Data on the Web Best Practices Working Group Teleconference

<scribe> scribe: deirdrelee

PROPOSED: Approve last week's minutes https://www.w3.org/2016/06/10-dwbp-minutes

+1

<annette_g> +1

<newton> +1

<RiccardoAlbertoni> +1

<phila> +1

RESOLUTION: Approve last week's minutes https://www.w3.org/2016/06/10-dwbp-minutes

<hadleybeeman> https://www.w3.org/2013/dwbp/wiki/Status_of_comments_about_the_last_call_working_draft

Public comments on docs

BernadetteLoscio: we didn't receive any new comments this week
... but we tried to address all the comments we received until now
... we updated the table and have a response and correspondance for each comment on the table
... comment 8, annette_g responded, made the change in the doc, and the commenter agreed, so we can consider this as resolved
... for comment 7 we have reached a response, but need to email Franz, but we can consider this resolved
... for the others, we have proposed resolutions, and if the group agrees we'll email the commenters
... there is a comment from Ivan on how to include gis in the bp doc. i looked at the sdwg doc, but can't see a resolution

<BernadetteLoscio> Ivan Herman

<BernadetteLoscio> comment 2

BernadetteLoscio: for comment 6, the group agreed to include a paragraph on BP22, but i made another proposal to add a paragraph to BP on access, as I think it's more suitable to access
... just a proposal

hadleybeeman: do we need any more comments? is this enough

phila: we should also show comments that we've received and addressed previously, in order to satisfy directors' need that doc has had widespread review
... issue is that comments mainly come from inside the group, we haven't necessary shown that we've had widespread review

<hadleybeeman> deirdrelee: Maybe we haven't received comments from it, but we can

<hadleybeeman> ...demonstrate we've disseminated widely.

<hadleybeeman> ...There is only so much we can do about getting comments

phila: the wiki page is good, and we can point to tweets and reaction, etc
... we can certainly say we've actively been disseminating

<Caroline_> we also disseminated a lot in Portuguese, specially to the Brazilian community https://www.w3.org/2013/dwbp/wiki/Dwbp-LC-campaign.PT

phila: we could receive comment that we've done this dissemination, but received hardly any comments?
... i wrote an email to members of the WG who haven't been active, asking for input/comments
... but received zero respone

hadleybeeman: have we received comments from brazilian community?

Caroline_: not on the bps, but we have received feedback that they will participate in implementation report
... there are three orgs that aren't nicbr

<hadleybeeman> deirdrelee: Phila described the situation... how unusual/usual are we in having so few comments?

phila: I'll talk to Ralph and ask him if I came to you with this level of comments, is that sufficient?
... to see if we're ready for CR from official side

BernadetteLoscio: because this is the last draft, perhaps we shouldn't have a lot of comments. We've also received a lot of comments saying that the doc looks good and is easy to read, etc.

<newton> this comment from @csarven https://twitter.com/csarven/status/741363851933974529 is really nice :-)

<annette_g> maybe we need a table of positive-only comments

<Caroline_> we also have some of that in Portuguese that were sent to our emails :)

<hadleybeeman> https://twitter.com/RichCarne/status/732147503253065728

BernadetteLoscio: this is feedback too, not that we have something to change, but that it's great

<laufer> maybe will be interesting to append this "good "comments in the table

<Caroline_> :)

hadleybeeman: we should add the positive comments to the table...marked as resolved :)

<Zakim> phila, you wanted to talk about who can test

phila: For CR, implementaion experience/tests matter just as much as anyone else

Caroline_: so testing from nicbr and other members matter?

phila: crikey yes
... if you think of it as a technical spec, then we need code, so this is usually writeen by WG members

hadleybeeman: i may have confused in the past by saying 1) it works in pracice, not just theory, and 2) that this works in the wild
... so even if the implemenation is by wg member, it still demonstrates value

laufer: who determines if the implementation is correct?

phila: i keep referring to tech specs, as that's what process was designed for
... e.g. csv on the web has a test suite
... we have written tests for BPs, we don't have to verify them independently
... we take them on their word

laufer: if WG members have provided the implementation is that okay?

phila: we need people to respond, and have 2 independent implementations of each BP
... but every instance doesn't have to implement ALL BPs

deirdrelee: will talk about implemenation form later

hadleybeeman: you need the entire WG to sign-off on entire doc, but we don't need to point to a resolution in meeting minutes for every comment/change

<RiccardoAlbertoni> larsen effect..

phila: we just have to show that the group has received, reacted to, and responded to comments

<hadleybeeman> @bernadette, what do you think we need to discuss (because it will cause problems with the group if we don't)?

<Zakim> phila, you wanted to male a comment about #2

<phila> phila: For comment 2, it needs just a little prose as well as the refernce. I'm happy to offer the prose

BernadetteLoscio: we need someone to review all the comments and all the changes,to make sure it's okay
... other question is about comment 6
... i've included paragraph at introduction of Data Access section, not BP22

phila: we did say that this was a good opportunity to link to data usage vocabulary

BernadetteLoscio: not sure how we can refer to DUV, because we don't mention it, not sure how to make link

<phila> ACTION: phila to offer text for comments 2 and 6 [recorded in http://www.w3.org/2016/06/17-dwbp-minutes.html#action01]

<trackbot> Created ACTION-285 - Offer text for comments 2 and 6 [on Phil Archer - due 2016-06-24].

phila: let me have a go

<annette_g> bernadette did already write up some nice text for 6

BernadetteLoscio: maybe we can have both,text in the intro and in the BP

<phila> Yes, I saw that, I just want to add in the DUV ref annette_g

laufer: the comment from andrea was about different uses of data,e.g. feedback.
... it would be intereesting to have this also in the feedback section,as they have examples of lots of different types of data they have
... maybe in the second part of the intro fo the feedback section

BernadetteLoscio: can someone check the commits to make sure they're ok?
... we could also just send a message to commenters to see if its ok

+1 if they're not controversial

<BernadetteLoscio> http://www.iso.org/iso/home/standards/iso8601.htm

BernadetteLoscio: for example, for comment 3, the commenter says we should use id for iso, but not sure if this is correct

hadleybeeman: better to check with commenter,

BernadetteLoscio: i'll send email to commenters to see if resolution is okay

<Zakim> newton, you wanted to talk about evidences form

Implementation Form

newton: thought it might be difficult to fill out form on google form, usability not great
... we're creating a new form (html of course :) ), hosted by nicbr
... we won't use google forms any more

<deirdrelee:> Will the structure and questions be the same?

<newton:> more or less. We'll have an index of BPs, and when you click on one it will show a form with the same questions. Pass/fail, and a field for comments.
...Main difference is structure.

<deirdrelee:> I think that makes sense. The google form feels like one big long thing, and is a bit difficult to navigate. For people whose datasets don't meet all...

<deirdrelee:> ...General comments on the form:

<deirdrelee:> ...For each of the tests. A person will nominate a dataset, and for that dataset they'll answer all of the tests (pass/fail)

<deirdrelee:> ...I think it would be good to give them the option for comments there too

<deirdrelee:> ...We still have pass/fail, so can analyse that way. But without it, we'd be missing out.

<newton:> We are doing that

<phila> +1 to deirdrelee on adding a comment field

<newton:> ...For the evidence URL or dataset URL, instead of using one for the whole form, we're putting fields on every part of the form (per BP)

<newton:> ...So people who are testing don't have to go again and again to test different datasets.

<deridrelee:> And people might want to test multiple datasets for one bp.
...And we say: minimum of 2. For example, "must have metadata", we could have hundreds of thousands of them. Does 2 make sense?
... I was looking for things that were high risk
...For some of the tests, especially the later ones, the phrasing isn't dataset specific anymore.
...It's focused on the publisher or the user instead of the dataset.
...The tests are written differently.
...I think we'll have to update the text of the tests.
...In its current form, it might not make sense. "For this dataset, did you provide feedback."

<Caroline_> deirdrelee - could you help us changing that on this table, please https://docs.google.com/spreadsheets/d/1a9cOGzWJTIhh2OrAemvWBR8f0rv5xqvL03pJeMrotCo/edit#gid=0

<deridrelee:> ...But then you might have different roles in relationship to one dataset.

<deridrelee:> Yes I can help, Caro, but I'm aware that changing the text of the test means changing the document.
...Is that okay?
...Or maybe a tool tip or explanation?

<Caroline:> thought you were looking at line 55 or line ## in the table
...but you were talking about the BP?

<deirdrelee:> No, it's more the BPs.

<BernadetteLoscio> +1 to add an explanation!

<deirdrelee:> ...But I think what Newton has described will help this. But putting it in as a tool tip will help.

<hadleybeeman> Okay, but if the document doesn't make sense...?

<deirdrelee:> OkAY. Let's see what Newton does.
...And I guess we're not going to vote to CR next week?

<newton:> we'll send a version to you this weekend or Monday. deirdrelee, if you can suggest changes on the form, that would be good.
I think it's okay to vote next friday

<BernadetteLoscio:> What needs to be done for the voting?

<phila:> Just needs a resolution in the minutes to say "We resolve to send this to CR"
...From this conversation — we could possibly do it next week.
...Given that we want to do it ASAP, we could test the water. Try it.

<BernadetteLoscio:> When is the meeting with the director?
...Now or when we have the implementations?

<phila:> Both
...To go into CR and to go from CR to PR. And then the group's work is largely done.
...First: widespread comments and review, addressed all issues and actions, any formal objections?
...Second: implémentation évidence.

RiccardoAlbertoni: before we go to CR, we have to decide which BP are at risk
... how will we do this?

<deirdrelee:> We have no record of any BPs being at risk

BernadetteLoscio: I was talking to Caroline_ and newton, we should try to get evidence
... if we do this it will be nice way to check it bp is at risk

<deirdrelee:> deirdrelee: The process is so that we go to CR, and before then, we have to flag if anything is at risk. Then we gather evidence, the implementations.

<deirdrelee:> ...If we have problems with any, we have to go back to working draft.

<BernadetteLoscio:> I meant that we, editors, would do this right now.
...If the group also does this exercise, it can be nice.

<deirdrelee:> That makes sense, but as a chair I'd be worried about timing. you're talking about gathering evidence. If we start that process, then we delay going to CR. We don't have time to delay.
...So if you want to do it in a day ,for yourselves, that's fine. But I wouldn't be in favour of making this a wide activity.

<phila> +1 to what Dee is saying

<deirdrelee:> ...Maybe before Friday, go through each test, take 5 minutes, and ask "can I think of a dataset that will satisfy it, and won't satisfy it?"

<deirdrelee:> ...There are a couple that are obvious. But there are others that are hard in some fields. But other fields might cover it.

<BernadetteLoscio:> I agree. I'll do this now.

<annette_g> what's to stop us from just using our use cases as evidence?

<laufer> bye all... nice wknd

<RiccardoAlbertoni> bye..

<annette_g> bye all!

<yaso> Bye!!

<Caroline_> annette_g: I think it is a good idea using our use cases as evidence