Data on the Web Best Practices Working Group Teleconference

<scribe> scribe: PhilA

<scribe> scribeNick: phila

<hadleybeeman> Phila is a bit early :)

[Discussion of the wish list] https://www.w3.org/2013/dwbp/wiki/Main_Page#Wish_List

Precious call minutes

<hadleybeeman> s/precious/previous

-> https://www.w3.org/2016/11/11-dwbp-minutes Minutes from 11/11/16

NOTUC on previous minutes?

RESOLUTION: Accept previous meeting minutes

BP Transition

BernadetteLoscio: I think we're ready and I think Carol and Newton agree

<BernadetteLoscio> http://w3c.github.io/dwbp/dwbp-implementation-report.html

-> http://w3c.github.io/dwbp/dwbp-implementation-report.html Implementation report

BernadetteLoscio: We tried to explain the methodology and the kinds of evidence that we collected
... In 2.1 we show the relation between BP and evidence
... considering datasets, guidelines and docs
... for some BPs it's easy to show, Others are more difficult
... BP28 is the most difficult on ewhich is about assessing datasets coverage as it's hard to find real implementation for this as it's very specific.
... But it is relevant and we showed that it is possible to do and we have agreement that it's important
... We have the lost of evidence, a link for each one
... That section 2.2 - 2.4
... Also have guidelines from Share-PSI etc.
... Included a section with some graphics to show which BPs have more evidence. Also tried to show that we have...
... First graph shows all evidence, 2nd for datasets and portals, 3rd for..
... Also made a cross ref between BPs and challenges
... After the graphics we tried to analyse the evidence that we collected. We want to show that preservation is not difficult to implement, just hard to find evidence of implementation.
... This section needs to be finished.
... We evaluated data catalogue solutions - CKAN and Socrata - you can see...

<BernadetteLoscio> https://docs.google.com/a/cin.ufpe.br/spreadsheets/d/18Gz0n9HOmeSPjo6qChhXdtPeOBf9kl1mFlA-KkqxP24/edit?usp=sharing

BernadetteLoscio: Plan is to convert the Google doc into an HTML table
... We wanted to show that CKAN and Socrata supports BPs even if a specific publishers doesn't follow them.
... We wanted to end with a summary of how to implement each BP and proof that we've done it.
... We're working on this implementation to show a step by step guide of how we implemented each BP.
... We can therefore show that each BP is implementable

newton: We are making subtle changes like swapping Google Docs for HTML but that's cosmetic

hadleybeeman: Thank you. This is a huge amount of work. The most thorough I've seen.
... The table at section 2.1 is what the Director will focus on.
... There may be questions like why does BP28 only have 3 when others have more - but you can answer that verbally.
... The last thing you said - about BPs being implementable - by showing that others have done it, we've already proved that.

BernadetteLoscio: It's because I'm worried about BP28 where we only have 1 evidence from a dataset and 2 from documents/blogs
... it's a site
... This implementation was made by someone else (not us)
... So we thought that it would be nice to show that someone else has done everything.

hadleybeeman: I think that's a user. Someone has used it - good. But you've already proved that it's implementable.

BernadetteLoscio: Can we stillwork on the implementation report up until the Director meeting?

<annette_g> Does anyone else see "The DocumentView interface is not supported

<annette_g> Non-W3C methods of obtaining "window" also failed" when they open the editor's draft?

BernadetteLoscio: We can probably add more evidence. I;ve written to Christophe, for example

hadleybeeman: The IR can be edited up until the Director's call, it's not a formal document.
... That said, we need to vote on the BPs based on what we have at the time of the vote
... I would also recommend... assuming we go ahead, as you prepare for the Director's call, be ready for questions you can predict you're going to get.
... You can put info in the doc if you want to but it may not be read.

BernadetteLoscio: It's not a Note, but we can still make it available, no?

phila: Yes, I'll put it in /2016/11/{blah}

<newton> +1 phil :-)

BernadetteLoscio: Do you think its worth including the data catalogues evaluation?
... IU think it shows that they're available in the solutions that are widely used. They're not currently in the evidence table.
... It's another level

hadleybeeman: Phil said it can be on the Web, so it's out there. You can do what you like with it.
... If you want to add, close off etc. you can

BernadetteLoscio: But for Director's call, it's the table in section 2.1 that's most important

hadleybeeman: Yep, and 2.2 explains 2.1
... It's a very good doc and very thorough

<Zakim> antoine, you wanted to discuss nitpicking

antoine: It's a very good doc.
... Just one comment - on the graphics in section 3
... I'm not sure why the no. evidences is in a differnet diagram
... It's because not all docs have references
... It looks a bit different from the other one which makes it look as if there's a difference in the methodology.
... It looks as if you're trying to hid something when in fact you're trying to explain.

BernadetteLoscio: Is it a problem that we don't havea docs and ref for each BP?

antoine: No, as long as there's another kind of evidence

BernadetteLoscio: This type of evidence is a kind of support.
... Our main concern was to have datasets and data portals for each BP

antoine: I'm already convinced.
... On the number of evidence per challenge
... I'm concerned that this diagram over emphasises the smaller level of evidence for BP28
... It highlights data preservation prob even more

BernadetteLoscio: I agree. I had similar thoughts when I saw it.

antoine: In 2.3 there's a really weird link for ???

<hadleybeeman> q/

[Discussion - the hash bang URL is fine]

<hadleybeeman> phila: This document... we can put it on the web. It is linked from the PR and the actual recommendation. It is part of the documentation.

<Makx> q

hadleybeeman: have all comments received been addressed? Is there a disposition of comments?

BernadetteLoscio: We received some comments but they were not to change the content
... We had a message from Christophe but he was thinking that we were still in the previous phase, but he made some commetns adn suggestions for fixing some minor mistakes

hadleybeeman: And do we have a message from the commenters that they're happy with our response.

BernadetteLoscio: They were about fixing a word or two
... I answered saying that we're going to fix it, it's not about making a proposal.

annette_g: I wanted to point out the restrictions on the Google doc prevent us getting in

<Caroline_> http://w3c.github.io/dwbp/dwbp-implementation-report.html

annette_g: Can it be linked from the regular WG homepage so we can get to it.
... There's a link to the IR but it goes to the Google doc
... Oh, no, it goes somewhere else

hadleybeeman: So we need a message from Christophe confirming that he's happy

BernadetteLoscio: I'll collect the messages
... I'll do it the same way as last time on a wiki page

<Zakim> phila, you wanted to talk about Wendy Carrera https://lists.w3.org/Archives/Public/public-dwbp-comments/2016Nov/0001.html

<hadleybeeman> phila: Wendy's comments are more than editorial. She runs the European Data Portal.

On Wendy's mail, yes, we can point to https://www.w3.org/TR/dwbp/#LocaleParametersMetadata about locale-neutral data. We could talk about her comment on multilingualism. I don't think that we can add in a line about multilingual labels in data at this stage but I'm not sure this is where that belongs anyway - I'd say that's in vocabulary development. I'd be happy to amend this sentence in BP 15:

" In the context of the Web, using unambiguous, Web-based identifiers (URIs) for standardized vocabulary resources is an efficient way to do this."

to say

" In the context of the Web, using unambiguous, Web-based identifiers (URIs) for standardized vocabulary resources is an efficient way to do this, noting that the same URI may have multilingaul labels attached for greater cross-border interoperability."

(Or some such small addition in that general area of the doc).

On DCAT-AP, I thought we'd mentioned it somewhere but a quick search shows me to be mistaken. We could perhaps amend BP 1 so that the current:

<hadleybeeman> Phila: She made a couple of comments on her email. I have a few suggestions on how to reply. But we do need her to then say, "that helps"

"when defining machine-readable metadata, reusing existing standard terms and popular vocabularies are strongly recommended. For example, Dublin Core Metadata (DCMI) terms [ DCTERMS] and Data Catalog Vocabulary [VOCAB-DCAT] can be used to provide descriptive metadata."

becomes

"when defining machine-readable metadata, reusing existing standard terms and popular vocabularies are strongly recommended. For example, Dublin Core Metadata (DCMI) terms [ DCTERMS] and Data Catalog Vocabulary [VOCAB-DCAT] can be used to provide descriptive metadata. Such vocabularies are designed to be very flexible so it is often helpful to use a specific /profile/ of a vocabulary such as the European Commission's DCAT-AP (link)."

<hadleybeeman> ...Her second comment was to comment DCAT-AP. Which we haven't yet.

<hadleybeeman> ...I think that's as far as we can go to address those at this point.

<hadleybeeman> BernadetteLoscio: I thought her comment meant changes to the implementation report

<hadleybeeman> phila: I think we can get away with these minor changes to the BP doc.

<hadleybeeman> BernadetteLoscio: So you can help us to answer her message?

<hadleybeeman> phila: I'll make those changes to the doc and confirm them with Wendy

<scribe> ACTION: phila to act on Wendy Carrera's comments suggested and write to her [recorded in http://www.w3.org/2016/11/25-dwbp-minutes.html#action01]

<trackbot> Created ACTION-303 - Act on wendy carrera's comments suggested and write to her [on Phil Archer - due 2016-12-02].

BernadetteLoscio: I'll take another look at the mailing list.

<Caroline_> https://github.com/w3c/dwbp/pull/473

newton: We had pull requests from Andrew Kirkpatrick, from Adobe that changed a link. Do we include that as a comment

hadleybeeman: I'd point to that and say that we accepted their proposal

Makx: Just one comment on section 4 of the implementation report. You're making statements about a product that might not be in line with their view. If you publish without asking them, you might get into trouble.
... You're matching a BP against a product - the vendor might object to that.

BernadetteLoscio: I see that. We've finished this evaluation yesterday. We're going to contact the vendors and see if they agree with this.
... If not, we won't include it.

<riccardoAlbertoni> just delete the column about fail

hadleybeeman: A suggestion - we need to show evidence - not all evidence will be relevant.

<ericstephan> +1 hadleybeeman

<Makx> +1 to hadley

hadleybeeman: You could remove the fail column

<riccardoAlbertoni> +1

<Caroline_> +1 :)

hadleybeeman: Just keep provides evidence, provides partial evidence instead of "pass" and "partial pass"
... That's less confrontational

BernadetteLoscio: Instead of pass/fail, we can just say which solutions implement which BPs
... Like in 2.2

<Makx> that's OK

hadleybeeman: That would work too.

<riccardoAlbertoni> \me the same here

annette_g: One little thing that I think we can take care of. The editor's draft in Safari I get javascript errors. But it works in Chrome and FF

phila: I believe that problem will disappear in the published version in which ReSpec disappears

<newton> @annette_g, would you mind chatting a little on skype to show me those errors?

PROPOSED: Accept comments from Christophe G and Wendy C that will lead to very minor tweaks to the BP doc

<annette_g> what if commenters aren't happy with our changes?

<annette_g> :)

+1

<ericstephan> +1

<riccardoAlbertoni> +1

<newton> +1

<hadleybeeman> +1

<annette_g> +1

<deirdrelee> +1

RESOLUTION: Accept comments from Christophe G and Wendy C that will lead to very minor tweaks to the BP doc

<Makx> +1

<BernadetteLoscio> +1

<Caroline_> +1

<annette_g> but Wendy may be trying to get us to add a BP about multilingual publishing

<hadleybeeman> @annette_g, in which case we'd have to back and do CR again.

PROPOSED: That subject to positive replies from Wendy C and Christophe G, the WG will seek transition of the BP doc to Proposed Recommendation, noting the extensive evidence of implementation that has been gathered

<hadleybeeman> +1

<annette_g> +1

<ericstephan> +1 in the season of hope

<Caroline_> +1

<riccardoAlbertoni> +1

<newton> +1

<Makx> +1

<BernadetteLoscio> +1

<antoine> +1

<deirdrelee> +1

RESOLUTION: That subject to positive replies from Wendy C and Christophe G, the WG will seek transition of the BP doc to Proposed Recommendation, noting the extensive evidence of implementation that has been gathered

<BernadetteLoscio> uhhuhuhuuhuhuhuhuhu

<riccardoAlbertoni> congrats!!!!

<ericstephan> woot woot

<newton> :-) :-)

<annette_g> PROPOSED: A vote of thanks to editors for putting together a fantastic implementation report!

<hadleybeeman> +1

<riccardoAlbertoni> +1

<annette_g> +1

+1

<ericstephan> +1

<Makx> +1

<deirdrelee> +1 :)

<antoine> +1 :-)

<newton> thanks! :-)

RESOLUTION: A vote of thanks to editors for putting together a fantastic implementation report!

Data Quality vocabulary

<BernadetteLoscio> thanks a lot everybody!

<Caroline_> Thank you! Great job all members of the WG :))))))

hadleybeeman: DQV editors, would you like to publish a new verrsion of DQV?

antoine: Yes
... All we changed was a couple of mappings in mapping to the ISO quality dimensions

riccardoAlbertoni: And we changed data usage to dataset usage

hadleybeeman: Are there any other changes?
... Are there any other comments?

riccardoAlbertoni: We collected some new implementations in the wiki

antoine: There are some old comments for which we didn't receive precise feedback to your replies, but we're thinking of adding to the WG's wishlist

hadleybeeman: So you're saying that even though you don't have confirmation from the commenters, after allowing good time, you want to go ahead

PROPOSED: That the current Editor's Draft of the Data Quality Vocabulary be published as an updated Note

<hadleybeeman> +1

<deirdrelee> +1

<Makx> +1

<annette_g> +1

<antoine> +1

<newton> +1

<riccardoAlbertoni> +1

<BernadetteLoscio> +1

<ericstephan> +1

<Caroline_> +1

RESOLUTION: That the current Editor's Draft of the Data Quality Vocabulary be published as an updated Note

Dataset usage Vocabulary

<annette_g> yay!

ericstephan: I'd like to mention the ongoing discussion with Andrea P
... Andrea has been excellent at providing feedback on the DUV
... He pointed out an inconsistency in data usage and dataset usage
... He was asking about the relationship with DQV
... I was thinking these might be good topics for a summary document that we could publish cf. putting in the doc.
... Andrea has also pointed out some errors in usage of dct:identifier
... So I'd like permission to check with the SPARK ontology editors
... We have a stable version of the doc, it's been stable since August. What I'd like is permission to publish with the examples fixed.

hadleybeeman: Typos, yes. What's the discussion with SPARK ontology editors. How different could that make the doc?

ericstephan: Our team has been working with SPARK for a while.
... It's just making sure that we use the correct property in the way that SPARK would like to use it
... If it's not something that we can verify, I'd just take the property out of the example.

<hadleybeeman> phila: is this dcterms:identifier?

<hadleybeeman> eric: no

<hadleybeeman> phila: In the resolution, we can note tis is an ongoing discussion. Subject to simple resolution of the issue with vero:isReferencedby, and the one on dcterms:identifier, the group is happy.

<hadleybeeman> ...If that isn't simply resolved, we could reconvene just before the end of Dec.

Makx: I read Andrea's comment, I think biro is an indirection (pointing to a record)

ericstephan: That's what I think we're doing
... I just want to double check
... Worst case, we go back and use dct:isReferencedBy but I don't think that's correct

hadleybeeman: We could resolve to publish now and come back if we have to

phila: True

<hadleybeeman> phila: Eric, if you make a snapshot copy before you make any changes, I'll use that one.

<hadleybeeman> ...But you're hoping not to make changes, right?

<hadleybeeman> ericstephan: Right

<Makx> Eric, I think you do use biro:isReferencedBy incorrectly; your object is a fabio:InstructionalWork not a bibliographic record

hadleybeeman: We could vote on what there is, and come back for another meeting is we have to

ericstephan: Makx also provided some guidance in IRC
... If it blows up, we'd just remove anything that seems to be controversial, that's how I'd contain it

hadleybeeman: Are you talking about removing a term, that's substantial.

ericstephan: I want to do the right thing but I also want to respect the timing

Makx: Just to say to Eric, it's not a question of changing the vocab, it's just changing the example
... I was looking at the diagram - that is correct. The example is in conflict with the diagram
... The diagram references a biro prop, in the example, you make a reference to a ?? work
... The consequence is for the example

ericstephan: That gives me hope.

<BernadetteLoscio> +1

PROPOSED: To publish the editor's draft of the Dataset Usage Vocabulary, pending the fixing of typos

<Caroline_> +1

<hadleybeeman> +1

<ericstephan> +1

hadleybeeman: Please, ericstephan, get back to us by Wednesday (Europe) if you need a meting next Friday.

<BernadetteLoscio> +1

<annette_g> +1

<Makx> +1

phila: Either way, we're only going to publish one new version, not two.

<ericstephan> https://lov.okfn.org/dataset/lov/vocabs/duv

RESOLUTION: To publish the editor's draft of the Dataset Usage Vocabulary, pending the fixing of typos

<hadleybeeman> yeay!!!

ericstephan: LOV did publish our vocab but since we didn't publish the other vocabs, it looks tiny. Can we improve that

<Zakim> newton, you wanted to discuss the last and very quick question - can we add more evidence for Data Preservation BPs?

<Zakim> phila, you wanted to talk LOV

<hadleybeeman> phila: sure

<hadleybeeman> phila: I see what ericstephan means. The number of terms DUV defines is small.

-> https://www.w3.org/ns/duv Namespace

<hadleybeeman> ...Therefore, your namespace file — that's what they look at.

<hadleybeeman> ...They don't look at the TR space document.

<hadleybeeman> ...There's got to be a way to do something bout that.

<hadleybeeman> ericstephan: I can go back to them and see what we might do. I'll copy phila

<hadleybeeman> phila: I can easily ad to the NS document at any time. IT's not locked down.

antoine: I suspect that we might be able to use a specific field in the metadata to force the links to be recognised

<ericstephan> yes will do!

hadleybeeman: The queue is empty...
... Eric, you mentioned...
... Topics for a summary document
... You can write a doc like an implementation report, i.e. not a formal doc

ericstephan: Its just a web page about the relationships or background info
... You can put it in the wiki etc.

<BernadetteLoscio> I can help you Eric ;)

hadleybeeman: The wiki is frozen when the WG closes

<BernadetteLoscio> ok ;) thanks!

hadleybeeman: We're 30 mins over...
... Thank you to editors, contributors, participants

<BernadetteLoscio> sure!!!

hadleybeeman: We'll work on other stuff coming up

<ericstephan> Take care and thank you all for everything!

<deirdrelee> yay! Great work

<Makx> OK bye bye!

Caroline_: Thank you the chairs, Obrigado

<Makx> Hope to see some of you next week

<deirdrelee> see some of you next week!

<annette_g> bye folks!

<deirdrelee> Thanks for chairing hadleybeeman

<hadleybeeman> bye all :)

<hadleybeeman> thanks for co-chairing, deirdrelee and yaso!

<hadleybeeman> :)

<Makx> Bye

<newton> bye and thank you all!