– DRAFT –
DXWG Plenary

Admin

proposed: accept minutes of last week - https://‌www.w3.org/‌2019/‌06/‌18-dxwg-minutes

+1

<kcoyle> +1

<alejandra> +1

TomB: +1

Resolved: accept minutes of last week - https://‌www.w3.org/‌2019/‌06/‌18-dxwg-minutes

DCAT

<kcoyle> PWinstanley: wide review for 3+ weeks; some input coming in

PWinstanley: sent around a reminder today; danbri is passing around post in Google
… also trying to get some responses from European Commission
… need to study their upgrade path

<alejandra> assessment meeting - sounds good

PWinstanley: was talking to Philippe and he suggested an assessment meeting before going to candidate recommendation
… also, evidence for implementation only needed for end of CR
… which take 4 weeks minimum
… suggests meeting 3 July 20 UtC
… topic will be checking that all is organized and relation to feedback
… suggestions from i18n seem easy to do

alejandra: seem to be things that we can address

PWinstanley: please prepare for the 3rd

Profiles

PWinstanley: meaning of profiles in the different documents

<PWinstanley> kcoyle: To remind people that we got a response on the link between Conneg and IETF, and it seems that they can be different, but not contradict

PWinstanley: links from public email on profiles are in agenda
… reactions to this?
… kcoyle's comment on it being a mouthful. It is a good idea to give people more specifics

tomB: never liked this definition; parts I don't like is : to accomplish a particular function
… but also do not like a named set of constraints
… a book could be a named set of pages
… guidance document enumerates many types of profiles
… not just constraints, sometimes extensions
… understand that Rob argues that an extension is a constraint

PWinstanley: Antoine talks about data profiles

tomB: what I don't want to do is to limit profiles to the DCMI sense

<PWinstanley> kcoyle: I am now beginning to see profiles as specifications, if they explain how something is done

<PWinstanley> +1 to kcoyle

<PWinstanley> kcoyle: so maybe it is a specification with certain qualities

PWinstanley: yes, agree, you can have a schema that describes what the book is
… there can be a set of rules

tomB: rfc 6906 has a definition "a profile ... does not alter..." (will need to complete)

<PWinstanley> TomB: https://‌tools.ietf.org/‌html/‌rfc6906

<PWinstanley> This specification defines the 'profile' link relation type that allows resource representations to indicate that they are following one or more profiles. A profile is defined not to alter the semantics of the resource representation itself, but to allow clients to learn about additional semantics (constraints, conventions, extensions) that are associated with the resource representation, in addition to those defined by the media [CUT]

tomB: extensions are not understood as constraints
… if anything is a constraint then nothing is a constraint
… rfc definition is not bad; main intention is to provide additional info

PWinstanley: please respond to the emails from Antoine and kcoyle

tomb: saying "a set of constraints" is more limited than what profiles are

AndreaPerego: trying to make up my mind; not ready to be constructive

PWinstanley: anything you are uneasy about?

AndreaPerego: for me the issue was the inconsistent way we were using profile when talking about conneg and profile guidance
… if we are able to say that there is one definition for conneg and another for prof guidance
… if so we would need to add a qualifier to make the distinction clear

riccardoAlbertoni: Nothing I am uneasy about
… my impression is that there are different types of profiles so
… good to use a qualifier when talking about profile
… but do not favor a particular definition

PWinstanley: each dealing with particular subsets of a broader concept of profle
… would be helpful to move from the total universality of constraints to
… show a family of profile types
… then there would be things that pass tests, like data profiles
… then there are profiles of patterns, like DCAT; and other things that are more loosely structured

tomb: arguing against overly broad use of constraints
… "constraints, conventions, extensions" is a start of a list of things that characterize a profile
… does not alter semantics is also a good "rule"
… constraints are just one possible function

PWinstanley: the other thing about extensions is that they bring in something that was defined elsewhere

PWinstanley: in a guidance document we would want to say something about those cases

<PWinstanley> kcoyle: one think I noticed in Antoine's note that concerned me was that DCAT says something like that DCAT-AP must conform to DCAT

<PWinstanley> ... there is something in the conformance area of the DCAT doc that raises concerns

<alejandra> https://‌w3c.github.io/‌dxwg/‌dcat/#profiles

<PWinstanley> alejandra: I think what we say is that in the conformance section we are not discussing DCAT-AP,

<alejandra> " Each of these new specifications constitutes a DCAT profile, i.e. a named set of constraints based on DCAT (see § 4. Conformance)"

2. "conformance of data catalogue descriptions wrt. DCAT profiles" (DCAT-AP-whatever).

AndreaPerego: about conformance - depends on context; is different if open or closed world assumption
… dcat-ap is conformant because dcat does not say you cannot add
… if you can read dcat you can also read dcat-ap ignoring the extensions

<riccardoAlbertoni> +1 to andrea ... extension does not imply "non conformance .."

<PWinstanley> kcoyle: I don't disagree, but think that it needs to be in the DCAT doc when discussing conformance -

<PWinstanley> ... what you said makes sense, but just saying 'conformance' needs more specification

<PWinstanley> AndreaPerego: there is a bullet point saying that when metadata is conformant with DCAT then you must use the properties and classes of DCAT, but if you want to say something else then you can add more

<PWinstanley> ... but perhaps we need to look at it again

<PWinstanley> kcoyle: the question is how you would test conformance. The doc doesn't discuss open world assumption etc

PWinstanley: where are we? we have some work to do on DCAT and other docs that extend from discussion
… still need a paragraph or 2 for general public when we discuss profiles
… need to think about naive user as the audience

tomB: +1

PWinstanley: (general pep talk about DCAT)

PWinstanley: let it go for another week or so?

<riccardoAlbertoni> +1

tomB: good idea; no problem defining conformance in DCAT just in defining for profiles generally
… do DCAT and conneg, then guidance; worry about how much group can do

PWinstanley: this will become clearer after the assessment meeting
… we are also looking toward another PWD of conneg

<PWinstanley> kcoyle: I asked on the list where it might be a good idea to discuss, github or the email?

<PWinstanley> tomB: there is a google doc, but that is only good for a specification

<PWinstanley> kcoyle: this is a sentence or 2, so github should work

<PWinstanley> tomB: agreed

alejandra: github is a good place to do this
… reiterate point of meeting on July 3

PWinstanley: July 3 meeting is about DCAT move to CR, so need a meeting at same time as subgroup
… to work out where we are with DCAT
… while in CR can make only stylistic changes
… also, evidence of implementation is due at END of CR period (about 4 weeks)
… in addition to wide review need to ensure we've had a comprehensive plenary review
… so please join in that meeting

AndreaPerego: cannot be there on the 3rd - at a workshop about discoverability on the web

<AndreaPerego> Workshop on making spatial data discoverable on the Web: https://‌webgate.ec.europa.eu/‌fpfis/‌wikis/‌x/‌Y9kuF

AndreaPerego: can there be a prep meeting before it?

<riccardoAlbertoni> +1 to andrea

riccardoAlbertoni: can join next week but need a prep meeting, esp. because there are github
… discussions that haven't concluded
… can we arrange an agenda for next week or week after to see when we can do it
… it would be good to have a DCAT meeting before discussing with Philippe

<alejandra> +1 to preparatory meeting

PWinstanley: ok, will ask philippe about week 3 of July (e.g. 15th)

<riccardoAlbertoni> thank you all, bye