Data on the Web Best Practices Working Group Teleconference

<annette_g> Hi hadleybeeman

<hadleybeeman> Good morning, Deirdrelee, laufer and Ig_Bittencourt

<hadleybeeman> Trackbot, start meeting

<Ig_Bittencourt> Good Morning hadleybeeman

<laufer> good morning, Hadley

<deirdrelee> hi annette_g we have an electricity cut here!

<deirdrelee> hi hadleybeeman

<hadleybeeman> No electricity, Deirdrelee?! Are you all operating on laptop batteries?

<deirdrelee> but wifi (o)

<hadleybeeman> Morning, newton. And thanks for the explanation :)

<BernadetteLoscio> link to DWBP open issues: https://www.w3.org/2013/dwbp/wiki/F2F_Zagreb_-_Agenda_Proposal_DWBP#Open_Issues_to_be_closed_during_the_F2F_Zagreb

<deirdrelee> scribe: deirdrelee

<Caroline_> Hello! We have no audio for now :(

<phila> meeting: DWBP Face to Face Day 2

yaso_: we will start with discussion with sdw

<Caroline_> we will start disscusion on DWBP and SDW

<yaso_> yes, Caroline_

BernadetteLoscio: we are collaborating with sdw around best practices, we had some meetings together around commonalities

<newton> Current SDW Editor's draft: http://w3c.github.io/sdw/bp/

BernadetteLoscio: we discussed how best to manage this overlap
... we will finish this discussion now

<phila> scribe: phila

<deirdrelee> ... for resource identifiers, there is some intersection, and for apis,

yaso_: We're going to start by looking at the discussion with the SDW

<deirdrelee> ... we'd like opinion of group of how to deal with this

<deirdrelee> Caroline_: there are 3 BPs in DWBP - BP11 we use persistent URIs

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

<BernadetteLoscio> sdw BP: https://www.w3.org/TR/sdw-bp/#bp-identifiers

<deirdrelee> ... SDW their BP1 says you should use persistent identifiers

<deirdrelee> ...this could make reference to our bP 11

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

<deirdrelee> ...we discussed that they could keep their BPs but that they could make reference to ours

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

<deirdrelee> ...the second one that matches, our bp12 - Use persistent URIs as identifiers within datasets

<BernadetteLoscio> SDWBP: https://www.w3.org/TR/sdw-bp/#reuse-ids

<deirdrelee> their BP has a similar one

<deirdrelee> ... on reuse ids. but it also points to that there are some things within our bp that could be rewrote

<deirdrelee> ... i'm suggesting we rewrite our bp slightly

<deirdrelee> ... and they could reference our bp, but go deeper as well

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

<BernadetteLoscio> SDWBP: https://www.w3.org/TR/sdw-bp/#ids-for-changing-things

<deirdrelee> ... another bp is our bp 13 Assign URIs to dataset versions and series

<deirdrelee> ... which matches their bp 4

<deirdrelee> ... their bp 4 already reference our bp on versioning

<deirdrelee> ... hthey have 5 bps on identifiers

<deirdrelee> ... so other ones that are for only spatial

<deirdrelee> ... however out of scope for us. we recommend they keep their ones, but we don't have to include

<deirdrelee> ... there is one sdw bp 5, provide identifiers for large information resource

<deirdrelee> ... i recommend that we should include this

<deirdrelee> BernadetteLoscio: question is whether they should be more specific and refer to our more generic ones

<deirdrelee> ... they should focus more on spatial, in more detail

<deirdrelee> yaso_: so there is one that may be changed

<deirdrelee> BernadetteLoscio: we as a group, should give an answer to sdw group

<Caroline_> Best Practice 5: Provide identifiers for parts of larger information resources

<deirdrelee> Ig_Bittencourt: you say bp3 we shouldn't use, because its more specific, but bp5 we will look at

<deirdrelee> BernadetteLoscio: yes, this is related to subsetting discussion, if we have that bp, we can ask sdw to reference it

<deirdrelee> laufer: an identifier is identifying a resource. you can access a subset in two ways, either by a uri or an api.

<deirdrelee> ... we don't say this is an api

<deirdrelee> ... when we talk about we should have access to an api

<deirdrelee> ... how our bp will be, it will be a joint of two of our existing bps

<deirdrelee> Caroline_: it's just a matter of seaparating two things

<deirdrelee> ... let's go to sdw and say our bps 11, 12 and 13 match to their 1, 2, 4

<deirdrelee> PROPOSED: To say to SDWG that 11, 12 and 13 match to their 1, 2, 4

<deirdrelee> Ig_Bittencourt: there are other BPs that are there as well, that we should look at

<deirdrelee> PROPOSED: To say to SDWG that their BPs 11, 12 and 13 match to our 1, 2, 4, but that it would be a good idea if they could be more specific on spatial, as our BPs are more general

<deirdrelee> (this is only a start, we will say them more to)

<deirdrelee> newton: we shouldn't vote now because phila is busy

<deirdrelee> Caroline_: annette_g can you look at proposal

<annette_g> I'm fine with it

<Caroline_> Annette, sorry but the mic is terrible here

<yaso_> annette_g, here the same

<deirdrelee> PROPOSED: To say to SDWG that their BPs 1, 2 and 4 match to our 11, 12, 13, but that it would be a good idea if they could be more specific on spatial, as our BPs are more general [08:36] <newton> q+

<yaso_> hadleybeeman we can hear you

-> http://w3c.github.io/sdw/bp/ is at SDW BP

<laufer> +1

<Caroline_> we have the mic off hadleybeeman and annette_g

<Caroline_> sorry

<Caroline_> it is too bad here

<annette_g> yes, I hear you Newton

<annette_g> HI

<hadleybeeman> hi

<Caroline_> we don't hear you know

<hadleybeeman> yes, we can hear you :)

<Caroline_> when you are going to say something we will turn on the speaker again

<yaso_> +1

<deirdrelee> +1

<Ig_Bittencourt> +1

<BernadetteLoscio> +1

<Caroline_> +1

+1

<annette_g> +1

<PWinstanley> +1

<newton> +1

<hadleybeeman> +1

RESOLUTION: To say to SDWG that their BPs 11, 12 and 13 match to our 1, 2, 4, but that it would be a good idea if they could be more specific on spatial, as our BPs are more general

<BernadetteLoscio> https://www.w3.org/2013/dwbp/wiki/F2F_Zagreb_-_Agenda_Proposal_DWBP#Open_Issues_to_be_closed_during_the_F2F_Zagreb

<deirdrelee> yaso_: we are talking about subsetting now, because this will feed back into SDW discussion

<deirdrelee> BernadetteLoscio: we had discussion yest about subsetting data. look at open issues in link

issue-160?

<trackbot> issue-160 -- Should we add at BP about subsetting data? -- open

<trackbot> http://www.w3.org/2013/dwbp/track/issues/160

<deirdrelee> ... there are 3 issues, first one 160

<deirdrelee> ... added by annette_g should we add a new bp

issue-208?

<trackbot> issue-208 -- What to say about subsetting data, and identifying those subsets. this to be done in conjunction with sdw wg about this issue -- open

<trackbot> http://www.w3.org/2013/dwbp/track/issues/208

<deirdrelee> ... discussion in last meeting, but i wasn't there

<deirdrelee> ... first thing is to decide if we're going to include this bp in our document

<yaso_> This is the Best Practice proposed http://w3c.github.io/dwbp/bp.html#EnableDataSubsetting

<deirdrelee> ... if we do we can go back to SDW to link to theirs

<deirdrelee> ... otherwise we see it's out of scope

<deirdrelee> ... i tried to look if their was resolution in last meeting's minutes, could someone who was there report?

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

<deirdrelee> hadleybeeman: to answer BernadetteLoscio's question on previous discussion, it didn't end with a a decision, it ended with annette_g having to write the bp to bring back to the group

<deirdrelee> ... which she did send on

<deirdrelee> phila: to clarify, we have some general wording from annette_g on subsetting which is generic

<deirdrelee> ... I think it has to be generic because it's such a big thing

<deirdrelee> ... sdw will specialise that, that's exactly what we want to be done

<yaso_> This is the Best Practice proposed http://w3c.github.io/dwbp/bp.html#EnableDataSubsetting

<newton> @annette_g did you see the comment I left on your PR? https://github.com/w3c/dwbp/pull/319

<deirdrelee> laufer: the BP says that data is used in the most flexible way. if we accept this bp, we should rethink about API access

<annette_g> @newton, no, I will look

<newton> I haven't merged all commits of that PR, I merged two of them: data enrichment and data subsetting BPs

<deirdrelee> ... we don't know what the granularity is

<annette_g> @newton, okay

<annette_g> The BP as written is not just about APIs. The example is an API one, but that doesn't mean the BP is not about other things

<newton> @annette_g you mean the BP 21?

<annette_g> yes

<hadleybeeman> I agree with Laufer -- but also, I think the current (proposed) text covers more than just APIs

<deirdrelee> laufer: I'm not saying that the bp states that only apis are used to access subsets

<annette_g> using an API is also a set of defined URIs

<laufer> I do not agree, annette. It could be. But is not mandatory.

<deirdrelee> BernadetteLoscio: I would like to propose that we review the bp, we leave 3 issues open, then we vote if we should keep the bp

<annette_g> true

<hadleybeeman> Seems sensible, BernadetteLoscio.

<annette_g> @laufer, did you want to suggest we add a sentence?

<deirdrelee> ... it's difficult now, maybe laufer could work with annette_g

<laufer> For me an API is commonly understand as provided by an unique URI with parameters

<deirdrelee> yaso_: it's in the official draft, should we vote

<deirdrelee> BernadetteLoscio: np because its's just a draft

<annette_g> I'm not opposed to adding a sentence to clarify Laufer's point

<Caroline_> great, annette_g!

<laufer> Ok, annette. My point is that providing a set of fixed URI´s is also a way to access. Even template URIs.

<hadleybeeman> Vote by email is a good idea. Or schedule an extra call.

<laufer> I vote by voting in an extra call.

<annette_g> tomorrow wednesday?

<deirdrelee> deirdrelee: consious of time frame, we should have the decision soon

<yaso_> next

<Caroline_> next week

<hadleybeeman> Wednesday the 23rd and Weds the 30th of March?

<deirdrelee> not tomorrow annette_g

<annette_g> okay

<Caroline_> yes, hadleybeeman

<newton> deirdrelee: is suggesting to reschedule the meeting for another week day, because we can't have meetings on the next two fridays

Wednesday 23

<newton> ... one we'll be traveling, and another one is Easter

<deirdrelee> deirdrelee: we'll have same time as friday's call

And Wednesday 30th

<Caroline_> same time we have Friday's call

<deirdrelee> PROPOSED: Move meetings for next two weeks on Wednesday instead of Friday, as there's important decisions to make, and there's holidays coming up

<hadleybeeman> I propose we stick to Eastern Time.... as usual.

<deirdrelee> phila: same time gmt or eastern standard, because america just put clocks forward

<deirdrelee> Caroline_: what's best for annette_g and eric,

<yaso_> +1

<BernadetteLoscio> +1

<Caroline_> +1

<deirdrelee> +1

<laufer> +1

<annette_g> GMT is better, I think

<newton> +1

<annette_g> 7 is better than 6 for me

RESOLUTION: Move meetings for next two weeks on Wednesday instead of Friday, as there's important decisions to make, and there's holidays coming up

-> http://www.timeanddate.com/worldclock/fixedtime.html?msg=DWBP+Call&iso=20160323T14&p1=1440&ah=1 Time for 23 March

<Caroline_> then we keep as it is?

<deirdrelee> phil: Time will be at 14:00 GMT next wed and following wed, i'll send around

<newton> I wonder if I have a dataset with all cities of Sao Paulo state. I publish that dataset and I decided to sort the cities by name, so I will have a few datasets sorted by letter. Are those subsets of the main dataset or are they only datasets?

<deirdrelee> PROPOSED: We will vote on inclusion of BP on subsetting data on meeting of 23rd March

<deirdrelee> +1

<yaso_> +1

<newton> +1

<annette_g> +1

<BernadetteLoscio> +1

<Ig_Bittencourt> +1

<laufer> +1

RESOLUTION: We will vote on inclusion of BP on subsetting data on meeting of 23rd March

<hadleybeeman> +1

<annette_g> @newton, that is another dataset, not a subset

API open issues

<annette_g> sure, what do we need to discuss

<deirdrelee> https://www.w3.org/2013/dwbp/wiki/F2F_Zagreb_-_Agenda_Proposal_DWBP#Open_Issues_to_be_closed_during_the_F2F_Zagreb

<deirdrelee> newton: I raised the issues when annette_g and eric wilde were working on api bps

<annette_g> that all got rewritten along the lines you were suggesting

<deirdrelee> ... bp included that apis should use REST, new

<yaso_> https://www.w3.org/2013/dwbp/track/issues/213

issue-213?

<trackbot> issue-213 -- Changes on BP about REST APIs -- open

<trackbot> http://www.w3.org/2013/dwbp/track/issues/213

<annette_g> it's open, but I think it can be closed

<deirdrelee> yaso_: can we close this issue?

<newton> @annette_g but if I retrieve all cities of letter A by an API, is it a subset or a whole dataset?

<deirdrelee> BernadetteLoscio: i would like to know, after the changes on bp on api, if these issues are still valid or not?

<deirdrelee> ... we had a lot of discussion if these issues are still valid

<annette_g> @newton, that is a subset

<annette_g> I think we can close that. We rewrote to make REST an implementation

<BernadetteLoscio> ISSUE-213: Changes on BP about REST APIs

<trackbot> Notes added to ISSUE-213 Changes on BP about REST APIs.

<deirdrelee> CLOSE ISSUE-213

<trackbot> Closed ISSUE-213.

<deirdrelee> It's addressed

<deirdrelee> issue-2233

<trackbot> Sorry, but issue-2233 does not exist.

<deirdrelee> issue-233

<trackbot> issue-233 -- Amendments to BP: Use Web Standardized Interfaces -- open

<trackbot> http://www.w3.org/2013/dwbp/track/issues/233

<deirdrelee> yaso_: can this be closed?

<annette_g> I think so

<BernadetteLoscio> https://www.w3.org/2013/dwbp/track/issues/233

<Caroline_> thank you annette_g! :)

<deirdrelee> close issue-233

<trackbot> Closed issue-233.

<deirdrelee> issue-242

<trackbot> issue-242 -- APIs on the Web (for publishing Data on the Web) Best Practices -- open

<trackbot> http://www.w3.org/2013/dwbp/track/issues/242

<deirdrelee> close issue-242

<trackbot> Closed issue-242.

<deirdrelee> newton: issue-233 was raised by phila

<annette_g> right

<deirdrelee> issue-195

<trackbot> issue-195 -- Provide data up to date -- open

<trackbot> http://www.w3.org/2013/dwbp/track/issues/195

<newton> @annette_g so the difference isn't in the content, but the difference is in the way the content is retrieved?

<annette_g> I think it's okay now

<annette_g> yes

<deirdrelee> close issue-195

<trackbot> Closed issue-195.

<deirdrelee> issue-240

<trackbot> issue-240 -- To consider a new bp around numeric accuracy. comes from sdw's concern about useless extra decmimal places -- open

<trackbot> http://www.w3.org/2013/dwbp/track/issues/240

<Caroline_> great!

<newton> https://lists.w3.org/Archives/Public/public-dwbp-wg/2016Mar/0120.html

<yaso_> issue-248

<trackbot> issue-248 -- To change the Feedback section to Data Reuse section. -- open

<trackbot> http://www.w3.org/2013/dwbp/track/issues/248

<deirdrelee> phila: we discussed this issue yesterday, i got an action to include in the bp doc a reference to example from dqv that antoine and ricardo are writing

<deirdrelee> close issue-240

<trackbot> Closed issue-240.

<Caroline_> https://www.w3.org/2013/dwbp/track/issues/248

<deirdrelee> BernadetteLoscio: annette_g sent a message about including a new BP abour data reuse

<Caroline_> the link above is the one BernadetteLoscio is talking about

<newton> Annette's message https://lists.w3.org/Archives/Public/public-dwbp-wg/2016Mar/0087.html

<deirdrelee> ... editors thought that bp on feedback could be renamed to bp on reuse

<deirdrelee> ... we brainstormed on better name for section

<deirdrelee> ... we could change the name to data usage, because then there's a direct link to the vocab we're devleoping

<Caroline_> deirdrelee: why do we need this about data usage?

<Caroline_> ... we talked aobut reusing data is a benefit for all BPs

<yaso_> deirdrelee: reuse is a BP or a Benefit?

<annette_g> I think the BP itself is important

<annette_g> to give balance to the ideas for publishers to do

<annette_g> this is soemthing for people reusing other people's data to do

<annette_g> reusing in this sense is another type of publishing

<deirdrelee> BernadetteLoscio: if we have this bp, we should focus on dataset citation

<annette_g> the BP for feedback is different, it tells the original publisher to accept feedback from users

<annette_g> This tells users to use that

<deirdrelee> ... we already have bp on feedback, we could rewrite this bp to include ref to citation and not be so generic

<newton> https://github.com/w3c/dwbp/pull/319

<annette_g> citation is only one part of their responsibility

<annette_g> why so much resistant to a new section?

<annette_g> We don't need to scrunch it into a section if doesn't belong with just to avoid creating a new section

-> http://agreiner.github.io/dwbp/bp.html Annette's version of the BP doc

<deirdrelee> laufer: the bps are about publishing data, not about using

<annette_g> this is for re-use, not use

<annette_g> re-use is re-publishing

<deirdrelee> ... if we talk about this, then we'll have to talk about a lot of other things

<annette_g> re-use is publishing

<deirdrelee> BernadetteLoscio: we had a lot of discussions about this, we decided bps should be for data publisher, not data users

<deirdrelee> PWinstanley: the concern i have, is when we think about data we think about tables, we haven't gotten into media files, etc. there is lots more to think about around this

<yaso_> annette_g, when you're in the q, can I ask you to put an "]" when you finishing typing?

<annette_g> give an example of reuse that is not publishing something for other people to use

<annette_g> I think we can just use the one BP

<annette_g> that's all that's needed to be covered for this.

<BernadetteLoscio> and to create a new section?

<annette_g> I feel strongly that this side of the conversation should be addressed

<annette_g> see our charter for why

<deirdrelee> deirdrelee: close topic for now, vote on wed 23rd

<annette_g> I've already done the work in the pull request

<deirdrelee> BernadetteLoscio: agree we vote next week, updating section requires a lot of work, because it messes with structure of doc

Personally, I'll be voting in favour. I like what annette_g has written

<newton> but the name of the section is Data Re-use... Deirdree concerns about this name as well

<annette_g> let's not leave out useful info because we don't want to do the work

<newton> http://agreiner.github.io/dwbp/bp.html#Re-use

<deirdrelee> BernadetteLoscio: if we'd like to add a section, we need corresponding challenge, benefits, etc.

<deirdrelee> ... would need help

<annette_g> we already have benefits in the BP

<annette_g> we can add the challenge

<hadleybeeman> I'm with bernadette -- in that: it is good to tie our BPs back to use cases. To demonstrate we haven't just made them up.

<deirdrelee> newton: the new section is called data-reuse, talked to dee yesterday, she was concerned with name

<deirdrelee> phila: annette_g has written new section, think it's useful and well-written

<yaso_> +1 too phila

<annette_g> I've already added requirements in the new content

<deirdrelee> BernadetteLoscio: if we add new bp, we are going to add a new challenge, and change the use-case doc?

<deirdrelee> yaso_: not necessary, we can use existing challenge

<laufer> I vote for voting next wed

<deirdrelee> PROPOSED: We will vote on inclusion of bp on reuse during next call on wed 23rd

<deirdrelee> +1

<yaso_> +1

<newton> -1

<annette_g> +1

<BernadetteLoscio> +1

<laufer> +1 to +1

<Ig_Bittencourt> +1

<PWinstanley> +1

RESOLUTION: We will vote on inclusion of bp on reuse during next call on wed 23rd

<hadleybeeman> +1

<hadleybeeman> s/resolved/UNRESOLVED

<deirdrelee> newton: wants to know where this bp should be included?

<deirdrelee> ... a new section or in an existing section

<annette_g> I mostly just want it to go in

<annette_g> then we can decide where

<deirdrelee> laufer: we will discuss this during week

<annette_g> I agree the title "Re-use" is problematic, we should discuss via email other ideas

issue-249?

<trackbot> issue-249 -- Discussion about fragment identifiers -- open

<trackbot> http://www.w3.org/2013/dwbp/track/issues/249

<deirdrelee> issue-249

<trackbot> issue-249 -- Discussion about fragment identifiers -- open

<trackbot> http://www.w3.org/2013/dwbp/track/issues/249

<newton> @annette_g ok

<deirdrelee> BernadetteLoscio: we raissed this issue in relation to eric wilde's comments on comment tracker, we weren't sure what to tell him

<deirdrelee> ... link to comment tracker, and link to his message in the issue

<annette_g> he said that he had dropped the issues because he got no response

<deirdrelee> Caroline_: what do you mean by fragmented identifiers

<Caroline_> he had a response

<annette_g> no, he gave up

<annette_g> on us responding

<BernadetteLoscio> https://lists.w3.org/Archives/Public/public-dwbp-comments/2015Jul/0002.html

<deirdrelee> BernadetteLoscio: he had an answer, i had a discussion with him

<deirdrelee> phila: in his world fragmented identifer don't exist

<deirdrelee> ... that's his issue i guess

<deirdrelee> newton: you only use frag identifers inside environment

<deirdrelee> ... i think he was tlaking about content negotiation, when you add the file extension, e.g. .rdf or .ttl

<deirdrelee> phila: nothing to do with frag extensions

<annette_g> He's talking about providing ids that others can link to. He says "*linkable* means more than just URIs. "

<annette_g> right

<hadleybeeman> We can refer to this: https://www.w3.org/DesignIssues/Fragment.html

<deirdrelee> phila: he's saying, if you provide data in multiple formats, frag extensions should be common across them. that's fine

<deirdrelee> ... i can take an action to write it

<hadleybeeman> phila, there is a section in https://www.w3.org/DesignIssues/Fragment.html entitled "Fragment IDs and Content negotiation - known bug"

<hadleybeeman> might be the source of Eric Wilde's comments?

<deirdrelee> action for phila to add line on frag identifiers in Best Practice 15: Provide data in multiple formats

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

<deirdrelee> action phila to add line on frag identifiers in Best Practice 15: Provide data in multiple formats

<trackbot> Created ACTION-265 - Add line on frag identifiers in best practice 15: provide data in multiple formats [on Phil Archer - due 2016-03-22].

<annette_g> in a hypertext API, there are links, too

<deirdrelee> close issue-249

<trackbot> Closed issue-249.

<deirdrelee> issue-250

<trackbot> issue-250 -- Should we create a new section to include the namespaces that are being used in the examples? -- open

<trackbot> http://www.w3.org/2013/dwbp/track/issues/250

<annette_g> what namespace is that?

<deirdrelee> BernadetteLoscio: antoine sent a message that namespace should be documented somewhere

<deirdrelee> ... what is the best way to include this

<newton> @annette_g all namespaces used in the document

<deirdrelee> phila: it's a two minute job, easy

<annette_g> ooooh, okay

<Caroline_> :)

<newton> @annette_g are you ok with that?

<deirdrelee> phila: just to add a namespace table

<deirdrelee> action BernadetteLoscio to include namespace table at top

<trackbot> Created ACTION-266 - Include namespace table at top [on Bernadette Farias Loscio - due 2016-03-22].

<deirdrelee> close issue-150

<trackbot> Closed issue-150.

BP Status table

<Caroline_> https://www.w3.org/2013/dwbp/wiki/BP_Tasks_for_CR

<newton> https://www.w3.org/2013/dwbp/wiki/BP_Tasks_for_CR

<deirdrelee> BernadetteLoscio: we created a table with tasks that still have to be completed

<deirdrelee> ... these tasks are based on open actions and things we discussed during F2F

<newton> http://w3c.github.io/dwbp/bp-status.html

<Caroline_> scribe: Caroline_

<deirdrelee> newton: I just merged antoine's pull request

<deirdrelee> ... there are still some bps that are open

<scribe> scribe: deirdrelee

<yaso_> hadleybeeman, are you there?

<hadleybeeman> @newton, no I haven't yet. Was hoping to do that today.

UNKNOWN_SPEAKER: question for hadleybeeman if she has reviewed bp 22 yet

<BernadetteLoscio> https://docs.google.com/document/d/1HCPCDD5QbSVQrlAI2v6HT6iF_-N-M8KuTysgh7ZrTaM/edit?ts=56e73e23#

<annette_g> sorry to go backward in subject, but if people want to see the reuse bp, they can go to http://agreiner.github.io/dwbp/bp.html#Re-use

<annette_g> just to give the link

<newton> thanks

<yaso_> tks, annette_g

<BernadetteLoscio> thanks!

newton: when we reviewed doc, some of the intended outcomes are not so clear

<phila> s/gand/grand

an overview of if intended outcomes and tests match are here https://docs.google.com/document/d/1HCPCDD5QbSVQrlAI2v6HT6iF_-N-M8KuTysgh7ZrTaM/edit?ts=56e73e23#

<Zakim> deirdrelee, you wanted to talk about bp6

Caroline_: for us, it's clear the how to test should match with intended outcome, this should directly reflect

<Zakim> phila, you wanted to go back a bit

<phila> action-239?

<trackbot> action-239 -- Peter Winstanley to Check on using an example about real-time data for bp23 -- due 2016-03-25 -- OPEN

<trackbot> http://www.w3.org/2013/dwbp/track/actions/239

<BernadetteLoscio> https://www.w3.org/2013/dwbp/wiki/BP_Tasks_for_CR

BernadetteLoscio: this table is for editors, just so people know what we're doing

<annette_g> The table needs to be updated

BernadetteLoscio: we will work on this starting tomorrow

PWinstanley: what i was proposing is that rivers data, we have detailed response from dave reynolds
... eric emailed this morning to say woopdedoopdedo

<annette_g> lots of them can be ticked off, though. For example, I reviewed the ones I'm listed next to and issued a pull request.

deirdrelee: we should use actions for recording tasks

BernadetteLoscio: it's just informal, our own todo list

<Caroline_> https://docs.google.com/document/d/1HCPCDD5QbSVQrlAI2v6HT6iF_-N-M8KuTysgh7ZrTaM/edit?ts=56e73e23#heading=h.z5puontm4h1f

<Caroline_> that is the table BernadetteLoscio is talking about

table on intended outcomes and tests

BernadetteLoscio: it's really important that we can test each bp, these need to be reviewed
... there is always a risk that a bp can't bet tested

deirdrelee: should we have a vote on this on the 30th March, htat way we can be sure to review it

laufer: i agree there should be a strong relation between tests and intended outcomes
... sometimes we have to rephrase the intended outcomes, sometimes the test

<yaso_> deirdrelee: do you want feedback as github pull requests or in the doc?

<Caroline_> please, just change directly at https://docs.google.com/document/d/1HCPCDD5QbSVQrlAI2v6HT6iF_-N-M8KuTysgh7ZrTaM/edit?ts=56e73e23#heading=h.z5puontm4h1f

Caroline_: if everyone agrees to edit the google docs, it's better, it makes it easier to see edits and comments

<Caroline_> :)

yaso_: it might be easier to use github? you can see history and comments

<Zakim> phila, you wanted to talk about Google docs (don't worry)

phila: if you want to use google docs, use google docs, after please save as pdf and publish on wiki

Ig_Bittencourt: what should we do with google doc? just give a comment?

BernadetteLoscio: give a comment, but give an alternative solution

Ig_Bittencourt: there are cases there the intended outcome is right, but the test is not possible

BernadetteLoscio: always give a proposal

laufer: if we have a bp that's impossible to test, should it be included?

<annette_g> I suggest we encourage people who are having a hard time writing a test to say so in the doc

<annette_g> then someone else may have an idea of how to test

<newton> +1 to annette

BernadetteLoscio: if there is a bp that is difficult to test, what do we do?

<Caroline_> +1 to annette_g

<laufer> why we have to orient a user to use the right formalization level...

<laufer> this is not a particular thing od data on the web

phila: when we get to CR, we flag it as at risk. if we don't get implementations, it's okay, we revmove bp

<yaso_> annette_g

<annette_g> nap time

phila: if it's not flagged as at risk. if we don't get implementations, we have to go back to working draft

<hadleybeeman> sleep well, annette_g :)

<annette_g> good night everybofy!

<phila> Thanks annette_g

<BernadetteLoscio> thanks a lot Annette!

<hadleybeeman> cheers all!

<annette_g> np

<yaso_> coffeebreak now :-)

<laufer> cheers, hadley

<phila> ==Coffee==

<phila> scribe: phila

Final Issues for BP

deirdrelee: We have an hour and 20 mins to work until lunch. Need to finish off BP and then look at DQV

<yaso_> scribe: yaso

<yaso_> scribe: yaso_

<phila> peterW: On the topic of the tests. There are some tests to do with whether people or machines can understand what the data is about.

<Ig_Bittencourt> http://www.smart-society-project.eu/essence/

<phila> ... We can't really do that well. I'm saying keep it in. We can't do it well quite yet, but a lot of stuff in projects I'm invilved in, like ESSNCE, using NLP and AI

<phila> ... there's good work there and elsewhere that mean that, yes, it's aspirational, but in a short time, it will become more functional.

acribe: phila

<scribe> scribe: phila

peterW: If we keep it, there are people who can do it better than others. In our BPs, we should understand that there is a bleeding edge that will become mainstream
... We can set things in place that will last.

Ig_Bittencourt: I was looking at the SDW BP document
... Their BP8, 29, 30
... their 29 is closely related to our work.
... BP29 http://w3c.github.io/sdw/bp/#self-describing-apis

<newton> SDW29 and DWBP26

Ig_Bittencourt: Maybe we can talk to them about this as it is so similar to ours
... And their BP 30 http://w3c.github.io/sdw/bp/#include-search-api is about users
... So their 29 is my concern.

deirdrelee: We should recommend to SDW that they don't get into general BPs, so that we can differentiate and not replicate.

laufer: I was in the meetings with SDW. What I got from them - they have concerns about things that they need in the APIs, like s eearch funtionality
... is this something we'll have in our BPs.
... We should say that we won't do everything. They need a set of specific things that they need. They shouldn't need to put in general stuff.

<Caroline_> scribe: Caroline_

phila: what do you suggest they do if they come across our document should have?
... should them send us a complete BP and suggest we put in DWB?
... they said that if they can specialize they will
... but it they cross on some BP that should be on DWP, should they send us?

laufer: we have to analize it

<phila> laufer: If they have a general BP, that is generic, not just for them, that's missing in our BPs, we have to analyse and decide what to do.

<phila> laufer: They have sub groups of BPs?

<scribe> scribe: phila

laufer: It's a strange thing that there aren't otehr WGs that need this.
... Maybe there will be future groups and we'll need to link between them or whatever.

deirdrelee: Of course if thyere is sometehing that they need, the reality of the timing means that we're coming to an end. We can't incorporate lots of new BPs. They can specialise ours, if they need others, they can add them of course.

Ig_Bittencourt: There are some BPs that are not specially spatial. They have some about search engines for e.g. That's not in scope for DWBP
... but I still think that BP29 is basically the same thing as our Document your API BP. In *this* case, it shoujld be one of ours.

BernadetteLoscio: I think that BP28 is something similar to what we have. Maybe BP27 is related to subsetting. We can maybe send a message that this BP says what we want to say

<scribe> ACTION: ig to attempt to merge the BPs on API documetnation [recorded in http://www.w3.org/2016/03/15-dwbp-minutes.html#action01]

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

<scribe> ACTION: bettencourt to attempt to merge the BPs on API documetnation [recorded in http://www.w3.org/2016/03/15-dwbp-minutes.html#action02]

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

<newton> ACTION: bittencourt to attempt to merge the BPs on API documetnation [recorded in http://www.w3.org/2016/03/15-dwbp-minutes.html#action03]

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

<scribe> ACTION: Ig Ibert Bittencourt Santana Pinto to attempt to merge the BPs on API documetnation [recorded in http://www.w3.org/2016/03/15-dwbp-minutes.html#action04]

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

<newton> try Ig_Bittencourt

<scribe> ACTION: Ig_Bittencourt to attempt to merge the BPs on API documetnation [recorded in http://www.w3.org/2016/03/15-dwbp-minutes.html#action05]

<trackbot> Created ACTION-267 - Attempt to merge the bps on api documetnation [on Ig Ibert Bittencourt Santana Pinto - due 2016-03-22].

laufer: I think SDW has to make the relation, not us

Ig_Bittencourt: wrt BP 28 about spatial data - I think theirs is fine, It relates to ours, but it's more specialised - but the title could make it clearer

BernadetteLoscio: If Ig can analyse SDW BP27, 28, 29 that would be helpful

deirdrelee: Whule we say in our opinion they should do XYZ, of course it's up to them to do as they see fit.
... Any more on gthe BP doc?

BernadetteLoscio: Nope. we're done

<PWinstanley> s/shuild/should

Data Quality Vocabulary

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

antoine: Introduces the DQV for the benefit of Share-PSi partners
... It's a very generic framework. Not providing complete metrics, it's a framework for metrics.
... users can add their metrics and exchange them meaningfully
... Really about building blocks between diff quality data.
... yesterday we did a lot of work on pending issues.
... This morning we need to tackle a couple more

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

antoine: There's an issue about aligning the ISO/IEC 25012. Thyat has been done in revent weeks but it's not recorded in the tracker yet.

close issue-200

<trackbot> Closed issue-200.

antoine: Next issue...

issue-221?

<trackbot> issue-221 -- What is the importance of the alignment between hcls-dataset and dqv/duv? -- open

<trackbot> http://www.w3.org/2013/dwbp/track/issues/221

antoine: That's abouit the alignment with the HCLS work.
... They created a profile for dataset descriptions
... That states a number of stats relevant for describing datasets. It also specifies some fields that are connected with quality dimensions.
... We haven't haad time to look at this in detail.
... We got feedback from Michel Dumontier who did some mappings. He also pointed out some families of stats that re important for him.
... We could try and show how to represent their stats in the DQV framework.
... The second would be a more formlaised e.g. between dimensions and ??
... We are working on this. But our discussion today shows there's a lot of work to do there - not sure if it's a high priority. riccardoAlbertoni and I are keen to do it, but it's a question of time.
... So we'd like WG approval to let it lie for now and pick up if we have time.

phila: So Michel didn't offer to do that?

antoine: We have to get back to himn on otehr issues so that might prompt some help.
... I might also see him next wweek at Stanford.
... And there are others st Stanford with a lot of good material we're using.

deirdrelee: Explains about the moving of hte calls for the next 2 weeks.

antoine: With the number of actions we decided yesterday, we put those as a higher priority.

deirdrelee: makes sense.

laufer: Yesterday you showed a solution to relate metrics.
... I'm wondering if we need a similar thing to relate dimensions or can that be extracted from SKOS relations

antoine: We'd rather use SKOS relationships

riccardoAlbertoni: We can have many relationships using SKOS.

antoine: Anotehr issue from HCLS - part of the action to align us with them was about aligning how we describe stats. That's going to be important for describing metrics.
... What they have in HCLS is more sophisticated - one of our goals would be how to publish stats in a way that can be used in complex models like data Cube to more simple ones like we're proposing.
... A QB is a lot of data - might not be clear what the quality measures are, but we can show how to use DQV properties to do it.
... We think that would be a good result.

laufer: In the case of QB, you have to use the subsetting/slicing concept to group quality metrics.
... Yesterday we were talking about quality not onkly about Dataset. Maybe you can have stats for a slice

issue-222?

<trackbot> issue-222 -- Multiple/Derived values of a metric -- open

<trackbot> http://www.w3.org/2013/dwbp/track/issues/222

antoine: We took a lot of actions around issue-222 yesterday so I think we can close it.

close issue-222

<trackbot> Closed issue-222.

issue-223

<trackbot> issue-223 -- Parameters for metrics -- open

<trackbot> http://www.w3.org/2013/dwbp/track/issues/223

antoine: This was the painful issue this morning. This came out of comments received
... asking how parameters can be expressed for a metric.
... Small discussion on the mailing list on this. Jedremy D suggested using a DAQ property
... How to express parameters etc. trying to align with DAQ and we basically failed.
... It cost us some headaches and we felt that we just can't align everything as JD suggested
... We have some options that are good. We can't quite align them.
... They assumed we'd use some modelling patterns with classes of metrics and that we'd use their properties, but their domains don't match.
... That comes back to yeterdayt's discussion about abstract classes

laufer: A metric is a property or a class?

riccardoAlbertoni: Class

laufer: So parameters could be peroperties from thsi class

<riccardoAlbertoni> https://www.w3.org/2013/dwbp/wiki/Examples_for_metrics_extra_parameters

riccardoAlbertoni: If we try to be compliant with JD's suggestion, we have to represent the parameters in the way his model requires and that doesn't work for us.
... We have 2 examples in the wiki page in which we try and see how it looks to extend, startng from the property that JD was suggesting.
... Maybe we have to wait for feedback from jeremy Debattista

antoine: There are several options... a Metric is a class, but we may have to create subclasses to give the properties. Like the no. of items in the dataset.
... If we want to create subclasses it gets complicated (paraphrase)

laufer: I think it wojuld be nice to have a property that could be specialised
... I might want to say more about metrics than parameters.
... You could have general properties for a Metric.

antoine: Then how do you create the link between specific properties. Subproperties? Other methods?
... Creating properties for an abstract class gets tricky. Open World Assumption comes in to mess things up.
... We may have a solution based on SHACL
... That can attach specific rules on how to give a metric

<scribe> ACTION: antoine to add intro to DAQ in the DQV [recorded in http://www.w3.org/2016/03/15-dwbp-minutes.html#action06]

<trackbot> Created ACTION-268 - Add intro to daq in the dqv [on Antoine Isaac - due 2016-03-22].

deirdrelee: I think properties are out of scope. It's an endless list. If there's a future update, then OK, but a profile defined in SHACL sounds good to me

antoine: We are choking on this aspect of how to create application profiles.

deirdrelee: There's nothing stop people creating parameters, sub classes etc. We're not closing doors by not saying how to do it.

antoine: We can afford to remain silent on that.

deirdrelee: The abstract way may not work. Examples/options are going to be important.#

issue-224?

<trackbot> issue-224 -- Expected Data type for metrics -- open

<trackbot> http://www.w3.org/2013/dwbp/track/issues/224

antoine: We didn't get to this today due to lack of time. We got a comment on that
... I mailed Werner about that last week and his answer may change.
... Not sure exactly what he's after. Need more discussion.

<Zakim> phila, you wanted to talk about RdfDatatypes, skos:notation etc.

phila: Would something like the expectation around skos:notation be sufficient?

antoine: Values tend to be a number. But Werner wants anything to be possible, including, for example, a graph.
... That feels a little strange to us as we see metrics as nunmbers.

PWinstanley: They can be categorical

riccardoAlbertoni: That would be fine

antoine: I'm not so sure
... Gives an example...
... We prefer to wait until the commenter comes back to us.

PWinstanley: Certificates?

antoine: Those are in a different part of the model
... If we start mixing things up it gets more complicated.

PWinstanley: The absence of a value doesan't mean the absence of a quality metric, it might be where I'm not looking.

antoine: Yes.

PWinstanley: So do need some sort of Boolean to say there's a quality value at X?
... if there are 4 places where I can find a quality metric, do I have to look in all 4?
... You have a place for categorical quality, a metric, a certificate. I just want to know is there a quality assessemnt. How do I find it?

antoine: Yes, a SPARQL query should check multiple places.

PWinstanley: Maybe 4 ASKS?

antoine: It depends what you're interested in.

laufer: It seems that if we talk about metrics, we're talking about atomic things. This comes back to yesterday. How do we link to derived. If there's a data record, is there a relation we can find (sorry, not well scribed)

<PWinstanley> s/qell/well

laufer: How to relate these things that are called metrics? Derived is one way, but someone may want to relate a lot of quality things.

antoine: yes, maybe. We're wokring on this.

deirdrelee: Any more for any more?
... OK, we're half an hour until lunch. maybe there's something else?

AOB

Caroline_: I was talking to phila - when people reference our BPs, it can be confusing. So our suggestion is to rename our BPs to DWBP_x
... anda SDW can do the same

<BernadetteLoscio> we have Best Practice 1 and they also have Best Practice 1

deirdrelee: Do our Bps have a URI

<BernadetteLoscio> the proposal is to have: DWBP1 and SDWBP1

phila: yes, they have a frag id
... Adding in'DWBP'; before the number is easy.

laufer: We said a long time ago that the number would be a problem
... Since the beginning, we've needed a kind of keyword that is not a nunber
... The full ID now is too long.
... we can do better than nunbering.
... better than saying the full name

<PWinstanley> @laufer +1 about the key being a pithy name

BernadetteLoscio: We have this now as the numbers aren't quite stable yet
... as we stabilise, we can do that.
... It's only a problem between us.

Caroline_: we'll keep the numbers because we need them

laufer: Maybe 'BP-URIs' or whatever
... 'DWBP is not obvious'

deirdrelee: I'd like to close this topic.
... We finalise the document.
... then we can think about shorter IDs

<Caroline_> DWBP would come before the numbers

<Caroline_> instead of "best practices"

<scribe> ACTION: phila top creatre wiki page to collect a wish list for a future WG [recorded in http://www.w3.org/2016/03/15-dwbp-minutes.html#action07]

<trackbot> Created ACTION-269 - Top creatre wiki page to collect a wish list for a future wg [on Phil Archer - due 2016-03-22].

Time scales

deirdrelee: When are going to go to CR on the BP doc?

newton: We need to work through the table.
... We have a lot of actions due around 1 April.
... Once that's done, we can freeze the document.

phila: Do you have enough to complete the doc?

BernadetteLoscio: We have a few issues that we'll resolve next Wendesday and the review of the tests
... So I think those are the open issues. If we can resolve them on Wednesday, we'll be OK.
... The following week, we can agree about the final BPs then we can make the 1/4/16 date
... If we can slose those issues as proposed we're OK.

deirdrelee: Next Friday 18 - no call
... Next week, is Easter, so move meeting to Wednesday 23
... Following week we can meet on Friday 1 April as normal, no need to change the date.
... So on 1 April. we freeze
... Can't freeze until all issues and actions closed, comments answered.

BernadetteLoscio: if someone wants to make a change following the 1 April review, then we'll have to take action.
... We won't touch GitHub after 1 April, then we can vote on 8 April

deirdrelee: So people should voice any concerns as soon as poss after 1 April
... If anyone has issues, we really want them before 1 April.

<Zakim> Caroline_, you wanted to ask about the exactly dates of next meetings

deirdrelee: So what's the time line on DQV?

[Mumbles]

antoine: It's up the Wg to decide priorities

deirdrelee: But by 1 April, you'll have all the issues addressed?

riccardoAlbertoni: Yes

antoine: yes

deirdrelee: I think DUV was in a similar state
... So we have final vesions on 1 April for review by the WG

antoine: If we decide we're done on 1 Arpil... the final final final may be 1 May - we'll have to decide as a WG
... IF we need something done by then, OK.

deirdrelee: OK, we'll stick to 1 April as the target.
... Then we'll be guided by process

antoine: we can use 1 May as the deadline

deirdrelee: That allows a month for feedback which may be too much.

antoine: I think I'd rather have a month than less.
... There may be an interesting time to get implementation feedback.

<deirdrelee> phila: it would make a nice public announcement if we have a new version of dqv and duv notes on same day as cr of bp

laufer: We're deciding this for DUV as well? Eric should be involved

deirdrelee: We talked yesterday, he agreed.
... Calls last DWBP F2F to a close.

<laufer> uhu!

<laufer> hey, ho, let´s go!

phila: Meeting ends with 11 open issues and 38 open actions

<scribe> meeting: DWBP F2F Day 2