GLD Face to Face meeting Day 1 -- 11 Apr 2013

<PhilA3> we are connected to zakim

<MakxDekkers> I am in Barcelona, late start 10 am

<MakxDekkers> can we get visual maybe through skype?

<MakxDekkers> my apologies for falling ill, would have loved to be there

Quick review of European data Forum

<PhilA> bhyland: Invites those who were at the EDF to give a quick review

<PhilA> Christophe: Lots of people saying that big data is just data. Goof talk from Siemens.

<PhilA> ... I did a demo on the 2nd day

<PhilA> ... had lots of visits from Dutch people

<PhilA> BartvanLeeuwen: It was interesting. less tech than I usually go to, more polcy than I'm used to

<PhilA> ... lots of stuff about big data but there was a lot about semantics of data

<PhilA> ... lots of talks about linked data and big data in the same project

<PhilA> ... blown away by talk from Daimler - who knew they were using linked data

<PhilA> bhyland: So good to hear the Daimler CEO coming out and using the term linked data.

<PhilA> Boris: +1 to Bart. People were asking about best practices for LD

<PhilA> bhyland: Have you put your company in the directory, Boris?

<PhilA> Boris: Not yet, I thought it was gov only. Will fix

<PhilA> cygri: I only saw the second half of the second day. Great venue

<PhilA> cygri: I've only known Croke Park as a sports venue. Weird irish sports get played here. Didn't know about the conference centre

<PhilA> bhyland: Europe's 4th largest stadium 82K people

<PhilA> cygri: Hurling is the biggie here - a cross between hockey and murder

<PhilA> HadleyBeeman: I only saw some of the presentations. Some of seemed the messages we've been saying and hearing for a while. Shows importance of our work and related work

<cygri> (here's an excellent recent hurling game played in croke park: http://www.youtube.com/watch?v=vai3Gzd-ilw )

<PhilA> HadleyBeeman: Lots of projects were EU funded - lots of projects seemed to be straight translations from one language to another. Lots of machine translation

<bhyland> PhilA: Multilingual issues are more prevalent in EU discussions than the greater LD community has dedicated time & effort to. This doesn't make sense given the depth of multi-lingual expertise that W3C WG's possesses.

<bhyland> … We determined we could make DCAT available in at least 6 languages by 12noon today.

<bhyland> Hadley: John Sheridan said legislation.gov.uk is about to be launched in Welsh shortly.

<bhyland> s/legislative.gov.uk/legislation.gov.uk

<PhilA> BartvanLeeuwen: Maybe we've failed to highlight the multilingual nature of LD

<MakxDekkers> quality not briklliant here

<MakxDekkers> sound quality I mean

<MakxDekkers> can hear about 50% of what is said

<PhilA> Deirdre: The feedback I was getting about the networking etc. was positive. Lots of use cases and the like shown

<PhilA> bhyland: What was the most challenging part

<PhilA> Deirdre: the EC was behind it so funding wasn't the biggest problem. Other sponsors brought validation to the event

<PhilA> ... we targeted different communities wrt. invited speakers

<PhilA> ... only the second year but the reputation is growing. Athens in 2014

<PhilA> bhyland: It was excellent. I went to a NIST conference a while back - it was free, in Maryland. Bumped into LDP WG, lots of standards bodies

<PhilA> ... very academic. Lots of PhDs, more like the Siemens, StatOil talks from EDF

<PhilA> bhyland: Maybe more tracks are useful at these things but it's also good to get everyone together. Conferences don't always do that

<PhilA> bhyland: Interesting that privacy and security issues are so important here. Not talked about nearly as much in the US

<PhilA> HadleyBeeman: The Data Protection Directive is an issue - not an American concept

<PhilA> bhyland: We have nothing like Directives, or FP7 funding

<PhilA> ... lots of cases where one project didn't know about what another was doing - when they actually are doing the same thing

<PhilA> bhyland: Makes me think better about the Community Directory as a poss tool to help that.

<PhilA> Deirdre: On the issue of it being less tech than expected that's useful feedback for next year

<PhilA> bhyland: Odd to hear about European Linked Data (isn't it global?)

<PhilA> Room explains EU funding to Bernadette

<HadleyBeeman> We're now reviewing the semantics of clotted vs whipped cream.

<HadleyBeeman> And what soda bread is/tastes like. Interesting ramifications of localisation there.

<PhilA> Deirdre: Interesting point about non European funding, how to get EU funded and others linking/working together

<PhilA> cygri: Just pointing to the agenda...

<PhilA> bhyland: We're agenda bashing...

<PhilA> ... and collecting observations from the conference

<MakxDekkers> can't really hear what he is saying

<MakxDekkers> maybe just me?

<PhilA> Ghislain: I thought it was excellent. I didn't get a huge output for myself but I wasn't expecting much. Really enjoyed it

<MakxDekkers> OK will try to follow on IRC

<MakxDekkers> what was that

<DaveReynolds> Only if it's easy, audio is the main thing

<PhilA> Discussion of setting up a Google Hangout for streaming

<PhilA> Please put your Google e-mail address so we can invite you in

Agenda bashing

<PhilA> Boris: I have to leave before lunch

<PhilA> Christophe: I leave shortly after lunch

<PhilA> Others here for the duration

<MakxDekkers> try to have luch when you ahve

<MakxDekkers> yes

<martinA> your 1600

<PhilA> cygri: Topic to discuss - what's going to happen with the WG?

<PhilA> cygri: We've been told that we are not getting a charter extension as things stand. So we're rushing to get things done in time

<PhilA> cygri: If going to a 2nd Last Call means that's the last thing we have time for then that affects tech decisions and that's not the right way round

<DaveReynolds> +1 to cygri request, key issue to understand

<bhyland> +1 because it impacts much of how we proceed today and for next month

<BartvanLeeuwen> +1

<HadleyBeeman> PhilA: for the group to get an extension (here), we have to get everything to CR.

<MakxDekkers> in hangout now, can see you!

<HadleyBeeman> … Then, come the end of May, we can be reasonably confident that we can have an extension.

<HadleyBeeman> …The extra effort that has gone on recently has made a big difference.

<HadleyBeeman> … What comes after that is a topic for debate (on agenda for tomorrow). What do you want to do next? We in the W3C have been discussing the possibilities.

<HadleyBeeman> … Conversations were primarily on "is RDF finished"? If so, what impact does that have on our Semantic Web Activity?

<bhyland> … W3C Mgmt having serious discussions regarding maturity of RDF specifications, as well as broader Open Data on the Web

<HadleyBeeman> … What I've been putting in is that we need to talk about open data, and linked data within it. We also need to talk about data validation (which is why we have the Open Data on the Web workshop)

<HadleyBeeman> … That workshop will be important in setting the agenda for what comes after this working group's work.

<HadleyBeeman> … And what working groups will come out of this. A number of people (including DaveReynolds) have been talking about closing the world for RDF. A workshop is being planned in the autumn to explore that.

<HadleyBeeman> … We need to broaden the scope to talk about open data. Break down the religious wars between JSON and XML, etc.

<HadleyBeeman> … This group is making very good progress to getting towards finishing its charter. By end of GLD charter, deliverables should be in CR. Most of what we're talking about is already implemented. We shouldn't have too much complicated discussions. Therefore, it's realistic to think that three months after the end of May, we could have the rec-track docs as full recommendations. At the end of that period, the new working group will be in play.

<HadleyBeeman> … This planning has to go through the W3C planning process, consultations with us, and the membership as well.

<HadleyBeeman> bhyland: summer is misleading; we're talking about may 31 onwards.

<HadleyBeeman> phila: Our charter extension will probably be for June, July and August, and will say something along the lines of "this is to get the rec-track documents to the end of the process."

<HadleyBeeman> PhilA: Perhaps I should say, the key is "are the documents at the stage, by the end of May, that we could reaonsably see them getting to Rec within three months"

<PhilA> DaveReynolds: So if, say, DCAT isn't quite at CR by 31 May, does that mean we don't get an extension?

<PhilA> PhilA: No, it's not that hard and fast. The test is, are documents sufficiently stable that they could realistically be moved to Rec by the end of a 3 month extension

<PhilA> cygri: So we have to decide whether we want to actively avoid a second last call. How much of an influence should process be?

<PhilA> cygri: I'm not as optimistic as PhilA to get everything to CR by the end of May if we go through a second LC

<PhilA> ... some of the comment were quite substantive and might take several weeks to get through

<PhilA> ... I don't want to jeopardise other specs because DCAT isn't quite finished on time.

<PhilA> bhyland: W3C is reasonable. It won't shut things down that are active and making progress

<MakxDekkers> point of order: how does this discussion relate to the agenda? Not trying to be pedantic but trying to organise my time today and tomorrow.

<MakxDekkers> OK

<PhilA> bhyland: The reason Richard asked is because it frames the discussion he's going to be leading soon

<PhilA> bhyland: We have similar issues around the BP doc. Timing affects some decisions on what is left in vs what we can get done

<PhilA> HadleyBeeman: I'd also add on DCAT that my priority is on a higher quality piece of work than rushing

<PhilA> cygri: And we'd all rather produce a quality piece of work than something rushed out to a schedule

<PhilA> HadleyBeeman: There are limits to that of course but in general, plus or minus a little then we're OK

<PhilA> cygri: We got a number of comments that we need to go through for DCAT and the speed is an important factor

Agenda bashing for today & tomorrow

<PhilA> bhyland: Anything we need to change

<PhilA> HadleyBeeman: I'm sure that a conversation on licensing etc. would be interesting but I'm not sure it's a priority for us

<bhyland> http://www.w3.org/2011/gld/wiki/F2F3#Thursday.2C_2013-04-11

<PhilA> bhyland: I took that out a while ago...

<PhilA> HadleyBeeman: Presses F5

<PhilA> bhyland: In theory we're going to do the BP doc in 10 mins time

<PhilA> ... we have an hour for Cube. Enough Richard?

<PhilA> cygri: Yes, we only have one real issue to discuss

<PhilA> DaveReynolds: Agree, an hour should be enough

<PhilA> fadmaa: Yes, an hour or so is enough

<PhilA> PhilA: Hides head in shame when suchkect of URI presistence comments somes up 'cos I keep promising and not delivering

<PhilA> bhyland: May cut to half an hour as we need John E

<PhilA> HadleyBeeman: Something I want to accomplish in the next 2 days is to re-do our timetable planning

<PhilA> ... a bash at the timetable page would be good before we finish

<PhilA> bhyland: Editor should take the responsibility to update their timetables

<HadleyBeeman> (Example for reminder: http://www.w3.org/2011/gld/wiki/DCAT_Timetable )

<PhilA> Coffee break

<bhyland> until 10.30 Dublin time

<HadleyBeeman> "Caesar didn't have the option of following some British guys on Twitter, which would have changed the invasion completely."

<PhilA> scribe: fadmaa

Best Practices

<bhyland> see http://www.w3.org/2011/gld/wiki/BP_Timetable

<bhyland> https://dvcs.w3.org/hg/gld/raw-file/default/bp/index.html#source-data

bhyland: The BP document has a summary, list of checklists, diagrams. It then has a section on URI construction
... the fact that these recommendation can be done or not makes it different from other W3C recommendation
... I am suggesting removing the "checklist" section
... making the BP more normative
... and move the checklist to a separate note
... this also enhances our opportunity to have the BP done as a recommendation within the time frame

PhilA: this raises the question again of whether the BP should be a recommendation
... a recommendation also requires some implementation (independtly two implementations)
... which is hard
... checklist is really helpful and having them in a separate document make them hard to find

<martinA> Checklist for WCAG is very useful and it's a Recommendation

PhilA: it has been done to have a community group note that feels exactly like a W3C recommendation if it is done well

HadleyBeeman: Are notes easier to be carried on by subsequent groups?

<PhilA> Community group producing things that look like, but are not, standards http://www.w3.org/community/odrl/

BartvanLeeuwen: best practice is hard to be put in a solid way as they need to be updated

cygri: agree. With note the process of updating it is more flexible

PhilA: it is better to try to finish the BP within the charter of this group

<DeirdreLee> Google hangout available for visuals. If you want to join, post your email here

bhyland: taking URI process as an example, it is hard to point to implementations

HadleyBeeman: is Linked government data different than linked data in general?
... can't we consider an enterprise implementation
... as an existing implementation for the best practices for linked government data

<HadleyBeeman> or possibly have the scope of Best Practices to cover all linked data?

<HadleyBeeman> (to be determined by the group)

cgueret: would it be appropriate to refer to existing implementations from within a note?

PhilA: I don't think this is good especially that implementations are not guaranteed to continue to exist

boris: it will be hard to point to implementations of BP

bhyland: if we take into account enterprise implementations that won't be hard

boris: it is hard to prove that the implementation was good and the practices were helpful

PhilA: if the customer(implementer) is happy, we can trust this

HadleyBeeman: looks like we have three options: one recommendation, a recommendation and a note or a note?

bhyland: one recommendation is not an option given the time frame
... this leaves two options: a recommendation and a checklist separate as a note
... or one note

<gatemezi> http://www.w3.org/TR/mobile-bp/

<PhilA> www.w3.org/TR/mobile-bp/

bhyland: an example of a good way to put best practices: http://www.w3.org/TR/mobile-bp/

BartvanLeeuwen: BP will not contain a MUST and SHOULD kind of recommendations
... but the like of "we think it is good to do this and this"

PhilA: in the mobile bp document referred to, we required feedback from two different implementers
... it might be good to suggest including a "persistent URI" section for the likely new W3C WG dealing with open data in general
... this means we can make the URI construction part of the BP as a note

<sandro> Step 1 for talking the JSON folks: Don't use the terms "URI" or "IRI". Stick with "URL". :-)

<HadleyBeeman> wise words, sandro

bhyland: URI policy and implementation is something that I think this WG can and should do
... we have the required technical capabilities

DaveReynolds: I support the suggestion for going for a note containing the URI construction section
... I don't think having two separate docs is a good idea

<HadleyBeeman> HTTP range 14 compliance can of worms

DaveReynolds: and achieving consensus for a rec is risky, we might have the http range discussion opened again

<HadleyBeeman> :)

bhyland: I don't mind putting the BP in a note
... but I don't like the idea of defering this for a next WG

<DaveReynolds> Don't think that PhilA's suggestion was to defer, but to build on.

<DaveReynolds> Agenda says this discussion has 26mins left

PROPOSAL: put a BP as a note

sandro: is the argument that because URI construction is controversial we shouldn't try to get consensus on it?

DaveReynolds: I didn't mean that having a document on URI construction is a bad idea
... but having this within the best practice as a rec is risky

bhyland: +1

<sandro> DaveReynolds:We're spending months on a revision of the UK policy and we're not nearly done, so this WG isn't ready to produce a REC on URI Construction.

bhyland: I think it is better to have the BP as a comprehensive document and as a WG note

boris: giving the time restrictions, I also think that putting this as a note is a good idea

<DaveReynolds> +1

<HadleyBeeman> +1 to putting BP as a note

<cygri> PROPOSAL: The WG aims to publish Best Practices as a W3C Note

<MakxDekkers> +1

<BartvanLeeuwen> +1

<DaveReynolds> +1

<HadleyBeeman> +1

<gatemezi> +1

<martinA> +1

<cygri> +1

<sandro> +1 given that I'm hearing the editor says we can't possibly make it to CR by the end of May, or even the URI Construction part of it.

<DaveReynolds> To be clear. Very happy to have URI construction and policy as part of BP note. Having a REC track document that just focusses on that issue will be a challenge in terms of community consensus and timescale. UK experience is that despite several years use of previous UK recommendations, they need substantial update and that achieving consensus on that update is time consuming.

<boris> +1

<sandro> DaveReynolds, my question is whether or not we should ATTEMPT to meet that challenge.

<sandro> DaveReynolds, ... and I'm hearing "No."

<DaveReynolds> sandro: no, I'm saying "yes" but not in the next 3 weeks just from GLD perspective

<sandro> What if we had 6 months?

<HadleyBeeman> Looking further at URI persistence in the context of open data (especially with regard to JSON and XML) would be useful. Perhaps subsequent working group? At that point, it's out of the realm of Linked Data and RDF… but still worth doing.

sandro: If we get a 6 month extension, should we try to work on the URI construction recommendations?

bhyland: we have enough technical recommendations and guidance to put in a section on URI construction in a best practice note
... it can serve as input for further work

<sandro> bhyland: It would be confusing to take URI Construction out of Best Practices, into another document.

<cgueret> +1

bhyland: there are examples of notes which are widely used and even referred to as standards sometimes especially for people not aware of the W3C processes

sandro: if we have the URI construction doc separate, people will pay more attention
... and this will help getting this thoroughly discussed

<bhyland> Sandro - We want BP well documented and this is important.

<sandro> sandro: This is a major decision for the WG, so it needs to be well-documented and well-justified.

sandro: we need to clearly record the reason for our decision re. working or not URI construction

<bhyland> cygri: What we say on URI Policy for Persistence has wide reaching impact… for example LDP and RDF WGs

<sandro> cygri: What we say on URI construction can have a pretty big impact and affect lots of other WGs. Making those recommendations might also be in-scope for other WGs, like LDP. The RDF WG has views and opinions on these issues.

cygri: this has broad implications. one might argue that this falls also in the scope of other WG e.g. the RDF WG, the LD platform

<sandro> (It came about because of the UK guidance document.)

bhyland: the reason this was part of the GLD charter, is because its charter was written before the LDP one

<PhilA> that's the one DaveReynolds tells us that months of work is going into updating. It's referred to by many people.

cgueret: it is important to clarify whether there is something specific with government URIs in particular

cygri: some topics in the charter are hard to be addressed exclusively in the government scope

<PhilA> cygri: Makes the point that URI construction is not gov-specific (or any other domain). LDP is equally interested, for example

cygri: many of the issues are general and the technologies can be applied outside government linked data

bhyland: it is fine to impact other people and WGs and we can get input from them

cygri: my concern is that we might not have the right composition in the WG to address some broad topics
... that might go beyond the scope.
... there is a danger of not having all related people looking into our output and providing feedback

<sandro> " The group will specify how to create good URIs for use in government linked data. "

sandro: the charter states that government linked data specifically
... in the part referring to the URI construction

<bhyland> To quote the charter: URI Construction. The group will specify how to create good URIs for use in government linked data.

cygri: if I look at our draft I don't see it specific to the government data

bhyland: yes it is general

<sandro> bhyland, I think I halted the vote by asking this question -- about 6 months, etc.

<PhilA> PhilA: There's a session on URIs at the London workshop btw http://www.w3.org/2013/04/odw/agenda#uris. Some gov, but not all

<HadleyBeeman> It sounds to me like we can all agree that URI persistence section of our BP draft isn't finished. I suggest we have other conversations about the nuances of the content.

<sandro> DaveReynolds: If URI Construction were split out, we'd have to be much more careful and clear about the scope. Meanwhile, the UK stuff is Best Practice suggestions, *not* formal conformance spec stuff.

<sandro> I'm comfortable deferring to DaveReynolds expertise on that.

<bhyland> @Dave - so what you just said further suggests the URI Policy guidance should be in a WG Note, not a Rec track doc because it is not "pass / fail"

<bhyland> @Dave, do I understand you correctly?

<cygri> cygri: Potentially, all LD is in scope for LDP

<cygri> BartvanLeeuwen: Dutch government has a linked data working group

<sandro> BartvanLeeuwen: It'd be helpful to the Dutch group on this to have URI construction be a separate WG Note

<cygri> … having a separate document for URI construction would help highlighting it so that national groups don't reinvent things

<sandro> bhyland: I would not want to see URI Construction taken out of Best Practices.

<DaveReynolds> bhyland: yes, if it gets into lower details as the UK stuff does then that should be framed as a range of possible approaches rather than a one true way

<sandro> HadleyBeeman: I think URI Construction needs a lot more time and attention than we've got.

<DaveReynolds> PhilA: Don't know, in principle maybe, would need to discuss with the sponsors (defra)

<sandro> HadleyBeeman: Part of what needs to happen going forward with URIs is to (1) make it more government specific and (2) get more reviews from the WGs that might care about this

<bhyland> @Dave, then I agree with you & in discussion this AM discussion, believe even more so that it should be a WG Note.

<sandro> bhyland: yes and yes

<bhyland> Proposed: Publish Best Practices as as Working Group Note, inclusive of URI Construction guidance and URI Policy for Persistence, and various checklists for vocabs, etc.

<DaveReynolds> +1

<sandro> +1

<HadleyBeeman> RESOLVED: The WG aims to publish Best Practices as a W3C Note.

<MakxDekkers> +1

<cgueret> +1

<sandro> (We understand this is different from the Charter -- in that the charter says this will be a Recommendation.)

<HadleyBeeman> Charter: 2.2 Best Practices for Publishing Linked Data www.w3.org/2011/gld/charter

bhyland: reviewing the BP document as stated in the charter:
... Procurement we decided before not to include it in the BP doc

<DaveReynolds> Maintain my -1 on procurement checklist

bhyland: procurement as a checklist

<DaveReynolds> Well at least -0

cygri: it is hard to vote on whether it should be a checklist or no
... the group can vote based on teh content. I think there is no need for the WG to micro-manage this

HadleyBeeman: do we need formal process to drop parts of the charter?

sandro: we have discussions on emails but not a formal decision

<bhyland> Discussed that the Best Practices WG Note contain checklist information for procurement, vocab selection, and thoroughly URI Construction and URI Policy for Persistence. This WG Note does not expect to cover in detail Versioning, Stability, Legacy Data. Finally, the LD Cookbook will live as a wiki.

sandro: at some point we need some text explaining why we are not working on some parts of the charter

<HadleyBeeman> The charter commits us to delivering BP recommendations that we have discussed not having.

PhilA: we need to record that the WG has decided not to include some parts of the charter

<HadleyBeeman> We planned for sections on Procurement, Versioning, Stability and Legacy Data. WG members who were authorities on those topics have left, and we have further discussed that they are not relevant/feasible to produce to a high level.

<HadleyBeeman> bhyland: they are relevant, but we are only going to produce high-level summaries of them.

<sandro> PROPOSED: Best Practices will only very briefly discuss "4. Versioning", "5.Stability", and "6. Legacy Data. ". Also, "7. Cookbook" will be in a wiki. We don't have the time/expertise to do more.

<bhyland> +1

<HadleyBeeman> +1

<DaveReynolds> 0 (sounds like procurement checklist remains)

<sandro> PROPOSED: Best Practices will only very briefly discuss "1. Procurement", "4. Versioning", "5.Stability", and "6. Legacy Data. ". Also, "7. Cookbook" will be in a wiki. We don't have the time/expertise to do more.

<sandro> PROPOSED: Best Practices will (at most) only very briefly discuss "1. Procurement", "4. Versioning", "5.Stability", and "6. Legacy Data. ". Also, "7. Cookbook" will be in a wiki. We don't have the time/expertise to do more.

<DaveReynolds> +1

<sandro> +1

<sandro> (for DaveReynolds, the addition of "at most" clarifies that we might drop the Procurement section entirely.)

bhyland: the current section on procurement is not specific to government

<cygri> 0 (I don't see the point of this vote.)

<gatemezi> +1

<HadleyBeeman> +1

<boris> +1

<MakxDekkers> +1

<BartvanLeeuwen> +1

<cgueret> +1

bhyland: if we can't make it better I'd suggest not having it

<sandro> RESOLVED: Best Practices will (at most) only very briefly discuss "1. Procurement", "4. Versioning", "5.Stability", and "6. Legacy Data. ". Also, "7. Cookbook" will be in a wiki. We don't have the time/expertise to do more.

<sandro> We can change our mind in light of new information -- eg some Awesome new text.

<sandro> bhyland: This is largely guidance to the editors

<bhyland> Thanks all, very helpful to editors.

<sandro> HadleyBeeman: And this documents for the world that/why we're not doing these things in our charter.

<MakxDekkers> please try to keep break times the same. I am planning around that

<HadleyBeeman> noted, makxdekkers

Data Cube vocabulary LC feedback review

<DaveReynolds> http://www.w3.org/2011/gld/wiki/Data_Cube_LC_comments

<cygri> http://www.w3.org/2011/gld/wiki/Data_Cube_LC_comments

cygri: we have all the comments we received in teh Wiki page

Data Cube

cygri: many of them were expressing their used of the vocabulary and their satisfaction
... there are some editorial issues that the editors can take care of
... one thing that needs some discussion, is the question re. the SDMX part relation to the Spec itself
... and whether it is part of the Rec
... the main issue was regarding the heirarchical code list there was some push back

<cygri> ISSUE-59?

<trackbot> ISSUE-59 -- Last Call comment. Frank Cotton on qb:HierarchicalCodeList -- raised

<trackbot> http://www.w3.org/2011/gld/track/issues/59

cygri: the issue is mainly with hierarchical list which is common in statistical data e.g. geographic areas
... it is a frequent requirement to have this hierarchy represented in their data
... we use SKOS currently which can express hierarchy with broader/narrower
... the issue with this, is the existing domain specific hierarchies that are not represented using SKOS
... can we enable organisations to use the hierarchies they already have?
... one possible solution is to require these lists to be represented into SKOS
... but this will prohibit people from reusing their already defined data and URIs
... the hierarchicalcodelist property allows pointing to non-SKOS lists

<PhilA> scribe: DeirdreLee

cgueret2: shouldn't force use of skos for hierarchies, because skos is not enough

PhilA: how common is the problem that skos is not enough

cygri: 2 issues, first is forcing people to use skos and the second that skos is not rich enough
... if skos is not rich enough for your need, you can just extend it
... datacube does not specify that skos has to be used for hierarchy, however is you use skos discovering the hierarchical information is clear
... so datacube only 'forces' you to use skos if you are looking for this predictability
... more difficult to discover properties if skos is not used

PhilA: do we need a WG to extend SKOS?

cygri: there is a group that is looking into this

<HadleyBeeman> XKOS at the DDI: https://github.com/linked-statistics/xkos/blob/master/xkos.ttl

PhilA: What is the benefit of using Datacube if they don't use skos?

cygri: yes
... different kinds of hierachies should be represented using different model so that the hieranchical subtleties can be captured

<PhilA> PhilA: I am convinced by cygri. Thank you

cygri: proposed response to Frank 'we understand the concern, but we would like to provide a more flexible approach'

<cygri> 1. we would like to provide a simple way of using qb, although we understand that this simplification is not always appropriate

<cygri> 2. it's important to allow re-use of existing identifiers that orgs already have; forcing creation of parallel skos hierarchies not good.

<DaveReynolds> http://lists.w3.org/Archives/Public/public-gld-wg/2013Apr/0017.html

DaveReynolds: agrees with cygri

<bhyland> +1 to cygri's summary, always provide a simple way to use QB

DaveReynolds: cygri's point misses that there is information that skos cannot capture

<cygri> 3. even in SKOS there can be multiple hierarchies (e.g. containment and admin reporting)

DaveReynolds: there are genuine reasons for hierarchical representations other than skos

<cygri> rephrasing… 3. there are genuine reasons for using other relationships than SKOS broader/narrower, for example when there are multiple hierarchies

cygri: hierarchical codelist feature allows us to address this

DaveReynolds: Franck is right that it opens the door to confused representations, but it also opens the door to good representations

PROPOSAL: We keep the qb:HierarchicalCodeList and DaveReynolds explaining the reasoning for this

<cygri> +1

<fadmaa> +1

<gatemezi> +1

<BartvanLeeuwen> +1

<DaveReynolds> +1 (I will improve email based on this discussion)

<boris> +1

<cgueret2> +1

<martinA> +1

<MakxDekkers> got disconnected, now passcode 4531# is invalid

<PhilA> hmm...

<bhyland> Sorry Makx!

<MakxDekkers> OK back in

<MakxDekkers> yes

<HadleyBeeman> Cygri, is this still up to date then? http://www.w3.org/2011/gld/wiki/Data_Cube_Timetable

<MakxDekkers> yes fine now

PhilA: in the table of LC comments, there is evidence of use. How extensive is this evidence? Is it enough to get out of CR?

cygri: there are 2 kinds of implenations we want to consider, one is datasets, the second is consuming applications

<HadleyBeeman> RESOLVED: We keep the qb:HierarchicalCodeList and DaveReynolds explaining the reasoning for this

cygri: there are plenty of datasets that use qb, whose well-formedness can easily be tested based on spec

PhilA: it's not up to datacube to prove that it can be consumed

cygri: agree, but, for integration purposes, it is useful to show that there are consuming applications

PhilA: how long would it take to create a document reporting this?

DaveReynolds: What is missing is exactly what terms have been used
... but this is based on the CR criteria of all the vocabularies

<sandro> +1 we need a broader discussion of CR criteria. (I'm not very comfortable with ignoring consumers.)

PhilA: answer from cygri on if we can skip CR is 'no'

<sandro> (CR is both a validation of the market and of the technology)

cygri: if we wanted to, we could create a small implementation report, reaching out to community using qb
... but this engagement would take more than 2 or 3 weeks

bhyland: we will add an item to the agenda 'CR exit criteria'

<MakxDekkers> have to leave -- lunch is being served. back at 13:30 your time

<HadleyBeeman> cheers, makxdekkers

cgueret: : is there plans in qb to include openannotation for tagging?

cygri: this issue comes up often with lots of different vocabularies, for example prov, how to we treat their relations with other vocabularies
... better to address this issue generally for a link do any other vocabulary
... these are orthogonal issues, so should not be closely tied within the vocab spec

DaveReynolds: might it be worthwhile to document this in the bp guide?

<cygri> DaveReynolds++ great phrasing

DaveReynolds: interlinking vocabularies should be loosely coupled so they may be mix and matched

bhyland: like lego :)

cygri: discussion on datacube done

<HadleyBeeman> And to lunch. We're back here for 13:45 (1 hour from now)

<cygri> DaveReynolds: vocabularies as modular building blocks… don't artificially tightly couple them

HadleyBeeman: will reconvene at 13:45 Irish time

<martinA> ok, bye

<sandro> Hey Mike. They're at lunch until :45

<Mike_Pendleton> Thanks Sandro

<MakxDekkers> let us know when you resume

<MakxDekkers> can we get hangout visual back?

<MakxDekkers> thanks, hangout is back

<PhilA> Derirde will sort the Google Hangout thing shortly MakxDekkers

<MakxDekkers> I can see you

<MakxDekkers> zkaim, mute me

<MakxDekkers> hearing lots of noise

<HadleyBeeman> Welcome back from lunch

<MakxDekkers> feedback ;)

hey, if you want to join visually via google hangout, post email address

<bhyland> Agenda: http://www.w3.org/2011/gld/wiki/F2F3#DRAFT_Agenda

<martinA> Please, DeirdreLee martin.alvarez@fundacionctic.org

<MakxDekkers> OK now

<fadmaa> presentation for DCAT LC comments:

<fadmaa> http://www.w3.org/2011/gld/wiki/images/7/7b/Dcat-lc-comments.pptx

<HadleyBeeman> http://www.w3.org/2011/gld/wiki/images/7/7b/Dcat-lc-comments.pptx

<MakxDekkers> Not very well

<bhyland> We just moved the mic next to Fadi

<martinA> Better

<MakxDekkers> better

<martinA> good, thanks

DCAT Last Call Comments

<bhyland> Facilitator: Fadi

<bhyland> Scribe: Bernadette

<MakxDekkers> can faadi speak a little slower please

<bhyland> Slides for Fadi's pres: http://www.w3.org/2011/gld/wiki/images/7/7b/Dcat-lc-comments.pptx

<bhyland> Fadi plans to highlight the issues raised followed by more detailed discussion.

<MakxDekkers> please refer to slide number

<cygri> slide "Changes to existing terms"

<bhyland> Slide: 9

<gatemezi> slide 9: http://lists.w3.org/Archives/Public/public-gld-comments/2013Apr/0009.html

<MakxDekkers> what about slide 8?

<DaveReynolds> Very hard to follow. Was anything decided on dates?

<PhilA> no, not that I know of DaveReynolds

<bhyland> Slides 2-7 were all editorial mods that Fadi will handle.

<MakxDekkers> I strongly agree with Phil that using "01" is a bad idea

<bhyland> Slide 8: Noted but need to discuss.

<bhyland> Richard: As a procedural note, editors must be clear on action plan for 1) editorial changes (ed's just do it); 2) requires WG input - will be tough to do in R/T in F2F, suggest raising formal issues in Tracker

<bhyland> … Note: Substantial vs. non-substantial changes must be noted. In the case of *substantial* comments it implies we'll go to another LC so everyone has a chance of looking at the entire thing. No one should be surprised when it comes out as a Rec track deliverable.

<bhyland> … Modifications that cause changes to conformance are considered substantial.

<bhyland> … editorial changes are fixing: typos, improving clarification, and similar level changes — that is all non-substantive

<bhyland> Slide 9: Changes to existing terms, needs more discussion.

<bhyland> Slide 10: Also change to existing terms, submitted by Makx

<bhyland> Agreed, not substantive.

<bhyland> Slide: 11: Change to existing terms from Bill Roberts

<MakxDekkers> There is a proposed vocabulary but it was never finalised

<bhyland> Slide 12: Changes to existing terms from JeniT

<bhyland> Needs further discussion

<MakxDekkers> http://www.ukoln.ac.uk/metadata/dcmi/collection-DCCDAccrualPeriodicity/

<bhyland> Slide 13: Changes to existing terms

<MakxDekkers> going too fast for me

<cygri> http://www.w3.org/2011/gld/wiki/DCAT_LC_comments

<MakxDekkers> what was the conclusion on slide 13?

<bhyland> Needs further discussion (just arrived this AM)

<MakxDekkers> OK

<bhyland> Slide 14: Adding new terms from JeniT

<bhyland> Needs further discussion

<MakxDekkers> DCAT call would be good idea

<bhyland> ACTION: Fadi to organize a call with people who are able to help resolve open DCAT issues. [recorded in http://www.w3.org/2013/04/11-gld-minutes.html#action01]

<trackbot> Created ACTION-112 - Organize a call with people who are able to help resolve open DCAT issues. [on Fadi Maali - due 2013-04-18].

<bhyland> Slide 15: Adding new terms

<bhyland> Possible plan to follow what ADMS does but requires further discussion.

<bhyland> Slide 16: Adding new terms from ChristopherG

<bhyland> Possible plan to follow what ADMS or OpenOrg does but requires further discussion.

<MakxDekkers> Correct

<bhyland> Slide 17: Adding new terms from Bill R

<bhyland> There doesn't seem to be a vocab to describe granularity, so if DCAT offers it, we may have to come up with a new term.

<MakxDekkers> Not sure what "channelling someone" means?

<MakxDekkers> OK!

<bhyland> Slide 18: Adding new terms from Ed S.

<MakxDekkers> why not add a relationship to DCAT proper?

<MakxDekkers> lot of people seem to be asking for it

<bhyland> Fadi: Due to concerns of conformance, Fadi is concerned about adding "related" to dct because if will break things.

<MakxDekkers> can we further discuss this?

<bhyland> Yes!

<MakxDekkers> OK

<bhyland> Slide: 19

<bhyland> Scope of DCAT from Stuart H

<bhyland> Wants to broaden scope ...

<bhyland> Fadi: While many people ask for this, he is concerned about complexity as well as, others specs address this albeit not in RDF.

<bhyland> Requires further discussion

<bhyland> Slide 20: Scope of DCAT from AndreaP

<bhyland> From INSPIRE Directive team (see http://inspire.jrc.ec.europa.eu/)

<MakxDekkers> One issue is that peopel think that dataset = http://en.wikipedia.org/wiki/Data_set, so basically tabular data

<MakxDekkers> not defining is not always good

<bhyland> Fadi: Mailing list comments have inferred wider and broader definitions of dataset.

<MakxDekkers> waht about a dataset on paper?

<bhyland> ACTION: bhyland to add a dataset definition to the glossary (and cygri to help define it!!) [recorded in http://www.w3.org/2013/04/11-gld-minutes.html#action02]

<trackbot> Created ACTION-113 - Add a dataset definition to the glossary (and cygri to help define it!!) [on Bernadette Hyland - due 2013-04-18].

<bhyland> Discussion ensued about what the definition of a "dataset" is.

<MakxDekkers> isn't an API an access point to a dataset?

<PhilA> A favourite resolution of Dom in such situations is to 'remain silent'

<MakxDekkers> not defining creates confusion

<bhyland> Cygri: Believes we should avoid defining a "dataset", avoiding controversial edge cases about what a dataset is.

<bhyland> … the scope of DCAT is datasets (but we leave the definition of *what* a dataset is to the reader).

<gatemezi> @bhyland: And what will be that *definition* of dataset in a Glossary?

<MakxDekkers> if a dataset can be anything, just say that

<bhyland> cgueret: People can use it how they wish ...

<MakxDekkers> if a dataset can be anything, an API is a dataset

<MakxDekkers> never use circular definitions!

<DaveReynolds> Seems to me essence of DCAT is data resources that can be transmitted over a comms network. So code list, model etc is a dataset in that sense a person is not. Means that dcat:Dataset = information resource

<MakxDekkers> correction: an API is a distribution (access point) of a dataset

<DaveReynolds> So a dcat Dataset is anything for which http-range-14 doesn't bite :)

<cygri> Makx++

<bhyland> Fadi proposes dataset = "A collection of data, published or curated by a single source, and available for access or download in one or more formats."

<MakxDekkers> isn't an API not just an accessURL?

<bhyland> See above ;-)

<MakxDekkers> or API-URL as subproperty of accessURL

<sandro> "source" or "organization" ?

<sandro> "foaf:Agent" ?

<bhyland> Revision v1.0 - dataset = "A collection of data, published or curated by a single agent, and available for access or download in one or more formats."

<MakxDekkers> sandro, I made that comment

<MakxDekkers> agree with foaf:Agent

<bhyland> PhilA: Key stakeholders in greater Open Data on the Web world care about APIs and Access points … there are many ways to access data and we mustn't limit its use.

<gatemezi> according to the spec; dcat:Dataset dct:publisher foaf:Agent

<bhyland> cygri: Again, very concerned about having a precise definition because he doesn't want to exclude constituents who might otherwise use DCAT ...

<MakxDekkers> cygri makes things worse

<MakxDekkers> just say "a dataset can be anything"!

<bhyland> cygri - preferes to talk more about the purpose of DCAT rather than what a "dataset" is.

<bhyland> Fadi: Further clarification about properties for a distribution are necessary and will help reduce questions / confusion.

<bhyland> PhilA: Most people (at least PhilA) look at the diagram to reflect the spec.

<DaveReynolds> Discussion seems to be about APIs but Andrea Perego comment is not (just) about APIs, it's about whether dct Dataset includes code lists, video, software etc

<gatemezi> In the library world: a dataset is "collection of structured metadata — descriptions of things, such as books in a library. The equivalent of a dataset in the library world is a collection of library records." ..see http://www.w3.org/2005/Incubator/lld/XGR-lld-vocabdataset-20111025/

<DaveReynolds> Answer to that is "yes", isn't it?

<HadleyBeeman> To which, DaveReynolds?

<DaveReynolds> Andrea comment

<DaveReynolds> Discussion seems to be about APIs but Andrea Perego comment is not (just) about APIs, it's about whether dct Dataset includes code lists, video, software etc

<MakxDekkers> what about the text of a law? is that a dataset?

<bhyland> cygri: I have yet to see a proposed use for DCAT that won't work. Meaning, yes, you can use it for what you're planning but we hadn't previously anticipated your use case, but it would work.

<HadleyBeeman> makxdekkers I think it is on legislation.gov.uk

<MakxDekkers> OK fine, that's what I am going to need

<bhyland> cygri: DCAT can be used for a broad range of assets and DCAT is about creating a catalog of them.

<DaveReynolds> Andrea essentially asks if DCAT applies to "catalogues of any type of information resources". Jokes on http-range-14 aside, that seems right to me.

<MakxDekkers> Agree with Phil, of course!

<bhyland> PhilA: DCAT uses AccessURL to point to an API.

<MakxDekkers> Or define API-URL as a subproperty of accessURL

<cygri> "…by a single person, organisation or other kind of agent…"

<MakxDekkers> Thinking back to the discussion that led to downloadURL

<cgueret> Or define an API datatype

<HadleyBeeman> MakxDekkers: I asked JeniT (wrt her legislation.gov.uk work) re text of law as a dataset. Her response: "I'd more usually think of *all* the legislation making up the dataset, but for sure you could have subsets which could be as small as a single version of a single item of legislation"

<HadleyBeeman> Does that help?

<bhyland> PhilA: If there someday is a Open Data on the Web WG, would having WSDL for the open data community be helpful?

<MakxDekkers> Hadley, The initial answer was clearer: the text of a law is a dataset

<bhyland> sorry remote callers, Zakim rudely excused himself ...

<bhyland> we're dialing back in now

<sandro> we're here, yes

<HadleyBeeman> MakxDekkers: It was my definition, as opposed to representing theirs. (Which, it turns out, is a bit more nuanced)

<MakxDekkers> i'm still here

<PhilA> PROPOSAL - That the text be clarified to say that dcat:Distribution is very general. We provide properties defined to describe a particular Distribution type, downloads, and that it is hoped that future work will define extensions for other types of Distribution

<bhyland> cygri: Would like to see further editorial content to explain downloadable files in particular.

<bhyland> Fadi: Concerned that this change will break implementations.

<bhyland> Slide 21 & 22: Summary of Open Issues

<cygri> http://www.w3.org/2011/gld/wiki/DCAT_LC_comments

<bhyland> Summary: There are 23 issues on the wiki page and cygri proposes we get guidance from WG on what editors can safely edit vs. are possibly substantive ...

<DaveReynolds> Didn't we cover that while going through the slides?

<DaveReynolds> Most went down as "needs discussion".

<DaveReynolds> Probably those probably need to be opened as issues for tracking.

<PhilA> fadi: First point needs no discussion

<PhilA> .. next 3 do

<PhilA> List of issues is being updated live. Basic resolutions being recorded as to which ones need further discussion and which are editorial http://www.w3.org/2011/gld/wiki/DCAT_LC_comments

<MakxDekkers> agree with cygri, keep it as it is

<MakxDekkers> can't follwo

<cygri> http://www.w3.org/2011/gld/wiki/DCAT_LC_comments

<cygri> comment #17

<MakxDekkers> #17 OK

<MakxDekkers> #18 OK

<PhilA> for HadleyBeeman - you need http://lists.w3.org/Archives/Public/public-gld-wg/2013Apr/0009.html as evidence that 17 is resolved

<MakxDekkers> #19 OK

<MakxDekkers> #20 OK

<MakxDekkers> #21 OK

<MakxDekkers> #22 OK

<martinA> I'm sorry, I have to leave.

<HadleyBeeman> 10 minutes break

<MakxDekkers> be back in 10 mins

<martinA> See you tomorrow. Enjoy tonight's dinner

<MakxDekkers> passcode 4531# again invalid

<HadleyBeeman> Sorry, makxdekkers

<HadleyBeeman> makxdekkers, we're trying to find a way to get you in :)

<sandro> MakxDekkers, I bet it's just slightly-distorted DTMF. Try longer tones, or trying repeatedly.

<bhyland> http://www.w3.org/2011/gld/wiki/BP_Timetable#Detailed_timetable_and_checklist

Community Directory

<PhilA> scribe: PhilA

bhyland: Discussed timeline

HadleyBeeman: Updating the timeline itself

bhyland: It should be in good shape by the end of this meeting

HadleyBeeman: Suggest we stick to 29th as date of WG approval

sandro: We'll publish a FPWD earliest we can after 29 April and get comments
... then update the doc and publish that revised version as a Note

bhyland: Target date would be for 3 weeks for comments -> publication of Note on 21 May

<sandro> 1. FPWD 2. WGNOTE

<bhyland> See http://www.w3.org/2011/gld/wiki/BP_Timetable#Overview

sandro: The date to record really is the one for the meeting when the WG will make the resolution

<sandro> Hi PhilA

<HadleyBeeman> scribe: hadleybeeman

bhyland: Previous version of the Community Directory had a login; UX, admin and user headaches resulted. Over time 75 orgs registered. Beyond having a triple store and storing the data in RDF, it wasn't a linked data application. So we took on board the feedback and spent time re-doing it.

… Now it's what we have on dir.w3.org

<PhilA2> scribe: PhilA2

<gatemezi> ComDir: http://dir.w3.org/directory/pages/about.docbook?view

bhyland: Talks through the input form for the Directory (or rather the RDF generator, Foaf-o-matic style)

<sandro> bug report -- click on one of the contacts on http://dir.w3.org/directory/queries/org-view.rq?view&org=http%3A%2F%2Fwww.w3.org%2F2013%2F04%2Fgldcomdir.ttl%23org and you get five pop-up windows. Or I do, at least. (on firefox.)

bhyland: You get your dir.ttl file, publish that on a server somewhere and then tell the directory where to find it

PhilA2: I don't get those popups (Opera)

It's trivial, getting RDF on to people's sites etc.

<DaveReynolds> works on Chrome but do get 5 popups on firefox

bhyland: Putting the data there could be picked up by search etc.

Biplav: First of all thank you for putting the Directory online. I see IBM is there
... I wonder how a richer description could be created
... IBM's Web site has 130 country and language combinations. You see different things depending where you are
... the Web site is managed by multiple groups
... the basic entry - I'd like to update it with the homepages of the different regions
... the main homepage of ibm.com redirects you. Is there a way to extract out the dependencies from data, if provided, based on its location etc.

bhyland: I love the idea and I'd be happy for 3 Round Stones to compete against others to implement that for you
... But it goes well beyond what the directory is currently designed to do

<gatemezi> @biplav: why don't you suggest to describe IBM.com using org with all the sub-units?

<sandro> +1 vocab-org

Biplav: Part of managing the Web site is part of each unit's responsibility
... I was wondering whether we could use ORG etc. and so on to do this, distributed data management etc,

bhyland: So you'd like to see 130 different entries for IBM? That's going to skew the data set

<sandro> Shouldn't we have 10,000 listings, once things are going well?

<sandro> I mean, there are over 1,000,000 governments in the world.... :-)

<HadleyBeeman> +1 to Sandro

Biplav: There is no central catalogue for IBM, all the parts are locally managed

Furtehr discussion of Biplav's point about the way IBM's Web estate is managed

scribe: we want to offer the relevant product catalogue - there is no central data source

bhyland: It's well beyond the scope of the community directory
... we're just trying to provide a basic directory that can hook up different people working on the same thing

HadleyBeeman: You're pointing out that the directory only ever gives one result irrespective of the user's location. It would be great to be able to accommodate that power if we can.

PhilA2: wonders where GeoSPARQL is the way to go here?

bhyland: We need nice looking, easy to use, easy to understand stuff suitably branded that happens to be based on LD

I just got the popup bug sandro

BartvanLeeuwen: Does it only support turtle?

bhyland: I found out today that, for now, yes.

BartvanLeeuwen: I was thinking about RDFa coming from Drupal for example
... A Drupal plug in that you could say "here's our company contact page" and then the directory could go off and get it from the RDFa
... I notice that the update works (I've updated my entry)
... I wouldn't have the instructions for getting into the directory on the homepage. You want the directory itself
... And would adding more triples make it choke?

bhyland: Good question

BartvanLeeuwen: How about pointing to Open Corporates?

<sandro> and does opencorps provide a URI for each corp, so we can link?

<HadleyBeeman> Yes, sandro. Ex http://opencorporates.com/companies/ae_az/301890

bhyland: I like all these ideas

<sandro> cool, so what's the property to use to link to those..... hm.

DeirdreLee_: I thought it was easy to use, cool
... So it's not just gov linked data - good
... Can we provide a URL for an RDF catalogue for our products/projects?
... You want details of projects etc. We publish that already so it would be good just to point you to it

Deirdre and Bernadette continue to discuss various aspects

bhyland: The generator is just a foaf-o-matic. The directory will re-crawl the data every 6 hours so you can update it locally

BartvanLeeuwen: It would be good it it crawled so it would crawl multiple files/sources

bhyland: You want to be able to put ttl files across the site?

<sandro> or owl:import for something stronger than rdfs:seeAlso

BartvanLeeuwen: No, if my file references to other files then will it/can it harvest them too?

<gatemezi> @Sandro: owl:import is still in use out there ? :)

HadleyBeeman: Querying company type list. Seems a little limited so far

fadmaa: I like that I can get the entry specific RDF. Can I get all of it in one go?

<sandro> +1 folks should publish all the data they want, and let the directory filter!

<sandro> If the data is too big, use SPARQL.

<Zakim> HadleyBeeman, you wanted to start the conversation about what's next for the Community Directory

I did this recently - similar http://philarcher.org/diary/2013/euromap/

agreement that the problem is projection, not the implementation

<sandro> looking again at http://en.wikipedia.org/wiki/Map_projection#Equal-area

could make the colours less stark by reducing contrast

HadleyBeeman: You could use OSM with pins rather than colouring the whole country

<sandro> size of the pin

<sandro> yes

<Mike_Pendleton> yes

<sandro> size of the a circle.

<bhyland> Here is the map I'm projecting: http://usepa.3roundstones.net/rdf/2012/usepa/nuclear-demo/nuclear-map.xhtml?view

<Mike_Pendleton> +1

HadleyBeeman: What actually is the map for?

<DaveReynolds> Decision to make is whether you want to use this for navigation or to summarize where things are hot.

bhyland: To show potential decision makers what's available. Location often matters

<DaveReynolds> For navigation use pins. For density of activity use a heat map.

<gatemezi> It seems like http://dir.w3.org/directory/schema# is 404...

HadleyBeeman: Government departments like working with people who have experience of their area

BartvanLeeuwen: You could include social media presence (as org:Organization equivalent class foaf:Organization)

<gatemezi> http://leafletjs.com/examples/quick-start.html

gatemezi: You can include locations as well as addresses

<HadleyBeeman> http://xmlns.com/foaf/spec/#sec-glance for the Social Web

gatemezi: You can show the location on a map on the page
... Also the dir schema itslef is a 404

<MakxDekkers> all, just ran out of credit and got disconnected. planning to start fresh tomorrow at 10 your time. have a nice dinner

<HadleyBeeman> Bye, Makxdekkers!

thanks MakxDekkers - bye

HadleyBeeman: It's 5 o'clock

<fadmaa> I also suggest reconsidering the 3D pie charts... they've got awful reputation :)

The future of the Community Directory

<HadleyBeeman> http://www.w3.org/2011/gld/wiki/Community_Directory_Timetable

<PhilA> HadleyBeeman: We have the idea that the directory needs a new supporting group

<PhilA> bhyland: We're on the way to setting up a Community Group to support it

<PhilA> PhilA: Questions the long term sustainability of a CG dedicated to this

<PhilA> sandro: Is it only LD, is it only government? Is it anything vaguely touching on W3C?

<PhilA> sandro: If we can slide it that way then it's plenty sexy

<PhilA> sandro: It's a directory of people associated with W3C technology

<sandro> sandro: How about: Building a Decentralized Directory of People/Organizations/Projects Related to W3C Technologies (Starting With Government Linked Data)

<bhyland> +1 sandro

<sandro> +1 at some point flag W3C Members as such.

<PhilA> BartvanLeeuwen: Should the directoty highlight members?

<PhilA> +1

<BartvanLeeuwen> +1

<HadleyBeeman> +1

<PhilA> agreement on that

<sandro> https://www.w3.org/Member/ACList and the RDF is somewhere....

<DaveReynolds> 0

<PhilA> http://www.w3.org/Consortium/sup

<PhilA> bhyland: Aim was to highlight the ecosystem of LD

<PhilA> ... want 100s of projects etc.

<PhilA> HadleyBeeman: Are we putting the product ahead of the use case?

<PhilA> BartvanLeeuwen: It's in the charter, there is a demand for it

<PhilA> bhyland: It's about answering people in gov questioning whether this LD thing is actually supported.

<HadleyBeeman> for the government linked data space, yes. But expanding beyond that (to the entire W3C community, however we define that, or everyone who feels connected to the W3C) is very different situation. Merits some signficant scoping and requirements assessment.

<HadleyBeeman> (that was building on "Are we putting the product ahead of the use case?")

<PhilA> DeirdreLee_: It's useful for people finding out what other parts of the same organisation are doing, never mind what others are up to

<PhilA> HadleyBeeman: So what are we going to do with the community directory ahead of the charter end

<PhilA> bhyland: we can set up a CG but if there's a home it could be in that already exists then we can do that

<PhilA> PhilA: The putative Open Data WG could be a home for it, or yes, a CG

<PhilA> sandro: I think it would be a poor fit for ODWG. It's a specific thing and needs its own group. Not all open data people will want to support the dir

<PhilA> bhyland: We're not going to wait. We're going to set up a CG with 5+ members

<bhyland> Propose: Set up a community group with 5 members and plan to move the ongoing support & maintenance to the CG.

<HadleyBeeman> +1

<bhyland> +1

<PhilA> +1

<BartvanLeeuwen> +

<BartvanLeeuwen> +1

<PhilA> RESOLVED Set up a community group with 5 members and plan to move the ongoing support & maintenance to the CG.

<fadmaa> +1

<DeirdreLee_> +1

<bhyland> ACTION: Bernadette to get the W3 Dir-o-matic Community Group [recorded in http://www.w3.org/2013/04/11-gld-minutes.html#action03]

<trackbot> Created ACTION-114 - Get the W3 dir.w3.org Community Group [on Bernadette Hyland - due 2013-04-18].

<PhilA> sandro: Suggest calling it the dir.w3.org CG

<sandro> bhyland, that s/// command wont actually change the action, just how the action appears in these minutes. Messy.

Tomorrow's agenda

<bhyland> We are closing Day #1 with many things done, not the least of which is a working definition of a "dataset", see https://dvcs.w3.org/hg/gld/raw-file/default/glossary/index.html#dataset

<DaveReynolds> Don't think ORG will take an hour.

<PhilA> HadleyBeeman: Good news that ORG won't take an hour. How long will it take?

<HadleyBeeman> ^davereynolds?

<DaveReynolds> Hmm. Mic problem. Was trying to say it is probably 30min but might be surprised by what discussion is sparked.

<DaveReynolds> Fine

<PhilA> bhyland: We should have a tracker review and cleanup

- DRAFT -

GLD Face to Face meeting Day 1

11 Apr 2013

Attendees

Contents

Quick review of European data Forum

Agenda bashing

Agenda bashing for today & tomorrow

Best Practices

Data Cube vocabulary LC feedback review

Data Cube

DCAT Last Call Comments

Community Directory

The future of the Community Directory

Tomorrow's agenda

Summary of Action Items

Scribe.perl diagnostic output