Data on the Web Best Practices Working Group Teleconference

26 Feb 2016

See also: IRC log


annette_g, newton, antoine, deirdrelee, Riccardo, Caroline, laufer BernadetteLoscio


<hadleybeeman> present annette_g

<hadleybeeman> https://www.w3.org/2016/02/19-dwbp-minutes

<scribe> scribenick: antoine

<scribe> scribe: antoine

PROPOSED: to accept https://www.w3.org/2016/02/19-dwbp-minutes

<annette_g> +1

<deirdrelee> +1

<Caroline> +1


<laufer> +1

<riccardoAlbertoni> +1

<BernadetteLoscio> +1

<hadleybeeman> +1

<newton> +1

RESOLUTION: accept https://www.w3.org/2016/02/19-dwbp-minutes

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

BP-review status update

BernadetteLoscio: annette made some updates
... and we have aproposal for organizing the API section

hadleybeeman: first update on the table?

<annette_g> I did make changes to versioning, too

newton: we haven't updated the table much
... updated peter and riccardo's BP

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

Caroline: this is the link that has not been updated
... even though some of the BPs might have been updated

riccardoAlbertoni: I have pushed a revised version of the DQV example

<Caroline> :)

riccardoAlbertoni: and the BP table for my part

<Caroline> thank you riccardoAlbertoni

riccardoAlbertoni: please merge my pull request!
... (that's about BP7)

<BernadetteLoscio> thanks a lot!

riccardoAlbertoni: in the DQV section, we're refering to DQV
... the example BP is complete but depending on DQV issues we may have to update it.
... I promise I will maintain the two docs in synch.

BP1: providing metadata

<Caroline> but BernadetteLoscio worked a lot in DWBP during her vacation :)

BernadetteLoscio: will work on it asap. Really soon

hadleybeeman: how much work?

BernadetteLoscio: not a lot.
... especially as I have worked on examples for the following BPs

<riccardoAlbertoni> Just for the record, the revision on bp7 were suggested by antoine, Sorry for not giving the proper credit ;)

hadleybeeman: need help?

<Caroline> thank you antoine :)

<newton> action to newton to check if the turtle and RDFa examples are validated

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

<hadleybeeman> action newton to check if the turtle and RDFa examples are validated

<newton> action newton to check if the turtle and RDFa examples are validated

<trackbot> Created ACTION-233 - Check if the turtle and rdfa examples are validated [on Newton Calegari - due 2016-03-04].

<trackbot> Created ACTION-234 - Check if the turtle and rdfa examples are validated [on Newton Calegari - due 2016-03-04].

BernadetteLoscio: no need for help for now. I'm going to write example and then ask feedback.

hadleybeeman: before next week?

BernadetteLoscio: yes

hadleybeeman: are we in trouble because the things in this table is late?

BernadetteLoscio: beginning of next week we can have a general view about the document

BP 2: Provide descriptive metadata

hadleybeeman: do you need someone to help you review this?

newton: I can work with BernadetteLoscio
... get it done asap

BP 3: Provide locale parameters metadata

newton: same

BP 4: Provide structural metadata

<BernadetteLoscio> I'm gonna try to connect again

newton: we're ok with both examples
... the test will have to be changed

<BernadetteLoscio> I have connection probelms

BP 5: Provide data license information

newton: same. need to work on test and intented outcome

BP 6: Provide data provenance information

deirdrelee: issue was whether it needs a more detailed example
... I am considering extending the example early next week
... for the text I think what's there should be sufficient

newton: if it's necessary to change the intend outcome please let us know!
... the editors will work on this!

hadleybeeman: on all intended outcomes or only BP6?

BP 7: Provide data quality information

riccardoAlbertoni: example is finished
... for the test I didn't change, I wanted to see the other BP
... for now it seems aligned
... the question is whether we need to be more specific
... but that's a question for all BPs no ontoly this one!
... As soon as other BP's testing section are refined I could check and review

hadleybeeman: that's something to do for evey BP!

riccardoAlbertoni: we have to be careful: test mentions datasets, intended outcome mentions also distribution.
... I can try to align

BP 8: Provide versioning information

annette_g: there was already a good example

<Caroline> thank you annette_g for improving the English :)

annette_g: I think the test is ok.
... I've reviewed it but I may have written it too ;-)

hadleybeeman: who'd be a good person?

annette_g: I don't know

newton: I've not merged the pull request now

hadleybeeman: let's go BP by BP not step by step

annette_g: I've looked at the content, my PR also makes suggestions for the intro

newton: I will review the test

<scribe> ACTION: newton to review annette's test for BP8 [recorded in http://www.w3.org/2016/02/26-dwbp-minutes.html#action01]

<trackbot> Created ACTION-235 - Review annette's test for bp8 [on Newton Calegari - due 2016-03-04].

BP 9: Provide version history

annette_g: I think it's fine
... maybe need review from a vocabulary person

<scribe> ACTION: antoine to review the voc aspect of BP9 [recorded in http://www.w3.org/2016/02/26-dwbp-minutes.html#action02]

<trackbot> Created ACTION-236 - Review the voc aspect of bp9 [on Antoine Isaac - due 2016-03-04].

BP 10: Avoid Breaking Changes to Your API, Communicate Changes to Developers

annette_g: I did a lot of re-write, but let's discuss it later
... I need opinions on how to test
... sthg like 'send email to developers' would be really helpful?
... maybe scribe the actions one should take
... a notice on the API homepage

riccardoAlbertoni: is DUV keeping a list of users adopting a dataset or distribution?
... perhaps the BP example could re-use part of that info in DUV

hadleybeeman: finding your name in somebody else's metadata is a bit strange

BernadetteLoscio: we don't have a proeprty to capture this in DUV
... it shouldn't be mandatory

newton: annette_g maybe we can talk with SDW
... they have issues about APIs and examples
... we meet them next wednesday

annette_g: the issue here is really how one can write a BP for noticing users
... is it covered by SDW?

newton: not sure

hadleybeeman: I could imagine situations like this
... it might as well be solved the same way as everyone else
... there's nothing special about spatial data.
... Why don't you send an email to their list and our list?

BernadetteLoscio: examples says [cites the BP]

<hadleybeeman> action annette to email SDW (and DWBP) to ask about their API work (with regard to examples for BP 10)

<trackbot> Created ACTION-237 - Email sdw (and dwbp) to ask about their api work (with regard to examples for bp 10) [on Annette Greiner - due 2016-03-04].

BernadetteLoscio: maybe we can use an API as example and show waht should be done with that API.

annette_g: like, saying what change would break the API?

BernadetteLoscio: yes

annette_g: it's a great thought
... I was not sure we should have an example of a maintained example

BernadetteLoscio: we could have a documentation page for the API

annette_g: a fake documentation page?

BernadetteLoscio: yes

annette_g: who would build and maintain it?

BernadetteLoscio: I don't know, but if we have real data, we could have a real or fake API.

hadleybeeman: we need to show that it is practical
... other groups need implementations in the wild
... but we don't need to rely on sthg that needs to be maintained
... we just need enough code to show what we mean
... so that devlopers can adapt it to their case.

BernadetteLoscio: yes

annette_g: what can we put? An API is a system

hadleybeeman: we're not telling people how to build their APIs, aren't we?

annette_g: no, but we give ideas.

hadleybeeman: so we're pointing to other APIs?
... examples on how to use this in the context of data on the web?
... We need to provide examples of the parts we're describing, not what others are describing

annette_g: there is no spec

hadleybeeman: is there any normative references to REST work?

annette_g: ref to functioning APIs or work about how to create them?

hadleybeeman: looking at BP21
... it's becoming complicated now.

<laufer> I cannot hear anybody now...

<annette_g> me neither

<BernadetteLoscio> me neither

<annette_g> restart the call?

<deirdrelee> sound gone, i'm talking away

<hadleybeeman> Maybe WebEx has crashed? I can't seem to rejoin the call

<Caroline> me neither

<hadleybeeman> Wait, now I'm back in.

<laufer> neither me

<annette_g> just finally got audio connected

hadleybeeman: I'm suggesting that Annette has a go at putting down what a developer should have in mind as a bare minimum
... when they're implementing this
... instinctively
... (that's about the BP for making changes)

annette_g: I really like BernadetteLoscio 's suggestion

hadleybeeman: as you wish

<laufer> again...

<riccardoAlbertoni> no sound again..

<annette_g> no

<laufer> it is a ghost...

<Caroline> no audio :(

<hadleybeeman> Right. Well, I was saying — we have 8 mins left.

<annette_g> *holy cow, tempus fugit

<newton> @annette_g we're thinking in put all BPs related to APIs into a subsection under Data Access.

<hadleybeeman> I suggest we leave the BP table for now — but it would be great if everyone working on the rest of the BPs make an effort to complete the table by next week

<hadleybeeman> (since we're already a week overdue)

<annette_g> yes, I want to see them all together, with the one about having an API in the first place coming first

newton: can we avoid huge changes in the document at this stage?

<newton> @annette_g the APIs BP will be "Data Access API"

<laufer> I have raised an issue about the BPs I am reviewing

sorry I meant @newton

<hadleybeeman> okay ———— hang on please

<hadleybeeman> Let's finish this meeting here on IRC

<hadleybeeman> a) as above: I suggest we leave the BP table for now — but it would be great if everyone working on the rest of the BPs make an effort to complete the table by next week

<hadleybeeman> (since we're already a week overdue)

<annette_g> Could we have a section called "Data APIs"? I think people will want to search for that.

<newton> @antoine we would only replace them into a subsection into Data Access. Won't create new ones.

<laufer> +1 to annette

<newton> Under Data Access section, to have a "Data Section API" as a subsection

<BernadetteLoscio> yes! that's our idea!

<newton> It will not reflect into a huge change, I would be a structural change

antoine seriously object to structure change and any re-shuffling at the time you're asking text to 10 contributors

<annette_g> Maybe make the change and then send around an email telling us you're done

<annette_g> so we can do a pull

<hadleybeeman> antoine, what are your concerns?

<laufer> We have the problem of the BPs numbers...

I'm afraid that we lose time figuring out what the impact of the changes are on the individual BPs

I won't have time to figure this out

so I won't work on my BPs

<BernadetteLoscio> but the Data Access section needed to be reviewed

<hadleybeeman> Aren't we only talking about the API BPs?

<BernadetteLoscio> its just on the Data Access section

<BernadetteLoscio> yes!

review doesn't imply changes while everyone else is working on the doc

<BernadetteLoscio> its just the Data Access section

<riccardoAlbertoni> +1 to antoine, laufer ..

<BernadetteLoscio> Data vocabularies section wont be affected

<hadleybeeman> antoine, if their changes won't affect the areas you're looking at... Does it matter?

At a minimum *Zero* BP numbers should change

<annette_g> things are going to be changing if multiple people are editing anyway

<Caroline> @antoine, I don't think this would stop you working in your BPs, it would be only putting APIs' together

<BernadetteLoscio> we're gonna make the merge

<BernadetteLoscio> on github

<annette_g> I still think APIs deserve their own section

<hadleybeeman> Re BP numbers, I'd suggest putting a temporary header on any merged ones that say, for example, "BP 10 and 21."

<BernadetteLoscio> we're gonna review the numbers when necessary

<hadleybeeman> we can renumber at the very end

<BernadetteLoscio> yes

<newton> +1 to hadleybeeman suggestion

<hadleybeeman> antoine, does that work for you?

<annette_g> All the static, unwebby stuff in our doc gets so much text and multiple sections; I think we need to give the dynamic stuff fair treatment

I'm ok with renumbering at the very end. Just please don't pull the carpet under our feets while we're working.

<laufer> I agree that APIs deserve a section but we have to be careful in changing the BPs order now...

<hadleybeeman> Totally fair.

<hadleybeeman> Okay — we're out of time.

<newton> We can track every change and show which BP Numbers has changed

<hadleybeeman> You all are champions for continuing this without sound!!

I don't have the time to follow your tracking

<BernadetteLoscio> we wont change the structure of the document, its just the Data Access section

<annette_g> when moving things around, that is the time to re-order API BPs

<BernadetteLoscio> mainly API BP

<Caroline> thank you @antoine and @laufer for this concern, we will work on it

<laufer> Well, in this way we do not need a scribe... it is an advantage...

<hadleybeeman> @newton, I think it might keep everything more sensible now to keep all the numbers as they are, even if they end up in the wrong order or multiple numbers for one BP.

<hadleybeeman> True, laufer :)

<laufer> bye antoine...

<BernadetteLoscio> we dont refer to numbers in our document

<hadleybeeman> We will, I promise, make sure it's all lovely and pretty (with numbers in order!) before it goes out to the public.

<BernadetteLoscio> we always refer to the title of the BP

<BernadetteLoscio> to the identifier

<annette_g> I find it awkward that the idea of having a API at all comes so late. It seems strange to start telling people how to make an API first and then come back to that fundamental point

<hadleybeeman> True, BernadetteLoscio — but we do in our actions/issues and notes.

<newton> We will do it in a different branch or repository, and if the group agrees, we will merge it to the main document

<BernadetteLoscio> the number is created automatically

<hadleybeeman> that's frustrating. :(

<BernadetteLoscio> if we also use the title, then it wont be a problem

<hadleybeeman> We haven't always thus far, which is part of the issue.

<hadleybeeman> I think we need to keep them for the moment, if we can -- even if it ends up being an extra bit of text

<BernadetteLoscio> but just BP abou Data Access and APIs will change

<Caroline> we will do our best to not get in the way of what authors have been working on

<BernadetteLoscio> others will keep the same

<BernadetteLoscio> the order of the rest wont change

<hadleybeeman> That sounds great, Caroline. Please do that, for now.

<annette_g> others will be renumbered, though

<annette_g> because BP 10 has to move

<BernadetteLoscio> ah ok

<Caroline> if we see that this change may get in someone's way, we will wait to do it at the end. Would that be okay @annette?

<annette_g> sure, as long as we do it

<Caroline> great! Thank you! :)

<hadleybeeman> Brilliant. I'm sorry we have to end this meeting!

<BernadetteLoscio> ok! thanks!

<BernadetteLoscio> thanks a lot!

<hadleybeeman> Have a wonderful weekend, all. And thank you again for switching to text!

<laufer> I will send and e-mail about the issue I raised...

<newton> @annette_g I have some other issues to talk to you, may we have a call early next week to talk about it?

<hadleybeeman> Talk next week.

<Caroline> thank you all!!

<hadleybeeman> thanks, laufer.

<hadleybeeman> bye for now :)

<newton> @annette_g it's about the Content Negotiation issue

<annette_g> @Newton, sure, want to use Skype?

<riccardoAlbertoni> good weekend !!

<newton> @annette_g yes, Skype :-)

<annette_g> yay!

<Caroline> :))))

<annette_g> @newton, let's exchange email to coordinate

<newton> @annette_g ok, thanks :)

<annette_g> okay, bye!

<hadleybeeman> trackbot, end meeting

Summary of Action Items

[NEW] ACTION: antoine to review the voc aspect of BP9 [recorded in http://www.w3.org/2016/02/26-dwbp-minutes.html#action02]
[NEW] ACTION: newton to review annette's test for BP8 [recorded in http://www.w3.org/2016/02/26-dwbp-minutes.html#action01]

Summary of Resolutions

  1. accept https://www.w3.org/2016/02/19-dwbp-minutes
[End of minutes]