Re: DXWG Plenary: 2019-06-18T20:00:00

Hi,

Sorry what I'm trying to write has taken me too much time and I'll have to skip tonight's call to focus on other work I have to do for tomorrow :-(
In the meantime I'm going to try to react to Karen's point.
(anyway it's probably better to have my written reaction rather than adding myself to a queue that's probably going to be long!)


>> As for any thoughts of a "canonical definition" - we agreed on a
>> definition over a year ago [1] and much has happened since then. In
>> particular, we have decided to de-emphasize all deliverables except DCAT
>> and Conneg for the time being. 


Yes but I believe these many changes have essentially changed the approach to profile in these deliverables (some elements of which I've tried to point to in my mail to Tom)


>> Even at the time we appear to have seen
>> it as a working definition that could be revisited.


I agree this shouldn't be set in stone. But we may still want to not re-do all discussions.


> - Is a definition needed needed for DCAT beyond what is already there?


Yes. Because DCAT uses profiles in a way that's not similar to CONNEG (DCAT profiles are based on something while Conneg profiles can be self-standing), and it uses conformance (by means of dc:conformsTo [1]) for datasets themselves in a way the generalized the conformances that it introduces for data catalogue descriptions.


> - Are people happy with this definition for conneg

No because as said above Conneg can employ specifications that are not based on other specifications, and thus do not directly fall in scope of DCMI's application profiles, which is what DCAT currently refers (and then it may actually branch away from it. I'm not sure that all the parliance about constraints in DCAT profile section [2] is 100% compatible with DCMI's notion application profiles. It's actually quite close to our DXWG specification [3] in this part (where DCAT uses the expression "named set of constraints")


> - If not, do we think there will be time to revisit the definition by
> the time conneg reaches CR?


yes


> - Is there likely to be a Profiles Guidance document by the end of the 6
> month extension? (Note: I believe it was this document that was the
> mechanism for a unified definition of profile.)


if we can agree on a basic unification of the notion of profiles we came with [3], then yes. We already have a lot, especially on the functions of profiles, and that is quite useful for setting the scene to a unification.

Best

Antoine


[1] https://w3c.github.io/dxwg/dcat/#Property:resource_conforms_to
[2] https://w3c.github.io/dxwg/dcat/#profiles
[3] https://www.w3.org/2017/dxwg/wiki/ProfileContext#Definitions_of_Profile.2C_Application_Profile.2C_Metadata_Application_Profile


On 18/06/2019 20:14, Karen Coyle wrote:
> Rob, et al.
> 
> I'm actively pinging everyone involved in the IETF work to see what the
> thoughts are on how closely it needs to coordinate with DXWG. This may
> be a question that does not currently have a clear answer.
> 
> As for any thoughts of a "canonical definition" - we agreed on a
> definition over a year ago [1] and much has happened since then. In
> particular, we have decided to de-emphasize all deliverables except DCAT
> and Conneg for the time being. Even at the time we appear to have seen
> it as a working definition that could be revisited. Where we seem to be is:
> 
> - Is a definition needed needed for DCAT beyond what is already there?
> - Are people happy with this definition for conneg?
> - If not, do we think there will be time to revisit the definition by
> the time conneg reaches CR?
> - Is there likely to be a Profiles Guidance document by the end of the 6
> month extension? (Note: I believe it was this document that was the
> mechanism for a unified definition of profile.)
> 
> It seems to me that all of these questions should inform our discussion.
> In particular, I would rather not see the group commit to a process that
> cannot be resolved within the charter extension.
> 
> kc
> [1] https://www.w3.org/2018/02/06-dxwg-minutes#ResolutionSummary
> 
> On 6/17/19 4:48 PM, Rob Atkinson wrote:
>> This is somewhat misguided a discussion as formulated, as the IETF
>> document has no official status and the intent has always been to update
>> it to reflect the IETF relevant parts of the broader conneg
>> approach. The canonical definition of profile, within current processes
>> and deliverables,  is the one in the conneg document, as that is
>> focussed on the functional requirements for behaviour based on profiles.
>> (The Profiles ontology will need to be updated if necessary to reflect
>> any substantive changes agreed in the conneg document. The DCAT (and any
>> other) use of profiles will by necessity be a narrower usage, just
>> because they start with a more specific constraint on what is being
>> profiled. So as long as DCAT, Dublin core, MIME type profiles etc are
>> understood within context, and functionally covered by the general
>> definition, such usages are consistent with the canonical definition and
>> conneg by profile mechanisms can be used in those contexts safely.
>>
>> If there is some other underlying goal or reason to have a different
>> approach to setting the context for defining profiles it should be
>> stated, although its really too late in the day to be introducing new
>> requirements, and never useful to assert new requirements without a
>> grounding Use Case. Otherwise, we have a working definition, and useful
>> community feedback, and we are examining its accuracy and interpretation
>> w.r.t. to functional requirements. The actual issue is whether we can
>> identify any improvements in that definition. Starting up another
>> process or discussion with a different scope around this is not useful
>> of feasible at this stage IMHO.
>>
>> Rob
>>
>>
>> On Tue, 18 Jun 2019 at 01:08, pedro winstley
>> <pedro.win.stan@googlemail.com <mailto:pedro.win.stan@googlemail.com>>
>> wrote:
>>
>>      Yes, let's do that Karen. I will update
>>
>>      On Mon, 17 Jun 2019, 15:58 Karen Coyle, <kcoyle@kcoyle.net
>>      <mailto:kcoyle@kcoyle.net>> wrote:
>>
>>          Could we add a discussion of Tom's email to the agenda? [1]
>>          Although he
>>          focuses on conformance, I think the email rounds up some
>>          important info
>>          about profile definitions.
>>
>>          First, DCAT has a definition (in the green note [3]) that
>>          presumably is
>>          sufficient for DCAT purposes.
>>
>>          Second, the IETF proposal has a definition that again is presumably
>>          sufficient for that proposal. [2] It would seem inappropriate
>>          for the
>>          conneg definition to be significantly different from the
>>          definition in
>>          the IETF proposal. In fact, it would probably be necessary for
>>          them to
>>          be the same or as close to the same as possible. (We should ping
>>          Lars on
>>          this.)
>>
>>          Because of this, I see no reason to work on a definition of profile
>>          UNLESS the intention is to continue work on the profile guidance
>>          document and to develop a definition that is focused on the
>>          creation of
>>          profiles. That definition could be more specific as it would be
>>          attempting to drive the creation of a specific concept of profile. I
>>          don't think that definition would be a substitute for the
>>          IETF/conneg
>>          definition, and for DCAT that would be a question to be posed for a
>>          future version.
>>
>>          kc
>>
>>          [1]
>>          https://lists.w3.org/Archives/Public/public-dxwg-wg/2019Jun/0035.html
>>          [2]
>>          https://profilenegotiation.github.io/I-D-Accept--Schema/I-D-accept-schema
>>          [3] https://w3c.github.io/dxwg/dcat/#conformance
>>
>>          On 6/17/19 3:31 AM, pedro winstley wrote:
>>          > Dear Colleagues
>>          >
>>          > The next plenary meeting of DXWG will be at
>>          2019-06-18T20:00:00 and
>>          > the agenda draft is at
>>          > https://www.w3.org/2017/dxwg/wiki/Meetings:Telecon2019.06.18
>>          >
>>          > The key discussion point will be to make some concrete plans and
>>          > schedule for the definition document describing 'profile' to be
>>          > sufficient for the requirements of the conneg and DCAT work, and
>>          > discussion of how well this covers the needs of the prof
>>          vocabulary.
>>          >
>>          > Cheers
>>          >
>>          > Peter
>>          >
>>          >
>>
>>          --
>>          Karen Coyle
>>          kcoyle@kcoyle.net <mailto:kcoyle@kcoyle.net> http://kcoyle.net
>>          skype: kcoylenet
>>
> 

Received on Tuesday, 18 June 2019 19:57:40 UTC