Data on the Web Best Practices Working Group Teleconference

<deirdrelee> trackbot, start meeting

<trackbot> Meeting: Data on the Web Best Practices Working Group Teleconference

<trackbot> Date: 12 February 2016

<deirdrelee> not too many people on call....anyone else from irc joining

<deirdrelee> scribe: PWinstanley

<phila> Decision - only a handful of people, so we'll have an informal chat

<riccardoAlbertoni> i am trying

deirdrelee: Last week's minutes: accepted?

<deirdrelee> PROPOSED: Accept minutes of last week's meeting https://www.w3.org/2016/02/05-dwbp-minutes

<Yaso> +1

+1

<deirdrelee> +1

<phila> +1

<Caroline> +1

RESOLUTION: Accept minutes of last week's meeting https://www.w3.org/2016/02/05-dwbp-minutes

deirdrelee: Agenda: go bak to the BP doc and talk aout implementation

<deirdrelee> PWinstanley: I'd offered to do some work on bps, but would like to know what is the current workflow?

<phila> PWinstanley: I'd offered to do some work on the BPs, but what is the current workflow for working with these?

PWinstanley: What is the workflow for working on BPs?

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

PWinstanley: how do I get access to Github?

phila: make edits in own repo, give a pull request, then let editors know and they will do the merging

PWinstanley: Fork repo, edit, and issue pull request

deirdrelee: In Chairs meeting we talked about implementations for the BP doc, so that it gets to candidate rec, we need examples of c. >2 of each BP

phila: because it is a BP rather than a tech spec, the issue is what do you mean by an example...
... mobile web BPs asked people to try to implement and then report back
... the other way (Share-PSI) is to link to handbooks
... we can make the case to the Director if an official doc presents a BP then if our advice is consistent then we can infer validity
... we need evidence that people agree with us

deirdrelee: 3 things: 1/ "evidence..."; should the evidence be from outside the group?

phila: external validation is always better. We can use the Zagreb F2F to get some of this
... however, internal evidence is also good. We can look at Scottish or Irish or other guides that WG members are involved in ... that is good too
... we need >2 implementations per BP

deirdrelee: 2/ when we go through there might be BPs that are more challenging to get examples - so we should pick up those early so that we are not rushing around at the last minute trying to get them

phila: before candidate rec we mark "at risk", and they can be removed. If we don't do that we are back to working draft if we cannot find the examples

BernadetteLoscio: wht if the implementation doesn't work?

phila: if it becomes obvious that it is not working we rewrite or remove
... there is no minimum period for candidate rec.... we can turn round in a few weeks

BernadetteLoscio: so we can edit the doc?

phila: yes, but ideally you don't edit much. Once cand rec ends, there is a call with a Director, and the Chairs/Editors prove wide review and examples. Thereafter it is pretty much done

deirdrelee: how should we start collecting implementations? google doc? Suggestions please

phila: use the wiki or github? but a google form might be a good idea. I have a model from share-PSI

<BernadetteLoscio> Phil, is this one: https://docs.google.com/forms/d/1I-pGDOotOR4adCHwnayX6qYqme1K-UUHQgxuADNPQA0/viewform?c=0&w=1?

PWinstanley: if thereis a BP that doesn't 'work' then add it as an antipattern

<phila> Yes, that one BernadetteLoscio

annette_g: are our use cases illustrations of implementation?

phila: I guess so, but we need validation from others having implemented them

Caroline: it is important to make the google doc/form - I don't think we will have time to handle emailed material

BernadetteLoscio: about the Use Cases. I am afraid of using these as evidence as we used these for requirements and the challenge from which we developed BP. We need to avoid the circularty
... it is a circular argument

deirdrelee: to create an action for BP editors to develop a process that we can start using from next week

Caroline: we can have 2 actions, one for editors and another developed from the Share-PSI quesitons (so we are not starting from the ground up)

<deirdrelee> ACTION for phila to send bp editors implementation-questionairre template

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

<deirdrelee> ACTION: phila to send bp editors implementation-questionaire template [recorded in http://www.w3.org/2016/02/12-dwbp-minutes.html#action01]

<trackbot> Created ACTION-229 - Send bp editors implementation-questionaire template [on Phil Archer - due 2016-02-19].

<deirdrelee> ACTION: Caroline to create process for gathering evidence of implementations, e.g. wiki, google form [recorded in http://www.w3.org/2016/02/12-dwbp-minutes.html#action02]

<trackbot> Created ACTION-230 - Create process for gathering evidence of implementations, e.g. wiki, google form [on Caroline Burle - due 2016-02-19].

deirdrelee: back to BP issues...

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

BernadetteLoscio: Yaso and Annette - I need help with the section on APIs, it is still open and there needs to be agreement re: the proposed changes. Needs to be done ASAP

annette_g: I was planning to help

BernadetteLoscio: do we need an action? when we are going to ask for implementations, the BPs need to be stable. We have many BPs in this section. The work needs to be done before we ask for evidence

<Caroline> thank you annette_g :)

deirdrelee: target date is 19 Feb - next Friday. The following week we should be able to put out the call for implementations

phila: This needs a convo with the Director. It may only be done after Zagreb. Transition to CR is done after isses and actions are all closed. The WG has to think it is finished.

deirdrelee: so the main focus is on clising issues and comments

BernadetteLoscio: the table by Newton is for examples and test. The BP needs to be reviewed too - the whole thing needs a review, not just examples and tests

deirdrelee: for the next 30 mins - open issues?
... comments first please

<deirdrelee> Open comments: https://www.w3.org/2006/02/lc-comments-tracker/68239/WD-dwbp-20150224/

deirdrelee: 10 open comments. What is the status?

Caroline: I think we need to go down the list ....
... most of them are old

phila: given that they are so old, ...

BernadetteLoscio: some can be closed - the ones from Christoph, for example
... the ones from Eric remained open because I didn't know how to answer
... others have a resolution already

phila: I think this needs to be discussed on the Chair's call on wednesday. they need to be closed, but that depends on whether they have been answered

BernadetteLoscio: I think most already have a resolution.

deirdrelee: is there a need to contact the author again.

phila: in the case of Eric's then the comments are being worked on, but we need to focus on process - the group needs to be satisfied that the comments have been worked on

deirdrelee: going through them now

annette_g: can I wrap things up with Eric? what's the protocol

<Caroline> :))))

phila: just chat with eric and sort it out, then reply back. that will work

<deirdrelee> ACTION: annette_g to talk to Eric Wilde about open comments and reach resolution [recorded in http://www.w3.org/2016/02/12-dwbp-minutes.html#action03]

<trackbot> Created ACTION-231 - Talk to eric wilde about open comments and reach resolution [on Annette Greiner - due 2016-02-19].

<deirdrelee> https://www.w3.org/2006/02/lc-comments-tracker/68239/WD-dwbp-20150224/3048

<annette_g> but but but

<Zakim> phila, you wanted to highlight http://w3c.github.io/dwbp/bp.html#EnrichData

Caroline: the data enrichment document will be prepared elsewhere and the link put into the BP, but this has not yet been done

<phila> BP on enrichment

phila: annette, are you aware that there is a BP in the doc, it is light on content, but needs completing

annette_g: there are more ways of enriching data than simply adding more metadata

BernadetteLoscio: this BP needs to be complemented

annette_g: my initial problem was that it seemed to be looking only at adding metadata or doing some machine learning, but data enrichment can also be about e.g. segmenting a visual dataset to isolate patterns that have meaning. This whole side is missing

phila: rather than it being for Annette, it should be for Giselle and there is not enough interaction with the WG so we will have problem in finding evidence. The action is on UFMG but it is quesionable that this BP will survive without significant effort

<laufer> +1 to phil

phila: but it is for others to do this

<annette_g> I'm on the queue

annette_g: this is a BP I care about. maybe it isn't only UFMG, and perhaps we need to merge their point of view with mine

BernadetteLoscio: what annette is describing is different to what UFMG were thinking - they are more ML people so enrichment by addition of metadata was their 'thing'
... we need a definition here to help clarify the topic

Caroline: maybe annette could collaborate with them?
... would this work? do we still have time?

annette_g: I am worried about workload, but can give it a try

<phila> Personally, I think the conversation with Erik Wilde should take a higher priority

<BernadetteLoscio> It should be possible to perform some data enrichment tasks in order to aggregate value to data, therefore providing more value for user applications and services.

<Caroline> +1 to phila

BernadetteLoscio: we can have the discussion by mail - but we need clarification on the semantics of 'enrichment'/ I think the requirement is closer to what annette was describing
... i agree that the chat with eric is important, it should take priority

deirdrelee: 2 others for annette .... can these be looked at as well please.
... last few minutes....

BernadetteLoscio: we have comments from Maurino Andrea - I contacted him before publicaiton of the second draft. I think these can be closed. Chairs can confirm.
... I can message the list with the links

<Zakim> phila, you wanted to ask for help with a JSON problem

<phila> action-228

<trackbot> action-228 -- Phil Archer to Fix bpconfig.js to restore contributors to bp doc -- due 2016-02-12 -- OPEN

<trackbot> http://www.w3.org/2013/dwbp/track/actions/228

<phila> BPConfig file

<annette_g> I could look at it

<Yaso> I can help

phila: I need help with an action to fix (action 228) - I cannot debug it. The BP config file isn't showing up the list of contributors

<Yaso> :-) We need an action for that or it's ok?

annette_g: in talking with eric I will focus on the API stuff, but there is more that Eric commented on. Will that be dealt with at the Chairs' meeting?

deirdrelee: perhaps the BP editors should check first, then writing Eric a note

BernadetteLoscio: we already provided comments in the tracker. Open ones are because we don't know how to handle them

annette_g: we need to see if we have replied to Eric ....

BernadetteLoscio: we had a lot of discussion with him on the mailing list. his messages were replied to promptly
... I am going to go through all the comments and send the group a message indicating status for each

deirdrelee: thanks to all for the useful meeting.

<laufer> bye all...

<riccardoAlbertoni> bye, have good weekend|

<annette_g> bye!

<Caroline> bye! great scribing PWinstanley!