Meeting Record

There are some format problems with the chatlog. Please correct them and reload this page. They are labeled on this page in a red box, like this message.

It may be helpful to

Zakim IRC Bot: ok, trackbot; I see DATA_DWBP()11:30AM scheduled to start 35 minutes ago ←

Zakim IRC Bot: phila, I see DATA_DWBP()11:30AM in the schedule but not yet started. Perhaps you mean "this will be dwbp". ←

Zakim IRC Bot: ok, phila; I see DATA_DWBP()11:30AM scheduled to start 38 minutes ago ←

Unknown Caroline: Nathalia and I have called, but we are the only ones on the phone! ←

Hadley Beeman: our goal for TPAC is to have a rough draft of each of our deliverables before we leave ←

Bernadette Farias Loscio: re best practices, all the issues that were raised yesterday concern the best practices document ←

… we can discuss the issues, but for TPAC we had in mind to discuss the ToC and to have an agreement about the themes/subjects/sections for the ToC ←

Jeremy Tandy: q+ to ask whether we would like to develop some "in scope / out of scope" tests ←

… this can come from the discussion about the issues, because that helps clarify the scope & get agreement about what we’re discussing ←

… we need to discuss the issues and from that I hope we can get agreement about the scope for each section ←

Hadley Beeman: is that best done talking through the issues rather than the ToC? ←

… then it will be easier to describe the best practices for each requirement ←

Newton Calegari: -> https://www.w3.org/2013/dwbp/wiki/Proposed_structure#Table_of_Contents_DWBP Proposed Structure ←

… I’m not sure if that’s best done all together or to break up based on section ←

Hadley Beeman: I propose we start that together, then we can ask the question again ←

… 15-30 minutes of you introducing the work, then we can decide whether to split or not ←

Eric Stephan: I like the idea of starting the discussion on the ToC as a group ←

… so that we know what we’re working on, and we feel like we have consensus ←

… I’d rather stay as a larger group today instead of splitting off, because I’m not sure what the split is ←

Phil Archer: yesterday we went through the requirements in the use case document ←

… while that was being prepared, and since then, Ricardo from Italy, Lewis from NASA, have both put in new use cases ←

… yesterday’s discussion helps me see that some of the requirements should be ditched ←

… but where those new use cases present possible new requirements, if we had a timeboxed discussion, it might be worth looking at those quickly ←

Phil Archer: -> https://www.w3.org/2013/dwbp/wiki/Second-Round_Use_Cases Second round use cases ←

Zakim IRC Bot: jtandy, you wanted to ask whether we would like to develop some "in scope / out of scope" tests ←

Jeremy Tandy: I thought it would be useful to agree between us a couple of simple statements about what’s in and out of scope ←

… could we spend a small amount of time in plenary to discuss what those tests should be ←

Vagner Diniz: I’m afraid if we review the issues we’ll go back to the same discussion as yesterday ←

… I’d like to avoid going back to those discussions, just resolve the issues ←

… focus on the issue and how to progress it, not just go back on discussion ←

… I’d like to split that up by hour or something, so other people can be chair for a bit ←

Eric Stephan: with fjh in the room, there’s a great opportunity here for us to be thinking about how we can use other capabilities in the W3C ←

Hadley Beeman: ok, 5-10 mins on scope questions; then review new use cases until 10am, then 30 mins for BernadetteLoscio to introduce best practices ToC & structure ←

… then 10:30-12:00 to find issues in best practices & do work on best practices ←

Riccardo Albertoni: I like the schedule, please consider I am available via IRC for the next one Hour, and if you need further explanation about the LuSTRE USE case I can try to Jump in the discussion via Skype... ←

… “is addressing it encouraging people to publish or use data on the web” ←

… for the best practices document, what I’m afraid is that the things we need to explain will be in a non-normative section or a note ←

… I think we should try not to just remove things because they’re out of scope because they’re not testable ←

… but it’s something we should discuss with people, or say something about in a note ←

… because when we have the issues we’re trying to say let’s not include this because it’s out of scope ←

… can some things be removed because it’s out of scope, but other things because we can’t test we can’t have a best practice about that ←

Jeremy Tandy: I think scoping questions should have three possible answers: rec track, note, or bin it ←

… I’d like to see a differentiation about best practices about publishing data vs those that are specific to on the web ←

… to keep the scope of the work manageable, not just generally publishing data ←

… that might be hard because we want to provide as much information to people as possible ←

… but there’s a lot to do if we take a broad approach, so I’m advocating a narrow approach ←

… I’d say if they’re related by context to being testable, I’d consider that in scope ←

… the other thing to think about for testing is we really have to have implementations that can show that not only we’re publishing data on the web, but some evidence that people found that useful ←

… a use case that I introduced today, I’m planning an implementation for it ←

… not only publishing data on the web but were we successful in people using & reusing that data ←

… people are worried about publishing data eg metadata, sometimes it’s difficult ←

… the second thought is that the same thing can encourage reuse as encouraging [scribe missed] ←

… the requirement will encourage people to use data, I don’t know if we can encourage them to publish it ←

… but from a publishing perspective I’d say that what should be in scope is not only publishing data on the web but in a good way ←

… and if we’re doing recommendations & creating best practices, the scope should be about publishing data on the web in a good way ←

… eg about vocabularies, we can make recommendations about the reuse of vocabularies when publishign data ←

Phil Archer: BernadetteLoscio talked about contextualising some of the best practices ←

… a note is simply a document that we produce that isn’t tested, not part of rec track, not normative ←

… an alternative would be to include that within the recommendation, but mark sections as non-normative ←

… I think it will emerge about whether it’s easier to have everything in one document or to have separate documents ←

… ie just because it’s in a rec document doesn’t mean it has to be tested ←

… to pick up on ericstephan’s point, we have to show evidence that two organisations have implemented each recommendation ←

… when it’s technical that’s clear how to do; with a best practice it’s harder to prove that it’s been used ←

… my expectation is that the evidence the group collects will be looking at what people have done, but the evidence for it being used outside the people in this room is it being used in other people’s guidelines ←

… eg in SharePSI, 40 EU partners, including governments, academics, enterprise, SMEs etc ←

… their contractual obligation to you is that before June 2016 they’ll update their guidelines to refer what we’re writing here ←

… we have people from Spanish government, people from Albania, through to PwC, ODI, Open Knowledge, OGC, it’s a big group ←

… if there are recommendations that no one can actually use, you shouldn’t include it ←

Hadley Beeman: I’m interested in the distinction between description & context ie telling the story about what’s going on vs advice that can be tested ←

… I worry about suggesting updates to data at a particular frequency, or creating feedback mechanisms, where we can’t test it ←

… I feel like there are lots of other organisations that are spending time on the behavioural/economic/democratic implications stuff ←

… but I don’t think most of those discussions will help us make data on the web more reusable in a technical way ←

Eric Kauz: At GS1 we’re putting together guidelines about publishing product data on the web ←

Hadley Beeman: should we try to get wording like these questions that we can all agree to? ←

… it would be great to have some set of rules, whatever they are, because these questions do keep coming up ←

Bernadette Farias Loscio: if the requirement are in scope & are testable then they will be normative, otherwise they’ll be non-normative ←

Caroline Burle: https://www.w3.org/2013/dwbp/wiki/Proposed_structure#Feedback.28.3F.29 ←

Caroline Burle: this link I just put is Bernadette's suggestion to be discussed ←

Caroline Burle: I think BernadetteLoscio’s said, it would be nice if we could include in a document how we have these done already ←

… what I did is add in the link the questions that BernadetteLoscio was talking about ←

… they are doing 3 star, we want to say that 5 star open data is the best way to publish open data ←

Carlos Laufer: I agree with the third scope question, the word ‘testable’ is very different, the thing we have to see to confirm that the thing we’re doing is working ←

… when we have a standard, I can point at an application and prove that it’s working ←

… we have to confirm that people are using it; phila_ suggested that it’s to include it in their guidelines ←

… I don’t know if it’s a thing that W3C will see as the same thing as testing ←

Phil Archer: to a certain extent it’s up to you how you make that choice; it’s up to the Director whether he accepts that ←

… if it’s a guideline that is a nice idea, you have to prove that people agree and that it’s implementable ←

… you’re looking for evidence that the community finds it useful and has acted on it ←

… if you said ‘all datasets must be published in 3 formats with content negotiation’, you could test it but it isn’t going to happen ←

… saying something less concrete; you could offer evidence that people have included it in their guidelines, like GS1 or the geospatial community ←

Carlos Laufer: my worry is that they will have to put it in the guidelines before it gets approved ←

Hadley Beeman: it is a bit of a negotiation process because in Last Call we put out the document for comments & we get feedback ←

… but it might be that CR shows we don’t have implementations, which means we have to go back ←

Hadley Beeman: I think this needs to go into best practices, but it’s important to have somewhere else too ←

… these ideas & our commitment to them needs to be above our documents, it needs to feed into everything we do ←

Frederick Hirsch: my familiarity is that Rec track you have do testing of things in code & implementations ←

Jeremy Tandy: q+ to ask if we should review these scoping points against the charter ... is there anything we won't now deliver? ←

Phil Archer: From the charter - To advance to Proposed Recommendation, evidence will be adduced that each of the best practices have been recommended in at least two environments, such as data portals and formal policies. ←

Phil Archer: The vocabularies that will eventually be developed by the working group will be published via the W3C Vocabulary Management Process that offers stability of terms but the flexibility for future evolution in response to community demand. ←

… because everybody will produce some best practices so everyone needs to know how we’re going to test them ←

ACTION: newton to research and report ways that technical best practices have been tested ←

Trackbot IRC Bot: Created ACTION-112 - Research and report ways that technical best practices have been tested [on Newton Calegari - due 2014-11-07]. ←

Phil Archer: you should find every BP in mobile has at least 2 green bars in the table ←

PROPOSED: Do you agree that one of the scoping criteria is “Is it unique to publishing data on the web?”? ←

Phil Archer: if it's not unique to the Web then others are as able to make recommendations as we are ←

Phil Archer: But... it's a wiki, it can be changed over time if the Wg so decides of course ←

Unknown CarlosIglesias: yep but you can recommend BPs that are useful for the web but also for other means as well, not necessarily unique for the Web, no? ←

Phil Archer: it could end up that under each best practice we have a line that says ‘how this encourages reuse’ ←

Carlos Laufer: also the first question, we have to justify against these criteria ←

Hadley Beeman: @CarlosIglesias: We are trying to agree our scope, so that when someone suggests "We should include [a new topic]!", we can say "yes, it's in scope" or "no, it's not" ←

Phil Archer: one of the questions about the first one is whether the key word is ‘data’ or ‘web’ ←

Eric Stephan: We are talking about the second question about reuse of data on the web @Caroline ←

… ie is it about uniqueness wrt publishing data rather than documents, or wrt to publishing on the web rather than through other means ←

Zakim IRC Bot: jtandy, you wanted to ask if we should review these scoping points against the charter ... is there anything we won't now deliver? ←

Jeremy Tandy: once we’ve agreed these, maybe go through the charter and see if there’s anything that won’t then be delivered ←

ACTION: Hadley to check the scoping criteria against the deliverables/commitments in our charter ←

Trackbot IRC Bot: Created ACTION-113 - Check the scoping criteria against the deliverables/commitments in our charter [on Hadley Beeman - due 2014-11-07]. ←

Bernadette Farias Loscio: publishing data documents that I create & give to someone vs just publishing data ←

… just publishing on the web might be documents. They have to be together: “data on the web” ←

PROPOSED: Do you agree that one of the scoping criteria is “Is it unique to publishing data on the web?”? ←

Phil Archer: We've been discussing your point, Carlos. Consensus is that both data and Web are keywords in that statement ←

RESOLVED: Do you agree that one of the scoping criteria is “Is it unique to publishing data on the web?”? ←

Unknown CarlosIglesias: completely agree, my concern is with unique, not any other keyword ←

PROPOSED: Do you agree that one of the scoping criteria is “Does it encourage reuse of data on the Web?”? ←

… we are proposing for data on the web, but people can use in other contexts ←

… CarlosIglesias was concerned about the ‘unique’: it doesn’t mean it can’t be used in other contexts ←

Unknown CarlosIglesias: then it is perfectly ok for me, but don't understand why the unique is there ←

Carlos Laufer: I think we argue, we’ll talk about this for hours, so I accept ←

… to me you have to split into two: encourage publish yes, encourage reuse too ←

PROPOSED: Do you agree that one of the scoping criteria is “Does it encourage reuse or publication of data on the Web?”? ←

RESOLVED: Do you agree that one of the scoping criteria is “Does it encourage reuse or publication of data on the Web?”? ←

PROPOSED: Do you agree that one of the scoping criteria is “Is it testable?”? ←

RESOLVED: Do you agree that one of the scoping criteria is “Is it testable?”? ←

Jeremy Tandy: https://www.w3.org/2013/dwbp/wiki/Proposed_structure#Table_of_Contents_DWBP ←

Newton Calegari: @fjh this is the link: https://www.w3.org/2013/dwbp/wiki/Proposed_structure ←

Bernadette Farias Loscio: will proposal for table of oontents for best practices doc, audience, acope, backghround ←

Caroline Burle: the link for the Proposed Structure with the Table of Contents is https://www.w3.org/2013/dwbp/wiki/Proposed_structure#Mapping_of_Themes.5B1.5D ←

… note that we have many requirements of form, ABC is available, meaning that the metadata is available ←

ISSUE: Do we include versioning in the BP doc? Currently there are no use cases for it ←

ISSUE: Should we include data enrichment in the BP doc? Currently there are no use cases for it ←

ISSUE: We have use cases for feedback but is it in scope given the 3 aximoatic criteria resolved at TPAC ←

Phil Archer: The feedback section in the structure is at https://www.w3.org/2013/dwbp/wiki/Proposed_structure#Feedback.28.3F.29 ←

Caroline Burle: Feedback(?) Data usage feedback Data to be published Lifecycle ←

… data consumption side can use encriched data after aggregating with meta data ←

Carlos Laufer: concerned if encrichment is embedded in app, don’t have linked data yet ←

Phil Archer: q+ to talk about data enrichment at chairs' discretion. (I have raised the issue in tracker already) ←

Trackbot IRC Bot: ISSUE-72 -- Is data enrichment in scope for dwbp? -- raised ←

ACTION: Adrianoc to complete use case and requirements for data enrichment. Related to ISSUE-72 ←

Trackbot IRC Bot: Created ACTION-114 - Complete use case and requirements for data enrichment. related to ISSUE-72 [on Adriano Pereira - due 2014-11-07]. ←

Jeremy Tandy: outlines process to review use cases, requirements then make decision ←

… goals, common understanding of each section and secondly people should review sections and be prepared to contribute ←

Caroline Burle: https://www.w3.org/2013/dwbp/wiki/Proposed_structure#Mapping_of_Themes.5B1.5D ←

Eric Stephan: https://www.w3.org/2013/dwbp/wiki/Main_Page#Best_Practices (this is what we are looking at... ←

Caroline Burle: I suggest to also look at https://www.w3.org/2013/dwbp/wiki/Proposed_structure#Mapping_of_Themes.5B1.5D ←

Jeremy Tandy: proposing to organize around challenges rather than chartered deliverables ←

Phil Archer: Bernadette refers to the charter technology-agnostic approach to cover aspects such as: ←

Frederick Hirsch: to rephrase - no clear linkage between charter and challenge/use case documents ←

Phil Archer: this is very helpful progress, especially since we are reflecting use cases ←

… we updated use cases, so you are going to have to update this table of contents ←

Phil Archer: discvoery metadata vs content metadata is one example to think about ←

Jeremy Tandy: what is useful of this structure is that it is tied to data lifecyle ←

… are we sure we want to lose some requirements, Shakespeare wrote something that says we need a plan ←

Phil Archer: we should still have a top level non-normative policy section both to give advice and to enable buy in ←

Zakim IRC Bot: hadleybeeman, you wanted to question the wisdom of keeping the policy stuff ←

Hadley Beeman: disagree, thinking about other web standards like HTML, CSS etc ←

Eric Stephan: we are at a very high level, could add lots of non-normative text ←

ISSUE: Is it in scope to include mention of policy framework etc. as part of the non-normative discussion/editorialisation of the BP doc ←

Trackbot IRC Bot: Created ISSUE-74 - Is it in scope to include mention of policy framework etc. as part of the non-normative discussion/editorialisation of the bp doc. Please complete additional details at <http://www.w3.org/2013/dwbp/track/issues/74/edit>. ←

PROPOSED: the best practices document should focus on the technical audiaence ←

Hadley Beeman: I'm wondering if we all have different ideas of what "best practices" mean ←

Phil Archer: q+ to talk about 'difference' between BPs for publishing CSV and RDF ←

Unknown yaso: they are best practices for **publishing, using, reusing** data on the Web, aren’t they? ←

Ig Bittencourt Santana Pinto: The mission of the Data on the Web Best Practices Working Group, part of the Data Activity, is: ←

Vagner Diniz: just because W3C doc does not mean obliged to be technial, see this ←

Unknown CarlosIglesias: remember the wcag? one of the most famous and widespread w3c documents ←

… we are only addressing technical people, given how we are writing this document ←

Zakim IRC Bot: phila, you wanted to talk about 'difference' between BPs for publishing CSV and RDF ←

Phil Archer: should stay abstract and not go into detail on CSV, say that it should be discoverable and reusable through metadata. Reference CSV WG for details ←

… “these are meant to be technical guidelines for those implementing a policy" ←

Hadley Beeman: worried that policy will distract group and also consume time etc etc ←

Ig Bittencourt Santana Pinto: I sad I agree with Vagner_Br about technical audience and it is clear in the charter (mission). ←

… legal risks, ethical aspects, especially when international, this would be a huge job ←

Eric Stephan: +1 hadleybeeman - just a mention that publication is motivated and dictated by policy is okay with me ←

Hadley Beeman: most policy wonks don’t understand technology, but we could make our documents useful and reassuring ←

Phil Archer: Working on this... The primary audience for our work is technical. That technical work should be in context ←

Caroline Burle: I am wondering if is it worth to discuss the audience and if it would be publishers and users ←

Jeremy Tandy: @Caroline ... will come back to you question about publishers & users in a moment ←

Caroline Burle: need better explanation on feedback related to consumers and publishers, not sure there was an agreement ←

Trackbot IRC Bot: ISSUE-71 -- We have use cases for feedback but is it in scope given the 3 aximoatic criteria resolved at tpac -- raised ←

… defining interface between publishers and users and how to communicate the data that is shared ←

Jeremy Tandy: BernadetteLoscio can you please summarize next steps for today so we can move this forward, today ←

Bernadette Farias Loscio: 1. agreement from group that we can organize document around challenges from the use cases ←

Caroline Burle: BernadetteLoscio is talking about the Maping of Themes: https://www.w3.org/2013/dwbp/wiki/Proposed_structure#Mapping_of_Themes.5B1.5D ←

Caroline Burle: Would be great to have more people contributing where there are TBD ←

Caroline Burle: +1 to have a short description of each theme for the contributor ←

Ig Bittencourt Santana Pinto: The scructure of the XML Signature Best Practices is very interesting and split the best practices according to the user. ←

Jeremy Tandy: request that all think during lunch about volunteering and also specifically about what they need to understand ←

Unknown yaso_: https://www.w3.org/2013/dwbp/wiki/Proposed_structure#Feedback.28.3F.29 ←

... with the following structure: BP Title, What, How, Why, How to Test and Evidence ←

Jeremy Tandy: the WCAG have an example of how to test and actions for someone to validate that ←

Ig Bittencourt Santana Pinto: jtandy told that you way to approach in the BP document is that in spite of tell how we could describe what is the intended outcome. ←

ACTION: ericstephan to investigate how WCAG can be an example for our BP document ←

Trackbot IRC Bot: Created ACTION-115 - Investigate how wcag can be an example for our bp document [on Eric Stephan - due 2014-11-07]. ←

Nathalia Sautchuk Patrício: I think it would be cool to each best practice indicate which use case inspired the practice like I put in "Technical factors for consideration when choosing data sets for publication"" ←

Nathalia Sautchuk Patrício: https://www.w3.org/2013/dwbp/wiki/Technical_factors_for_consideration_when_choosing_data_sets_for_publication ←

Frederick Hirsch: spec-prod is the W3C mailing list to discuss ReSpec features, etc ←

Hadley Beeman: q+ to ask a different question about how we make best practices ←

Zakim IRC Bot: sorry, deirdrelee, I do not recognize a party named 'ipcaller' ←

Zakim IRC Bot: hadleybeeman, you wanted to ask a different question about how we make best practices ←

Carlos Laufer: what is the reason to do a BP? It's to encourage a person to use that thing ←

... the question is: how to take the conversation with Tomas (COMURI doc) and align that with this propose ←

Unknown yaso: I was going to propose to put this document on github and open it ←

Unknown yaso: should we make more clear about what to do about the doc of comuri ←

Phil Archer: is carrasco follows the best practices then the doc can be published ←

ACTION: phila to develop into something that we can reuse and make more like a best practice ←

Trackbot IRC Bot: Created ACTION-116 - Develop into something that we can reuse and make more like a best practice [on Phil Archer - due 2014-11-07]. ←

...there are two issues:one is the content and the second the separate doc and the format issue ←

... it is extremelly painful as an editor to keep changes of the changes in the doc ←

...the core reason is twofold: automatic generation is extremely valuable... ←

Zakim IRC Bot: hadleybeeman, you wanted to suggest a vote on PROPOSED: Each best practice will use this structure: http://philarcher.org/dwbp/bptemplate.html (so that we have a record of the ←

Frederick Hirsch: my key points - however you edit best practices, make sure you can automate numbering or it is painful as you add/remove/move practices. Also good to have a best practices index, a list of best practices linking to them ←

Frederick Hirsch: This can be automatically done in ReSpec, but let the ReSpec team know you are using the feature if you do! ←

... we need to apporve the set of documents or section..and we have to vote if the best practices are according... the group should vote ←

Hadley Beeman: we propose that all best practices will be in one doc, because it is easiers ←

PROPOSED: all of our best practices will be in one document, because it is easier for readers/implementers to understand and easier for editors/contributors to manage ←

Phil Archer: i want to say somtehing... one of the reasons for supporing the proposal that we have ←

Carlos Laufer: we have agreed that we have a template... to use just only document ←

Eric Stephan: it will be confuse if we have a set of documents... things should be in one place ←

Jeni Tennison: i think two points... i think the uri design is an issue...in a uk we have a huge discussion about the design of URIs ←

...the second is to make sure we dont just consider URI design for linked data... the purpose of publishing datasets ←

...to take an action. to explore actions that clarify that need for URI design ←

ACTION: phila to explore the use cases that might clarify the need for URI and URL design ←

Trackbot IRC Bot: Created ACTION-117 - Explore the use cases that might clarify the need for uri and url design [on Phil Archer - due 2014-11-07]. ←

PROPOSED: all of our best practices will be in one document, because it is easier for readers/implementers to understand and easier for editors/contributors to manage ←

...this is gonna make incorporating the ideas ofcarrasco's COMURI work easier ←

RESOLVED: all of our best practices will be in one document, because it is easier for readers/implementers to understand and easier for editors/contributors to manage ←

Phil Archer: s/http:\/\/philarcher.org\/dwbp\/bptemplate.html\/http:\/\/w3c.github.io\/dwbp\/bptemplate.html/g (warning: replacement failed) ←

… I’m not sure if we’re going to have enough people to write just one section ←

Hadley Beeman: We are looking at https://www.w3.org/2013/dwbp/wiki/Proposed_structure#Feedback.28.3F.29 ←

Eric Kauz: data type usage, and I was looking more for onthology perspective ←

… even when the format that you mentioned is obsolete the advice is still useful ←

Bernadette Farias Loscio: what will be the content of that section. We were thinking that we’re gonna have this today at tpac ←

… and we raised a lot of issues. Those issues will have a lot of impact in each item of the proposed structure ←

… I don’t know if we can discuss the items after or before discussing the issues that we raised before. ←

Newton Calegari: q+ to propose to take a look on raised issues and try to fit them on these themes ←

ig_bittencourt my question is related to the focus that we will give to the data vocabularies ←

… one of our perspectives was about the data vocabularies creation, another one about using data vocabs ←

DWBP f2f meeting, day 2 at TPAC

Minutes of 31 October 2014

1. Scoping Questions

2. Best Practices Table of Contents