W3C

Data on the Web Best Practices Working Group Teleconference

11 Mar 2016

See also: IRC log

Attendees

Present
PWinstanley, phila, newton, hadleybeeman, antoine, riccardoAlbertoni, ericstephan, deirdrelee, yaso, annette_g
Regrets
Carol, Bernadette, Laufer
Chair
Hadley
Scribe
annette_g

Contents


<hadleybeeman> https://www.w3.org/2016/03/04-dwbp-minutes

PROPOSED: Accept last week's minutes

<hadleybeeman> +1

<newton> +1

<antoine> +1

+1

<phila> +1

<ericstephan> +1

<RiccardoAlbertoni> +1

RESOLUTION: Accept last week's minutes

hadleybeeman: first item: closing items before the f2f. We need to try and do that by Monday

<hadleybeeman> html

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

hadleybeeman: BP status review?

newton: on the status table, we have some BPs that need to be done still.
... need some input from ericstephan

ericstephan: BP 30 is done, example and tests

<ericstephan> is this done now Newton? Or is there anything missing from the example I need to include?

<newton> Best Practice 30: Gather feedback from data consumers

newton: BP 30 still needs a little more work

<newton> Best Practice 31: Provide information about feedback

ericstephan: I provided an example in a google doc for one of the editors to include it. I think it's complete.
... I thought I provided BP31 also.

will check

ericstephan: I'm not sure the status of BP23.

hadleybeeman: are you the right person to look into that?

ericstephan: I've been kind of distracted with DUV stuff, could use some help from someone else on the example for real-time access.

hadleybeeman: anyone?

ericstephan: annette_g?

annette_g: I'm feeling similarly overbooked

PWinstanley__: I have some data that might work for this, from sensors

ericstephan: one problem I"m having is that there is a BP for keeping data up to date already.

PWinstanley__: I was thinking about river levels, not by the microsecond, but every 15 or 20 minutes. I can contact my colleagues to find out what the score it.

<ericstephan> PWinstanley__ thank you!

<scribe> ACTION: Peter to check on using an example about real-time data for BP23 [recorded in http://www.w3.org/2016/03/11-dwbp-minutes.html#action01]

<trackbot> Created ACTION-239 - Check on using an example about real-time data for bp23 [on Peter Winstanley - due 2016-03-18].

<hadleybeeman> phila: ericstephan's point about real-time and keeping data up to date... they can mean the same thing.

<hadleybeeman> ...For river levels, real-time can be every 15 minutes. The frequency is context-dependent.

<ericstephan> good points phila

phila: what's real-time depends on the data

<ericstephan> didn't catch the last part...

newton: is the real-time data the same as a data stream?

annette_g: I think it's not necessarily, but it can be

<ericstephan> streams like twitter feeds?

hadleybeeman: the "stream" only appears one place in our doc, talking about activity streams

<ericstephan> we could do this in real time

hadleybeeman: what else?

newton: BP18 si assigned to Joao Paolo

he seems to be tied up

hadleybeeman: can someone else take this on?

antoine: I was surprised that this was assigned to him. I can help, but not before next week.

newton: what does this do to the discussion next week?

<Zakim> phila, you wanted to talk about Joaoa Paulo

phila: I had an email from Joao Paulo saying he will join for a part of the f2f.

hadleybeeman: we can't really assign to him in absentia

newton: we can work on it during the f2f. That's better than no work on it. Berna and I can help, too.

antoine: Since I wrote that BP, that's why I was surprised it was assigned to him.

newton: should I change the name on the BP to Antoine?

hadleybeeman: sure.

newton: BP 32, enriching data
... Adriano and Adriano are assigned to it, but they aren't involved lately

yaso: they agreed that they are not responsible for the BP. It's up to the group now.

I agree with annette_g that it's an important one.

maybe we should close the action and try to talk about it in the f2f.

hadleybeeman: if we're to keep it as a BP, I'd like to have someone on it

yaso: I"m afraid to take it on and not be able to do it.

phila: my solution is Occam's razor.

it would take a significant effort to complete it.

phila: annette_g seems to be the remaining interested person

<hadleybeeman> annette_g: I'd rather try to fix it than just drop it.

<ericstephan> It could be picked again in a future community group

annette_g: I'd rather have it on my list than see it dropped

antoine: I can provide an example for this. I sent my example on the chat in a previous meeting, but there was no reaction. This may mean we should drop it.

<antoine> For info this is the example I mentioned: https://docs.google.com/document/d/1JvjrWMTpMIH7WnuieNqcT0zpJAXUPo6x4uMBj1pEx0Y/

ericstephan: I think the timing is not good. It might not be worth it for something that's more a cool idea.

I'm for dropping it

hadleybeeman: there is a way for the group to drop it and not lose it

<Zakim> hadleybeeman, you wanted to ask Newton when we need to stop accepting new text into the document

hadleybeeman: newton, when is the cutoff for new items?

newton: not sure we can add new text even now

* really???? :(

hadleybeeman: does it make sense to give annette_g a few days to try?

newton: I need to talk with the other editors

hadleybeeman: annette_g, are you going to zagreb?

annette_g: just remotely for a few hours

hadleybeeman: can you work on it over the weekend

<scribe> ACTION: Annette to work on the BP for enriching data before the f2f [recorded in http://www.w3.org/2016/03/11-dwbp-minutes.html#action02]

<trackbot> Created ACTION-240 - Work on the bp for enriching data before the f2f [on Annette Greiner - due 2016-03-18].

<newton> @annette_g are you going to improve the Data Enrichment section or only in the BP?

antoine: offers to help if needed as well.

annette_g: I think just the one BP

<newton> BP 22: Serving data and resources with different formats

newton: BP 22, serving data with different formats

<phila> - http://w3c.github.io/dwbp/bp.html#Conneg BP 22

<newton> http://w3c.github.io/dwbp/bp.html#Conneg

hadleybeeman: I can take it.

<scribe> ACTION: Hadley to review BP 22 [recorded in http://www.w3.org/2016/03/11-dwbp-minutes.html#action03]

<trackbot> Created ACTION-241 - Review bp 22 [on Hadley Beeman - due 2016-03-18].

<newton> https://www.w3.org/2013/dwbp/wiki/F2F_Zagreb_-_Agenda_Proposal_DWBP

newton: back to the table. If you have red around your name in it, you need to complete that action.

<deirdrelee> will do!

<deirdrelee> def finished by monday

<ericstephan> Or Sunday midnight here

newton: phila, can we close 94?

phila: yes

<newton> Close this issue. The group agrees that we shouldn't recommend specific tools in our BP. Maybe we can include an example about using git for dataset versioning. Should we include the example?

hadleybeeman: we nee to answer the question about whether to avoid specific technologies

phila: in the implementation section?

hadleybeeman: see issue 94

<phila> issue-94?

<trackbot> issue-94 -- Dataset versioning and dataset replication -- open

<trackbot> http://www.w3.org/2013/dwbp/track/issues/94

PWinstanley: an issue with versioning is whether you're doing versioning with diffs or with complete sets. we can point to using something like Git as an approach to the latter.

hadleybeeman: we need to get some language

PWinstanley: In the various approaches toward versioning systems, there are certain general principles and ways in which you can overcome issues like forking and merging.
... presumably we can come up with something that deals with that

phila: this was first raised on the use cases doc

hadleybeeman: does any part of this need to be settled for the BP doc?

newton: I don't think so.

<phila> close issue-94

<trackbot> Closed issue-94.

hadleybeeman: Let's just close it

<ericstephan> https://www.w3.org/2013/dwbp/track/issues/148 close

hadleybeeman: issue 148. ericstephan can we close it?

ericstephan: yes

<phila> close issue-148

<trackbot> Closed issue-148.

issue 220 was bernadette's and she wants to close it

<phila> issue-220?

<trackbot> issue-220 -- Should we include a more complexe example to illustrate provenance? -- open

<trackbot> http://www.w3.org/2013/dwbp/track/issues/220

<ericstephan> I am having troubles hearing

newton: I can create an action on myself to put in the example that Antoine suggested

from issue 220

<phila> close issue-228

<trackbot> Closed issue-228.

<phila> close issue-220

<trackbot> Closed issue-220.

<scribe> ACTION: Newton to follow up on issue 220 [recorded in http://www.w3.org/2016/03/11-dwbp-minutes.html#action04]

<trackbot> Created ACTION-242 - Follow up on issue 220 [on Newton Calegari - due 2016-03-18].

hadleybeeman: we closed 228

232 is done

<phila> close issue-232

<trackbot> Closed issue-232.

hadleybeeman: issue 149 is yours

<deirdrelee> yes

<phila> close issue-149

<trackbot> Closed issue-149.

<phila> issue-160?

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

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

<phila> Hmmm... big issue

<hadleybeeman> annette: I feel like we have it covered by having APIs... but if you're posting data outside of that, as a blob, maybe there is value in giving people accesss to fragments

<ericstephan> It is huge

<hadleybeeman> hadley: this is less of an challenge in linked data -- but we're talking about more than just that.

<ericstephan> Its really hard if you don't have a specific domain in mind

<hadleybeeman> phila: this is coming up in spatial data on the web

<hadleybeeman> ... My Occam's razor approach is to say the group recognises it's really important. Maybe mention it somewhere?

<PWinstanley> isn't data cube vocab an approach?

<hadleybeeman> ...But to be practical, we could hand it to sdw

<hadleybeeman> annette_g: Is is specifically spatila

<hadleybeeman> phila: no

<hadleybeeman> PWinstanley: Data cube slices fit in here too.

<ericstephan> if there is momentum in the spatial data then we might be able to grow something from there. We are dealing with subsetting on the temporal level

<hadleybeeman> phila: in spatial data, a slice isn't enough

<hadleybeeman> PWinstanley: We could have some simplistic cases that are widely used

<hadleybeeman> ...How you push data into a leaflet map ... critical for sortgin out good practice

<hadleybeeman> antoine: I'm not sure data cube slices sort the problem. they can exist in one dataset.

<ericstephan> subsetting a graph (subgraph) another issue, mathematical forests (trees of data another)

<hadleybeeman> ...Also, if it's an important issue, I think we should have a BP. Perhaps make it super simple, saying different domains have different criteria for subsetting;

<hadleybeeman> ...point to sdw for a specific example

<hadleybeeman> ericstephan: I'm concerned that if we don't have a really good approach, it may not have an impact.

<hadleybeeman> ...Maybe we should think about it more broadly in the future? I'm sceptical that it won't have much impact.

<hadleybeeman> phila: we could have an entire doc about this

<hadleybeeman> ...slicing isn't engouh; you also need to dice it

<hadleybeeman> ...You also need to think re page media

<hadleybeeman> ...In the spatial world, that could give you a number of tiles back.

<hadleybeeman> ...It could reduce the granularity of the data until it fitted... There are a lot of options here.

<hadleybeeman> ...I'm looking for something realistic we could say.

<hadleybeeman> ...Other groups have finished with a wishlist; CSV on the Web has set up a community group...

<hadleybeeman> ...This is a substantial topic in its own right.

<hadleybeeman> ...If we can say something substantial, we should -- but if not, we can do other things.

<Zakim> phila, you wanted to talk about CGs, wish lists, next steps

<ericstephan> Something that needs to be considered as well is the impact on the implementation report

antoine: I think we can still have an impact if we simply say that it's helpful to provide subsets
... if we just say something simple, that could bring a useful message.

hadleybeeman: we need to wrap this up.

who can take this forward?

<ericstephan> The only part might be to think of datasets and distributions. Thats the only thing that comes to mind

hadleybeeman: is there someone who wants to write now?

<ericstephan> address later

annette_g: I'd like to give it a try

hadleybeeman: something more for your weekend

also think about what could be left as future work

<phila> Zagreb agenda

Zagreb

hadleybeeman: last issue, during the f2f, we have times when we're just talking about BP. Do vocab editors want to do breakouts at those times?
... monday morning in particular

<RiccardoAlbertoni> +1 to break on monday morning

phila: thinking about Antoine in particular

<ericstephan> my concern is that Berna is going to be in the BP discussions, and Sumit won't be attending

antoine: I'm behind it

<ericstephan> from the DUV perspective

hadleybeeman: eric, that puts it in the middle of the night for you

ericstephan: crazy as it sounds, I still want to attend remotely
... not sure how much sense it makes to have overlap

might work for DQV more than DUV

<ericstephan> thank you and safe travels all

hadleybeeman: sounds like we should have DUV folks in with BP people and DQV separate

<deirdrelee> thanks hadleybeeman !

<newton> bye! thanks

<deirdrelee> see you all next week

<newton> see you on monday

<RiccardoAlbertoni> bye all!

*bye!

<yaso> thank you, Hadley! and tks annette_g ! Bye all

<newton> or maybe sunday

<hadleybeeman> bye all!

Summary of Action Items

[NEW] ACTION: Annette to work on the BP for enriching data before the f2f [recorded in http://www.w3.org/2016/03/11-dwbp-minutes.html#action02]
[NEW] ACTION: Hadley to review BP 22 [recorded in http://www.w3.org/2016/03/11-dwbp-minutes.html#action03]
[NEW] ACTION: Newton to follow up on issue 220 [recorded in http://www.w3.org/2016/03/11-dwbp-minutes.html#action04]
[NEW] ACTION: Peter to check on using an example about real-time data for BP23 [recorded in http://www.w3.org/2016/03/11-dwbp-minutes.html#action01]
 

Summary of Resolutions

  1. Accept last week's minutes
[End of minutes]