DXWG Weekly Telco

Preliminaries

First agenda item: approve meetings from last time

<Thomas> No corrections

No additions / corrections

Resolved: Last week's minutes approved

Kcoyle: we will spend some time on discussing how we are going to organise ourselves

DCAT Editors

kcoyle: discussing role of editors and contributors

<PWinstanley> I am up for editor

I'd like to be an editor

<phila> Data on the Web Best Practices had 3 editors and *loads* of contributors

<RubenVerborgh> Editor of what?

DCAT editor

<phila> DCAT, RubenVerborgh

<phila> phila has changed the topic to: DXWG Weekly telco

<Thomas> I'd be willing to paricipate also

<RubenVerborgh> *sorry, missed "topic", sounded more general*

<Thomas> As an editor

<SimonCox> I can help edit - have done one before

<Makx_Dekkers_> Can help with DCAT too.

<SimonCox> 4 editors is an OK number

potential editors: alejandra, thomas, Simon, Peter

Resolved: DCAT editors will be Peter, Thomas, Simon, Alejandra

<erics> +1

<Thomas> +1

<Makx_Dekkers_> +1

<Jaroslav_Pullmann> +1

<annette_g> +1

<present_Ixchel> +1

<MJ_Han> +1

+1

<nandana> +1

<Jaroslav_Pullmann> https://‌docs.google.com/‌document/‌d/‌1z8UVjMEPoqp69ZHXk6asY6tCHHECglk4-3Lznd5dxS0/

kcoyle: Next discussion - the proposal of use case template

kcoyle: another proposal on the reorganization of what the use cases are about and what the tasks are

kcoyle: many use cases and requirements that people are considering DCAT requirements but may be better as requirements for Application Profiles

Jaroslav_Pullmann: discussing the proposal for the use cases structures

Jaroslav_Pullmann: status of use cases: new, open, stable, closed

Jaroslav_Pullmann: describe the problem statement - current situation and missing aspect that should be considered and what is the motivation for an improvement

Jaroslav_Pullmann: optionally, consider existing approaches

Jaroslav_Pullmann: next mandatory part is 'requirements'

kcoyle: continues discussing the template structure - links to related use cases

RubenVerborgh: organization might require some tweaking

RubenVerborgh: separation might be artificial in some cases

RubenVerborgh: as some use cases are cross-cutting

<RubenVerborgh> in particular: overlap between DCAT distribution and profiles / conneg

<Jaroslav_Pullmann> seems I am muted, give me a second

kcoyle: Now discussing re-organization

<Makx_Dekkers_> My only pledge was to keep URLs to use cases persistent.

Jaroslav_Pullmann: did we reach an agreement on extending the template?

kcoyle: yes

Jaroslav_Pullmann: it will often be the case that we will need to close duplicate use cases and continue one of the threads

<RubenVerborgh> *ah woops, sorry for jumping ahead, thought that was settled*

<annette_g> +1 to having less repetition

kcoyle: moving to reorganization section

Jaroslav_Pullmann: Main category are deliverables - 3 deliverables at first

Jaroslav_Pullmann: DCAT 1.1. core

Jaroslav_Pullmann: catalog, datasets, distributions (representations of either static or dynamic datasets) - I agree that some elements are cross-cutting

Jaroslav_Pullmann: conneg and distribution may overlap

Jaroslav_Pullmann: so, we need to discuss the focus of the use case

<kcoyle> alejandra: if we have aspects from the use case, for a different section, should we add a new one?

kcoyle: Editors will be able to decide this after they have a look at the document

erics: I had a similar question - in DWBP there were cross-cutting use cases, so perhaps consider a section with these cross-cutting use cases across the major focus areas

SimonCox: I'd like to pull us back to the Use cases - the status flags are unclear to me

SimonCox: the use cases should yield requirements

SimonCox: does the status mean that the requirement has been satisfied?

roba: we need to decide if we put up the use cases and then the WG decides if they are in scope

roba: we also need to deal with the duplication

roba: stable means that those are the ones that we pay attention to

roba: I suggest we put cross-cutting use cases first

<SimonCox> The easiest way to detect duplication is through the requirments

roba: and then we deal with the others later

SimonCox: to identify the duplications we need to look at the requirements level, which is what it needs to be carried through

<Jaroslav_Pullmann> +Jaroslav_Pullmann

<Jaroslav_Pullmann> sorry I was offline for a moment

roba: we need to work out how much the group expects the editor to identify the duplication or we go through each use case one by one and the group decides

<Makx_Dekkers_> I'd also prefer to gather use cases without considering duplication, and deal with de-duplication later. The requirements are what matter.

Thomas: I was wondering if we are not focusing too much now on the use cases

Thomas: but we need to make sure in the future that the requirements are described through the use cases

<Makx_Dekkers_> I am seeing use cases that are similar to what I want to contribute, but I still would want to describe my use case. Otherwise, would I need to negotiate with the author of the similar use case to merge?

annette_g: we've got the use cases divided into 3 deliverables, but it seems to me that Profile/Conneg use cases might be a single group

<RubenVerborgh> +1 to tags (to an extent, they are already present now)

<RubenVerborgh> this would address my concerns on cross-cutting

<RubenVerborgh> we cannot pretend that conneg is entirely separate from the rest

annette_g: but maybe grouping is not the way to do it - I tend to maximise the use cases that are considered for a single deliverable

<phila> +1 to grouping/tagging requirements cf. grouping UCs

<Zakim> RiccardoAlbertoni, you wanted to say not sure that duplication should be avoided at this stage

RiccardoAlbertoni: I am not sure why we are focusing so much at duplication at this stage, as it can be very good now and it can be solved later on

RiccardoAlbertoni: later on

RiccardoAlbertoni: if we have more use cases, we have better ground for the requirements

<Makx_Dekkers_> +1 to riccardo

<PWinstanley> +1 to RiccardoAlbertoni comment about multiple forms of use cases

+1

<Makx_Dekkers_> we can do de-duplication on the requirements

Jaroslav_Pullmann: we are not worried about duplication at the moment

phila: tx

Jaroslav_Pullmann: we want to have focus deliverables

Jaroslav_Pullmann: cross-referencing is done via text

Thomas: we shouldn't shift the problem to text managing

<Zakim> phila, you wanted to talk about assumptions not in UCs

phila: UCs documents usually forget to say 'we are assuming X, etc' - we don't need to make explicit all assumptions

<SimonCox> I agree - need some pragmatism. But remember that the main purpose of UC is to generate requirements. It is the requirements which must be satisfied by the technology

<Thomas> +1 for Phil

phila: it allows to keep the number of UCs manageable

<Jaroslav_Pullmann> example on using tags: https://‌www.europeandataportal.eu/‌de/‌content/‌show-license

<Thomas> +1 to Simon

<Makx_Dekkers_> +1 to Simon

Jaroslav_Pullmann: the use cases will always have multiple dimensions
… refers to the example of using tags
… allowing a selective view of particular use cases

https://‌www.europeandataportal.eu/‌de/‌content/‌show-license

... we can switch on and switch off relevant use cases

<annette_g> nope

kcoyle: has anyone contributed their use cases?

<chile> no

<erics> No

<SimonCox> I have another!

<chile> will do

kcoyle: can we get all the use cases contributed as soon as possible?

I have more to add too

<annette_g> didn't we set a one month deadline last week?

<Thomas> I have some more to add

<Makx_Dekkers_> I am sitting on five to ten of them

Jaroslav_Pullmann: do we have a resolution on the reorganization proposal?

<Makx_Dekkers_> Need to find time to write them up

PROPOSAL: the group will consider this reorganization of use cases

<Thomas> +1

+1

<annette_g> +?

<roba> +1

<MJ_Han> +1

<phila> This reorganization

<Jaroslav_Pullmann> +1

<Colleen> +1

<RiccardoAlbertoni> +1

<present_Ixchel> +1

<erics> +1

<mbruemmer> +1

<Makx_Dekkers_> waht does 'consider' mean?

<Makx_Dekkers_> 'we'' think about it and decide later'?

PROPOSAL: the group will accept the reorganization

<Thomas> +1

<Colleen> +1

<Jaroslav_Pullmann> +1

<annette_g> -!

<annette_g> -1

<roba> i guess we can look at the results and then change our minds via a new proposal :-)

<roba> +1

<present_Ixchel> +1

annette_g: we discussed about doing it a little different

annette_g: one suggestion was combining those UCs that are to do with APs

annette_g: another option is to use the tags for grouping

annette_g: the last grouping from the document is looking rather small

annette_g: it seems to me that they cross-cut those too

annette_g: I think that tags would be better

Jaroslav_Pullmann: both approaches are proposed to be applied in parallel

Jaroslav_Pullmann: we have both: tags and the grouping

<SimonCox> {Jaroslav_Pullmann breaking up too much?}

alejandra: why don't consider the 3 deliverables as other tags?

<roba> one reason for reordering is to deduplicate easier - we shouldnt worry too much about final order perhaps.

Jaroslav_Pullmann: we want to harmonize UCs

<Thomas> @alejandra : isn't that the same in the end?

<phila> alejandra: The separation in the 3 deliverables is in the tags

<kcoyle> alejandra: one use case could be tagged with multiple tags

<phila> ... We can tag each UC/Req with the relevant deliverable(s)

<Makx_Dekkers_> @roba: I don't think we should try to deduplicate use cases; we can deduplicate requirements

Jaroslav_Pullmann: we were using the google doc for discussion, then we move this reorganization will be more apparent

Jaroslav_Pullmann: reorganizing this in the wiki page

kcoyle: I understood that we should use tags rather than reorganization

yes

<PWinstanley> +1 to tagging

annette_g: yes

<Makx_Dekkers_> +1 to tagging

<RiccardoAlbertoni> +1 to tagging

<PWinstanley> we should have a dynamic document

phila: there are bits where the javascript will take the requirements and generate a list of the use cases that yielded them

<PWinstanley> +1 to phila

phila: there are tables that are autogenerated

<Thomas> +1 to phila

<annette_g> +1 to phila

phila: it is a web document

<PWinstanley> tagging can help with curation

roba: I think we may be getting ahead of ourselves, we have a list of candidate use cases with duplication in them - group them to facilitate discussion is the first step

roba: if we had a technology where we could tag them and get dynamic grouping, then we could do it

<MJ_Han> like this? https://‌docs.google.com/‌spreadsheets/‌d/‌1EmojqR3nQo3ioj-qG_C0UeGJ0prcNZFom9th0AzaN0c/‌edit?usp=sharing

<phila> +1 to Rob

roba: but to organize the document to coherently look at them to reduce the number of use cases, it is the first step

+1 to Rob

<Thomas> +1 rob - keep it manageable and proceed for now

<MJ_Han> +1 to Rob

kcoyle: we should get all the use cases and continue the work, maybe we leave it to the UC editors to organize them

<SimonCox> I'm still concerned about focus on UCs as the end point. The UCs are there to expose requirements.

kcoyle: yes, requirements are at the core of what will define our work in the future

<Makx_Dekkers_> +1 to Simon; requirements is what we're looking for

<erics> I thought we voted last time that we were given a month?

kcoyle: can we stop here with the discussion then?

Jaroslav_Pullmann: I'm convinced that we will end up with some grouping

Jaroslav_Pullmann: so we need some simple structure, we can go on with our work and we'll see

<Makx_Dekkers_> Deadline for use cases is still end of June, correct?

<kcoyle> https://‌www.w3.org/‌2017/‌dxwg/‌wiki/‌Use_Case_Working_Space#Dataset_Versioning_Information

<Makx_Dekkers_> Can I reiterate request for persistent URLs for use cases?

kcoyle: deadline for providing the use cases is end of June

Before we finish, reminder for everyone to indicate their intention to attend the F2F meeting please

kcoyle: discussing use case on dataset versioning

<annette_g> +1 to the use case

phila: definitely support the use case, this brings me to mention that if the UC Editors find it helpful, they can look at the report of the workshop we had last year

<erics> +1 to the use case that covers dataset and distributions

<phila> SDSVoc Workshop report

Jaroslav_Pullmann: I support the versioning, but there is much discussion about what a dataset is

Jaroslav_Pullmann: discussion on the notion of dataset, as it clarifies what is going to be versioned

roba: I support the use case, I have some issues on how the requirement is expressed

roba: I personally think that the requirement is not an extension to DCAT

<phila> +1 to Rob - Reqs shouldn't include solutions

<Makx_Dekkers_> Versioning is much more complicated than in this use case. I have one on my list that identifies at least four types of versioning.

roba: it sounds to me as putting the solution forward too early on

kcoyle: can we accept the use case with the caveat that the requirements need to be reworded?

<erics> +1 to accepting a use case topic but not necessarily the content

<Jaroslav_Pullmann> +1

Jaroslav_Pullmann: we could put it in the Open state with a comment on it

<kcoyle> PROPOSED: accept UC ID4, with possible rewording of requirements

annette_g: just looking at the description, I didn't get the sense that we needed to do it in a specific way, but pointing out that there is some ways of dealing with it somewhere else

<phila> PROPOSED: Accept the versioning Use Case, modulo rewording the requirements not to include the solution https://‌www.w3.org/‌2017/‌dxwg/‌wiki/‌Use_Case_Working_Space#Dataset_Versioning_Information

roba: the requirement says 'an extension to DCAT'

+1

<erics> +1

<annette_g> +1

<roba> +1

<Thomas> +1

<mbruemmer> +1

<PWinstanley> +1

<Jaroslav_Pullmann> +1

<MJ_Han> +1

<RiccardoAlbertoni> +1

<present_Ixchel> +1

<SimonCox> +1

<Colleen> +1

Resolved: Accept the versioning Use Case, modulo rewording the requirements not to include the solution https://‌www.w3.org/‌2017/‌dxwg/‌wiki/‌Use_Case_Working_Space#Dataset_Versioning_Information

<SimonCox> (though still not fully clear what 'Accept' means in relation to a UC)

<Makx_Dekkers_> apologies, I do not agree with the versioning use case

<phila> Skeleton DCAT doc alejandra SimonCox Thomas PWinstanley

<Thomas> Thx phil

<RiccardoAlbertoni> thanks and bye

<erics> take care everyone

<PWinstanley> bye

<annette_g> bye!

<Makx_Dekkers_> bye

<Thomas> Bye