have now?

Makx: on the first paragraph it could be add to refer as terms or atributes

<phila> Makx: Rather than terms, refer to terms and attributes

<phila> Draft 3 - - That Re-use vocabularies be retained but that it should refer to 'terms or attributes' to broaden the acceptance beyond the LD community

antoine: if the proposal is that only adding attributes I am fine with it

<antoine> +1

<Makx> +1

<RiccardoAlbertoni> +1

<laufer> +1

<phila> antoine: Happy if we're talking about adding a few words rather than rewriting

<adrianov> +1

<phila> PROPOSED: That Re-use vocabularies be retained but that it should refer to 'terms or attributes' to broaden the acceptance beyond the LD community

<yaso> +1

<antoine> +1

<deirdrelee> +1

<RiccardoAlbertoni> +1

<annette_g> +1

<phila> +1

<ericstephan> +1

<laufer> +1

<Seiji> +1

<BernadetteLoscio> +1

<newtoncalegari> +1

<Giancarlo_Guizzardi> +1

+1

<phila> RESOLVED: That Re-use vocabularies be retained but that it should refer to 'terms or attributes' to broaden the acceptance beyond the LD community

<phila> PROPOSED: That Choose the right formalization level be reviewed

<deirdrelee> +1

<yaso> +1

<annette_g> +1

<BernadetteLoscio> +1

<Gisele> +1

<laufer> +1

<phila> +1

<RiccardoAlbertoni> +1

<Makx> +1

<phila> RESOLVED: That Choose the right formalization level be reviewed

<deirdrelee> close issue-166

<trackbot> Closed issue-166.

<phila> close issue-166

<trackbot> Closed issue-166.

<yaso> o/

<phila> Bernadette is happy!

<Sumit_Purohit> OK....

<yaso> we can return in 5 minutes!!

<deirdrelee> 5 minute break (tea has arrived!)

<ericstephan> back from coffee run

<Sumit_Purohit> No one wants to come back from tea ??? :-)

<yaso> http://www.w3.org/2013/dwbp/track/issues/145

open issues/actions about identification

<deirdrelee> issue-145

<trackbot> issue-145 -- It makes sense to have a BP "Use unique identifiers"? -- open

<trackbot> http://www.w3.org/2013/dwbp/track/issues/145

<deirdrelee> issue-163

<trackbot> issue-163 -- Should the bp document refer to uris or identifiers -- open

<trackbot> http://www.w3.org/2013/dwbp/track/issues/163

<deirdrelee> issue-194

<trackbot> issue-194 -- Data Identification -- open

<trackbot> http://www.w3.org/2013/dwbp/track/issues/194

<annette_g> * what one are we on? *

<deirdrelee> all three annette_g

<AdrianoCesar-InWeb> BernadetteLos

<deirdrelee> grouped together under agenda https://www.w3.org/2013/dwbp/wiki/F2F_S%C3%A3o_Paulo_-_Agenda_Proposal_DWBP

<newtoncalegari> BernadetteLoscio: the issues on the agenda are grouped by topic

<deirdrelee> all issues about identification

<deirdrelee> BartvanLeeuwen: let's look at all three together

<deirdrelee> ... lots of discussion on mailing list around uris

<AdrianoCesar-InWeb> BernadetteLoscio: next one BP 194

<deirdrelee> s/bartvanleeuwen/bernadetteloscio

<AdrianoCesar-InWeb> BernadetteLoscio: about data identification section, talking about some messages (33) about this topic

<AdrianoCesar-InWeb> BernadetteLoscio: is this issue still opened?

<AdrianoCesar-InWeb> BernadetteLoscio: ask Annette about her opinion

<yaso> http://www.w3.org/2013/dwbp/track/issues/145

<AdrianoCesar-InWeb> annette_g: the first one is solved (145)

<deirdrelee> close issue-145

<trackbot> Closed issue-145.

<yaso> http://www.w3.org/2013/dwbp/track/issues/163

<deirdrelee> close issue-163

<trackbot> Closed issue-163.

<yaso> http://www.w3.org/2013/dwbp/track/issues/194

<AdrianoCesar-InWeb> yaso: 194 is about limiting this section to information that applies to publishing *data*

<phila> PROPOSED: That the BP document will use the term 'URI' throughout

<Makx> +1

<Makx> +1 to annette

<AdrianoCesar-InWeb> deirdrelee: decision is to use URI or not? Let´s clarify it

<phila> PROPOSED: That the BP document will use the term 'URI' throughout, unless there is a clear reason to use a differfnet term (URL, IRI etc.)

<ericstephan> +1 to consistent and intentional use of uri

<yaso> +1

<phila> PROPOSED: That the BP document will use the term 'URI' throughout, unless there is a clear reason to use a different term (URL, IRI etc.)

<annette_g> +1

<Sumit_Purohit> +1

<yaso> +1

<RiccardoAlbertoni> +1 to phila about keeping the information

<AdrianoCesar-InWeb> BernadetteLoscio: include the description of the terms and other point is to use this terms in the document

<AdrianoCesar-InWeb> phila: it is important to have definitions informed in the document

<Makx> -1 to Phil

<annette_g> -1

<Makx> +1 to Ivan

<Makx> Let's not go there

<phila> +1 makx, let's not

<phila> PROPOSED: That the BP document will use the term 'URI' throughout, unless there is a clear reason to use a different term (URL, IRI etc.)

<annette_g> +1

<Makx> +1

<antoine> +1

<laufer> +1

<Sumit_Purohit> +1

<BartvanLeeuwen> +1

<annette_g> * I keep seeing URI *

<ericstephan> +1

+1

<annette_g> * in W3C space *

<annette_g> -1 to IRI

<scribe> scribe: Caroline

annette_g: using countries specific IRIs I don't think is a BP to include that

phila: it is better to use URI considering data

annette_g: I think is a BP of everything on the Web

<Sumit_Purohit> +q

annette_g: I want to keep it as data

phila: We can stick as URI
... a lot of people will ready any say "you mean URL"

<scribe> scribe: AdrianoCesar-InWeb

<Makx> I second Annette and I am not American

<newtoncalegari> http://sãopaulo.gov.br/example/dataset

<newtoncalegari> http://saopaulo.gov.br

newtoncalegari: give an example of saopaulo.gov.br, the idea is to use an international format

<deirdrelee> sa0pa0l0.gov.br

<phila> +1 to the security issue

<phila> +1 _ I can't write Sao Paulo properly and easily without copying and pasting from somewhere else

phila: W3C suggests to use international format, avoid to use special characters

<newtoncalegari> +1 deirdree

<yaso> +1 to deirdrelee

annette_g: this is important for everything on the Web, not only for data

<Sumit_Purohit> +1 deirdrelee

<yaso> +1 to newtoncalegari proposal

newtoncalegari: it is important to justify the need to suggest this

<annette_g> in a tiny little footnote

<Makx> Let's just keep the intro of section 9.7 as it is.

<newtoncalegari> newton: we can use URI and warn about the security issue. what do you think?

laufer: We are recommending to use URI?

<newtoncalegari> maybe we should recommend to avoid using special characters

+1 to newtoncalegari

<Sumit_Purohit> need to leave for a meeting..Will be back.

<newtoncalegari> annette_g, not using special characters we tend to avoid some security issues, right?

<annette_g> right

<yaso> +1 to newton

<yaso> annette_g is not happy with the definition

<phila> PROPOSED: That the definitions of URI, URL and IRI be removed from the draft section 9.7

annette_g: propose to describe this definition

<phila> Other proposal is to use the term URI throughout

BernadetteLoscio: we can record that the group agree to use URIs

<phila> +1 not to discuss this any more

<phila> annette_g: This is about our own writing, not what other people should do

<yaso> PROPOSED: That the BP document will use the term 'URI' throughout

<phila> PROPOSED: That the definitions of URI, URL and IRI be removed from the draft section 9.7

<newtoncalegari> -1

<Makx> keep the second part

BernadetteLoscio: proposing to rewrite the introduction of the section

<annette_g> PROPOSED: That the definitions of URI, URL and IRI be removed from the draft section 9.7

Caroline: suggest to explain the definitions, but to explain all definitions

<yaso> sorry, annette_g

<yaso> speak slowly, please!

<phila> annette_g: Says this is as crazy as including a definition of antidisestablishmentarianism because we think it's cool

annette_g: there is no reason to define if we are not going to discuss in the document

<laufer> +1 to phil proposal

<annette_g> maybe we need to decide first whether we are going to include anything about identifiers

<yaso> +1 to newtoncalegari

newtoncalegari: someone is reading the document to learn, as a W3C document we need to inform the difference between URI, URL etc
... therefore prefer to keep this in the document

phila: I desagree with that because the definition can generate more discussion, since there is no clear definition about these terms

<Seiji> +1 to phil

<BartvanLeeuwen> +1 to phil

<annette_g> +1 to phil

<laufer> completely agree +1

+1 to phil

<newtoncalegari> +1 for that proposal

<phila> PROPOSED: That the definitions of URI, URL and IRI be removed from the draft section 9.7

<yaso> +1 to phil

<newtoncalegari> +1

<yaso> +1

<antoine> +1

<Caroline> +1

<Makx> +1

<BernadetteLoscio> +1

<phila> +1

<laufer> +1

<Seiji> +1

<Gisele> +1

<adrianov> +1

<RiccardoAlbertoni> +1

<phila> RESOLVED: That the definitions of URI, URL and IRI be removed from the draft section 9.7

yaso: next item

<phila> PROPOSED: That the BP document will use the term 'URI' throughout, unless there is a clear reason to use a different term (URL, IRI etc.)

<phila> +1

<yaso> +1

<annette_g> +1

<newtoncalegari> +1

<BernadetteLoscio> +1

<BartvanLeeuwen> +1

<deirdrelee> +1

<RiccardoAlbertoni> +1

<Seiji> 0

<Gisele> +1

<Caroline> +1

<Makx> +1

<yaso> -q

<laufer> +1 (to define what is a clear reason)

<ericstephan> +1

<phila> RESOLVED: That the BP document will use the term 'URI' throughout, unless there is a clear reason to use a different term (URL, IRI etc.)

<phila> annette_g: What's there now is not specific about data on the Web, it's about anything on the Web

phila: there is a confussion about what each of these terms represent...

<yaso> +1 to phila

<phila> annette_g: It's the bulletted list I object to

<phila> phila: It's gone

<phila> annette_g: You can't put something on the web without using a URI so it's pointless saying that you need to give datasets a URI

<annette_g> PROPOSED: that the best practice about identifiers be rewritten to address issues when posting data on the web.

<Zakim> phila, you wanted to try and squatre this circle

<phila> some draft text - When any resource is put on the Web, it has a URI. Many URIs are generated automatically but when sharing data, it is useful to bear in mind the following factors

<yaso> +1 to phil

<deirdrelee> http://w3c.github.io/dwbp/bp.html#DocumentVocabularies

<phila> http://philarcher1.github.io/dwbp/bp.html#UniqueIdentifiers

<newtoncalegari> "So what should I do? Designing URIs" http://www.w3.org/Provider/Style/URI.html

laufer: if an information does not have an URI, then it is not on the Web...

yaso: proposing to try to finish the discussion about this issue

BernadetteLoscio: we are going to use URI as identifier, ok? Yes
... now we are discussing the best practice, is a different issue, right?

yaso: suggest to annette_g to describe this issue about the best practice...

annette_g: I can try it, describing this issue

<ericstephan> +1 phila

<phila> ACTION: phila to take another run at the BP Use persistent URIs as identifiers [recorded in http://www.w3.org/2015/09/24-dwbp-minutes.html#action03]

<trackbot> Created ACTION-204 - Take another run at the bp use persistent uris as identifiers [on Phil Archer - due 2015-10-01].

BernadetteLoscio: we can close the 3 open issues... ok

<phila> close issue-145

<trackbot> Closed issue-145.

<phila> close issue-163

<trackbot> Closed issue-163.

<yaso> http://w3c.github.io/dwbp/bp.html#dataVersioning

<phila> close issue-194

<trackbot> Closed issue-194.

yaso: next topic - Discuss the versioning section and resolve open issues (30 min.)

<yaso> ISSUE-193

<trackbot> ISSUE-193 -- Data Versioning -- open

<trackbot> http://www.w3.org/2013/dwbp/track/issues/193

BernadetteLoscio: in the last Draft (last F2F) we discuss about data versioning...

there was a diagram and we had discussed by email about this... about the meaning of a versioning

We agree that time series is not a case of versioning...

We try to explain better the meaning for versioning for this document

<yaso> http://www.w3.org/2013/dwbp/track/issues/193

<BernadetteLoscio> http://w3c.github.io/dwbp/bp.html#dataVersioning

<phila> http://philarcher1.github.io/dwbp/bp.html#dataVersioning

BernadetteLoscio: annette_g, do you agree with this proposal? Look in agenda, item 193
... one thing is our definition of data versioning

<phila> s/ http://philarcher1.github.io/dwbp/bp.html#dataVersioning//

<Caroline> https://www.w3.org/2013/dwbp/wiki/F2F_S%C3%A3o_Paulo_-_Agenda_Proposal_DWBP

<phila> http://w3c.github.io/dwbp/bp.html#dataVersioning

<Caroline> http://w3c.github.io/dwbp/bp.html#dataVersioning

<BernadetteLoscio> http://w3c.github.io/dwbp/bp.html#dataVersioning

<yaso> close issue-193

<trackbot> Closed issue-193.

yaso: is it ok to close this issue (193)?

<ericstephan> I have to leave in 20 minutes unfortunately...

<Makx> I need to leave in 20 minutes, dinner time

<RiccardoAlbertoni> I am going to stay only for 10/20 minutes

<Seiji> need to leave

<yaso> so we are going for more 20 min

<Caroline> scribe: WagnerMeiraJr

<phila> issue-168?

<trackbot> issue-168 -- Dataset versioning -- open

<trackbot> http://www.w3.org/2013/dwbp/track/issues/168

yaso: Going to issue 168. Newton, can you start?

newtoncalegari: I was working on this issue in the 1st part of the meeting: which vocabulary to use in versioning?

BernadetteLoscio: I'd like to know what is the suggestion for using when defining versions?

yaso: Any other suggestions?

phila: We cannot make a normative dependency. It is one possible way, but I do not want to do it.

BernadetteLoscio: It is just to give an example.

<annette_g> link?

<phila> PAV

yaso: I think that it is ok to use for sake of an example section.

<antoine> +1 to use Memento as a (quite different) alternative to PAV

newtoncalegari: There is an issue (3), the beginning of data version that motivates it.

yaso: Seems ok to me.

BernadetteLoscio: It is not just because it is an open issue that it is worth doing it. We will close issues 92 and 68. What's the opinion of the group regarding changes in the data. It is not clear whether updating the schema is a new version or not. Does this new attribute justify a new version?

<newtoncalegari> q

antoine: Not sure I understood the point.

<annette_g> +1 to letting the publisher decide

antoine: In the case you mentioned it does not sound to me it is the case to create a new version.

BernadetteLoscio: Sometimes the published does not know whether it is the case of creating a new version.

antoine: Let the publisher decide.

Giancarlo_Guizzardi: If we use the vocab and onto, there could be changes in them that do not change the semantics, but if the latter change, it should be anew evrsion.

Seiji: Same comment of Giancarlo_Guizzardi

<Zakim> phila, you wanted to say we shouldn't define when a new version is a new version

BernadetteLoscio: The publisher may decide, but we at least sign about it.

phila: When does something change enough? It is such a difficult question.

<Makx> yes, it's a can of worms

BernadetteLoscio: We tried to define based on the discussions.
... We should give some guidance to the publisher. If the definition is not good enough, we should not do it.

<phila> Mind you, I like the text in http://w3c.github.io/dwbp/bp.html#dataVersioning

<Makx> +q

laufer: The term version is used with several meanings. It also varies depending on the language. We need a way to talk about relations among datasets. People saying that one thing is a version of other is not just because it is a change, they may be in completely different languages.

BernadetteLoscio: There is a confusion among several terms that are used differently: version, distribution etc
... A time series is not a new version of a dataset. We are trying to define and let the publisher to decide.
... This new dataset is derived or is a version of other dataset. If the publisher does not know what is a version how come he will decide what to do.

annette_g: A language is not version specific. We may need to perform the same change in several languages at once.
... If you make a change in a dataset, you may create a new version depending on the publisher.
... It is sort of a editorial work to make a change and evaluate how different it is.

laufer: I'm not saying that different languages are different versions. We have to define our concept of version.

antoine: Imagine your updates are not used by the dataset in question, then it does not produce a new version of the dataset.
... What are the principles we want to follow when discussing versions? Should we discuss these?
... Versioning means that you change something that affects your dataset.

Makx: It is very hard to define what a version is for a particular person.
... They should not throw away old data upon a new version. People may be using it.

phila: Agree with Makx . I like the text and want to add: we can't antecipate everything and tell when a new version makes sense. I would encourage consistency. It may be every six weeks, publishing a new version.

<Makx> +1 to Phil

phila: They may decide and stick to it.

Giancarlo_Guizzardi: I agree that it is useful to give guidelines, including handling deprecated content and vocab changes. I agree that defining it is very interesting. But it is extremely hard. We tried recently to characterize versions for software and it is hard.

<Makx> We discussed this for a long time for the DCAT-AP in Europe and could not resolve it ;-(

laufer: I don't know whether it is feasible, but you may encourage the publisher to define what is a version.

<Makx> We were hoping that DWBP could give guidance ;-)

<ericstephan> Talk to you tomorrow everyone. Sorry to take off in the middle of discussion.

BernadetteLoscio: It is interesting and I agree that it is hard to define and I was trying to write about. I'm going to rewrite it considering our discussion here, towards help the publisher to decide. I'll rewrite the introduction and ask for feedback.

<BernadetteLoscio> bye bye Eric!

newtoncalegari: How to version data stream?
... DCAT uses dcterms:modified.
... A change in data stream will be treated as such just when the schema changes.

<annette_g> 2 separate issue

<annette_g> issues

<Makx> Even that is not agreed by everyone, Deirdre

deirdrelee: Just when the API changes?

<Makx> Apologies, I have to sign off. Will be back tomorrow.

BernadetteLoscio: Newton's concern is that you have a stream and there should not be a new version.

<deirdrelee> deirdrelee: if the API changes it doesn't matter, because an API is related to data access, not data structure or content

<BernadetteLoscio> bye bye Makx!!!

annette_g: API related versioning is really hard to define.

newtoncalegari: Don't we need to worry to track changes in data streams?

yaso: It does not seem to be necessary.

<yaso> http://www.w3.org/2013/dwbp/track/issues/192

yaso: We were discussing 192.

newtoncalegari: We are not going to recommend and we may close the issue.

<yaso> close issue-192

<trackbot> Closed issue-192.

newtoncalegari: We have no examples.

<phila> There is an alternative to PAV at http://www.w3.org/TR/vocab-adms/ but I prefer PAV which is better developed

<yaso> close ISSUE-168

<trackbot> Closed ISSUE-168.

yaso: It looks like we are done for today.
... Thanks everyone.

<annette_g> I won't be on until 6 PT

<annette_g> which is 10

<phila> Thanks everyone. Good night/good afternoon

<yaso> neither us, annette_g

deirdrelee: Is it fine to start at 8AM Sao Paulo.

<yaso> so tomorrow is 8:00 am

<laufer> bye annette... good lunch...

yaso: We will start at 8:00 AM.

<annette_g> bye!

<RiccardoAlbertoni> Enjoy your staying in sao paolo.. thanks for the interesting discussions , bye

<laufer> bye all...

<yaso> Thanks, RiccardoAlbertoni and annette_g and others attending remote :-)