OSLC CORE TC Minutes February 4, 2016

I re-submitted the meeting invite - please check to make sure you have calendar entries for Feb 18, Mar 3 et seq.

All TC members: At the next meeting we hope to be done with internal reviews. Now is your chance to review the full lineup of documents and raise any remaining issues. Typographical errors are fair game too. Please try to be done by Monday Feb 15.

Note one possible correction to the minutes - see the action items.

Scribe

Martin Sarabura (PTC)

Attendees

David Honey (IBM)
Harish Krishnaswamy (Sofware AG)
Ian Green (IBM)
Jim Amsden (IBM)
Martin Sarabura (PTC)
Jean-Luc Johnson (Airbus)
Nick Crossley (IBM)
Jad El-khoury (KTH)

Regrets

Resolutions

· Drop section 9 of resource shapes document

Actions

David to raise issue in core regarding vocabulary versioning

· Issue 43 - Nick to propose resolution

· Issue 39 - Nick to propose new names, and will check what they were intended to mean is the what he thinks it should mean

Ian to complete review of discovery document by Feb 11
Jim to propose appropriate non-normative wording regarding constraints in resource shapes to respect the fact that the normative requirement is SHOULD not MUST
Jim to propose wording to address pre-fill of delegated UIs
Jim to remove clause 6.4.6 (should that be 6.4.7?) from delegated dialog document

Chat transcript from room: oslc

[07:07] Martin Sarabura (PTC): Martin to scribe

[07:07] Martin Sarabura (PTC): Roll call:

[07:07] Martin Sarabura (PTC): David Honey (IBM) Harish (Software AG) Jad El-khoury (KTH) Jean-Luc Johnson (Airbus) Jim Amsden (IBM) Martin Sarabura (PTC) Nick Crossley (IBM)

[07:09] Martin Sarabura (PTC): https://issues.oasis-open.org/browse/OSLCCORE-45

[07:09] Martin Sarabura (PTC): Seems to be agreement on this one - further action not required - Nick to check that concolusions reflected in the specs

[07:09] Martin Sarabura (PTC): Jim has added to the specs

[07:09] Martin Sarabura (PTC): Should be closed

[07:10] Martin Sarabura (PTC): https://issues.oasis-open.org/browse/OSLCCORE-51

[07:10] Martin Sarabura (PTC): Nick: Also fixed

[07:11] Martin Sarabura (PTC): David: Will work on document using OSLC-CCM for vocabulary.

[07:11] Martin Sarabura (PTC): Will review with Nick. Still on his list

[07:12] Martin Sarabura (PTC): Nick: Vocabularies for OSLC not the same as domain vocabularies - can only add to it

[07:13] Martin Sarabura (PTC): David: Re resource shapes - no standard way to indicate evolution for query builder

[07:13] Martin Sarabura (PTC): Nick: Evolution of shapes is supported in theory - no recognized versioning system

[07:14] Martin Sarabura (PTC): URI structure can be used to associate versions of shapes

[07:16] Martin Sarabura (PTC): David: Shapes themselves are constructed dynamically from resources that have versions

[07:16] Martin Sarabura (PTC): No need to dig deeper at this time - will continue offline

[07:17] Martin Sarabura (PTC): https://wiki.oasis-open.org/oslc-core/CompletionTimeline

[07:17] David Honey (IBM): https://wiki.oasis-open.org/oslc-core/FrontPage#Working_Documents

[07:21] Martin Sarabura (PTC): Document statuses: Delegated dialogs still need some discussion

[07:21] Nick Crossley (IBM): Nick has completed a review of Attachments

[07:21] Martin Sarabura (PTC): Link guidance: No open actions

[07:22] Martin Sarabura (PTC): Delegated dialogs: Martin P still going to review?

[07:23] Martin Sarabura (PTC): Resource shapes: Looking for historical view from Nick and Martin P

[07:23] Martin Sarabura (PTC): Nick: Plan to do further review of vocab and resource shapes

[07:24] Martin Sarabura (PTC): Jim has run shape checker recently but could be run again

[07:24] Martin Sarabura (PTC): https://wiki.oasis-open.org/oslc-core/Meetings/Telecon2016.01.21

[07:25] Martin Sarabura (PTC): Minutes accepted from previous meeting

[07:25] Martin Sarabura (PTC): Jira issues

[07:25] Jim Amsden (IBM): [ ] OSLCCORE-43 OSLC Core 2.0 Appendix A: Common Properties defines things not currently in OSLC3

[07:25] Martin Sarabura (PTC): https://issues.oasis-open.org/browse/OSLCCORE-43

[07:26] Martin Sarabura (PTC): Jim: Nick to propose as part of review of shapes/vocab, still open

[07:26] Martin Sarabura (PTC): Nick: Wanted to make change to respec, lower priority than other fixes: Representation and bibliography

[07:26] Martin Sarabura (PTC): https://issues.oasis-open.org/browse/OSLCCORE-54

[07:26] Jim Amsden (IBM): OSLCCORE-54 Extend ReSpec to support oslc:range and oslc:allowValues for enumerations

[07:27] Martin Sarabura (PTC): Jim: This can be deferred, table has been created by hand as needed. Nick already marked it as such

[07:27] Jim Amsden (IBM): OSLCCORE-53 Does a ServiceProvider have one Service per oslc:domain value, or does it reflect the server's desired container structure.

[07:27] Martin Sarabura (PTC): https://issues.oasis-open.org/browse/OSLCCORE-53

[07:28] Martin Sarabura (PTC): Ian was going to review - come back when he joins the call

[07:28] Jim Amsden (IBM): OSLCCORE-55 oslc:describes on ResourceShape should be optional in ReSpec

[07:28] Martin Sarabura (PTC): https://issues.oasis-open.org/browse/OSLCCORE-55

[07:28] Martin Sarabura (PTC): Related to 43, respec should not crash if not specified

[07:29] Martin Sarabura (PTC): Nick: Not marked as deferred, 43 is the driver

[07:29] Jim Amsden (IBM): OSLCCORE-39 Names and descriptions of impact analysis direction properties are misleading

[07:29] Martin Sarabura (PTC): https://issues.oasis-open.org/browse/OSLCCORE-39

[07:30] Martin Sarabura (PTC): Jim: Property defines relation between subject and object nodes in a graph; impact perceived in same direction or opposite or both?

[07:31] Martin Sarabura (PTC): Nick: The current names are misleading

[07:31] Martin Sarabura (PTC): Impact always flows downstream - downstream is impacted artifact

[07:32] Martin Sarabura (PTC): Names don't imply impact

[07:32] Martin Sarabura (PTC): Proposal is to maintain old terms and mark them archaic

[07:33] Martin Sarabura (PTC): Recommend tools use new terms but accept old ones - some compatibilty cost

[07:33] Martin Sarabura (PTC): Nick: Don't know of tools that use them

[07:34] Martin Sarabura (PTC): Martin S: Agree, we don't use the terms

[07:35] Martin Sarabura (PTC): Jim: We may be creating the very incompatibility that we're trying to avoid

[07:36] Martin Sarabura (PTC): Jim: Propose define new terms, non-normatively specified, mark old terms archaic

[07:36] ian green: ian joined

[07:36] Martin Sarabura (PTC): Non-normatively specify how they are related to old terms

[07:37] Martin Sarabura (PTC): Don't use OWL same-as

[07:38] Martin Sarabura (PTC): Nick to propose new names, and will check what they were intended to mean is the what he thinks it should mean

[07:38] Jim Amsden (IBM): OSLCCORE-53 Does a ServiceProvider have one Service per oslc:domain value, or does it reflect the server's desired container structure.

[07:38] Martin Sarabura (PTC): https://issues.oasis-open.org/browse/OSLCCORE-53

[07:39] Martin Sarabura (PTC): Relationship between spc, sp, service, oslc usage property and relationship to underlying LDPCs

[07:39] Martin Sarabura (PTC): Jim: Can be completely orthogonal - may not have contents

[07:40] Martin Sarabura (PTC): Usage: Why a service should also be an LDPC.

[07:41] Martin Sarabura (PTC): Service can have multiple creation factories, multiple dialog uris, etc each marked with a different usage

[07:41] Martin Sarabura (PTC): Jim: Therefore, service should be an ldpc and service is free to decide its contents

[07:41] Martin Sarabura (PTC): Therefore no issue

[07:42] Martin Sarabura (PTC): Ian has not yet done full review of discovery document

[07:42] Martin Sarabura (PTC): Ian: By this time next week

[07:43] Martin Sarabura (PTC): Jim: All the open issues

[07:43] Martin Sarabura (PTC): Nick: Propose to drop entirely extensions part of shape document

[07:44] Martin Sarabura (PTC): Nick: OSLC2 shapes to be carried forward, but in doing so OSLC3 text largely taken out of W3C spec

[07:44] Martin Sarabura (PTC): Some recommendations speculative

[07:45] Martin Sarabura (PTC): No experimental implementation so we'd be standardizing something for which we have no prior art

[07:45] Martin Sarabura (PTC): Especially since we have SHACL coming

[07:45] Martin Sarabura (PTC): Jim: Propose to remove section 9 of https://tools.oasis-open.org/version-control/svn/oslc-core/trunk/specs/resource-shape.html#recommended-extensions

[07:46] Martin Sarabura (PTC): Jim: Read old spec to ensure nothing being introduced to make incompatible, found nothing

[07:47] Martin Sarabura (PTC): Jim: Only included because it was added to W3C

[07:48] Martin Sarabura (PTC): W3C has gone in direction of SHACL, which has lost some features but basically the right direction

[07:49] Martin Sarabura (PTC): Driver is compatibility with OSLC2

[07:49] Martin Sarabura (PTC): Not clear that SHACL is compatible, or that it has same coverage

[07:49] Martin Sarabura (PTC): Proposal: Drop section 9, seconded by Nick.

[07:50] David Honey (IBM): +1

[07:50] ian green: +1

[07:50] Harish (Software AG): +1

[07:50] Martin Sarabura (PTC): +1

[07:50] Jim Amsden (IBM): +

[07:50] Jean-Luc Johnson (Airbus): +1

[07:50] Nick Crossley (IBM): +1

[07:50] Jad El-khoury (KTH): +1

[07:50] Martin Sarabura (PTC): Carried

[07:51] Martin Sarabura (PTC): https://wiki.oasis-open.org/oslc-core/PerDocumentStatus

[07:51] Martin Sarabura (PTC): Skip core vocabulary discussion for now

[07:51] Martin Sarabura (PTC): Resource shapes: Feedback from Jad

[07:51] Martin Sarabura (PTC): Use of oslc:range. Seems like it is rarely used

[07:52] Martin Sarabura (PTC): Nick: Not many discussions about recommended use in early days

[07:53] Martin Sarabura (PTC): Since then some of the IBM tooling has used oslc:range - would be useful for any other tool that would like to integrate at least with IBM tools

[07:53] Martin Sarabura (PTC): But generally useful for all applications

[07:53] Martin Sarabura (PTC): Jim: Why were they not included? Nick: Initially there was no shape at all, just evolved over time

[07:54] Martin Sarabura (PTC): Nick: Just didn't come up

[07:54] Martin Sarabura (PTC): Jim: Seems beneficial to add

[07:55] Martin Sarabura (PTC): Jim: eg., validated by property in requirement amangement domain

[07:55] Martin Sarabura (PTC): Probably over-constraining target value of property

[07:56] Martin Sarabura (PTC): eg., why is created by not a foaf?

[07:57] Martin Sarabura (PTC): Tools should be allowed to constrain further. Maybe the issue is that it's hard to know what should be constrained and in what way?

[07:58] Martin Sarabura (PTC): Nick: If we word it with a degree of permissiveness then we can talk about valid constraints

[07:58] Martin Sarabura (PTC): Jim: How to distinguish between normative or non-normative?

[07:59] Martin Sarabura (PTC): Ian: Can still leave it "soft" in the spec

[07:59] Martin Sarabura (PTC): Jim: What does "applicable" mean? Not consistent, tools do what they want

[08:00] Martin Sarabura (PTC): Ian: Is there any expectation to act differently?

[08:00] Martin Sarabura (PTC): Nick: Guides the UI

[08:02] Martin Sarabura (PTC): Martin: Maybe helps guide the UI design, project plan

[08:03] Martin Sarabura (PTC): Ian: It's a good start anyway

[08:03] Martin Sarabura (PTC): Section 6.2: constraint worded as SHOULD

[08:04] Martin Sarabura (PTC): Jim: Server can do whatever it wants. Therefore oslc:range can be applied as a guide

[08:05] Martin Sarabura (PTC): Jim to raise an issue on this and solicit further feedback

[08:06] Martin Sarabura (PTC): Next: Delegated dialogs

[08:06] Martin Sarabura (PTC): Feedback from Ian: Guidance on oslc:label?

[08:08] Martin Sarabura (PTC): Clickjacking section has been augmented

[08:08] Martin Sarabura (PTC): Prefill on selection dialog?

[08:09] ian green: see "constraints" in https://wiki.oasis-open.org/oslc-core/DelegatedUIUseCases

[08:10] Martin Sarabura (PTC): Martin: context used to inform dialog

[08:12] Martin Sarabura (PTC): We don't rule it out, nor do we describe it

[08:13] Martin Sarabura (PTC): Jim: Non-normative section saying prefill is allowed, servers can choose to support it, spec silent on the details

[08:14] Martin Sarabura (PTC): Ian: Concern for forward compatility?

[08:15] Martin Sarabura (PTC): Pre-fill just means post to selection dialog factory

[08:16] Martin Sarabura (PTC): Implementation that chooses to use this ability does not constrain the spec going forward

[08:17] Martin Sarabura (PTC): Jim: Could imagine allowing the URI to be augmented by query fragment to satisfy Martin's use case

[08:18] Martin Sarabura (PTC): Ian: Already implementations out there that allow POST with pre-fills for selection dialog uri

[08:19] Martin Sarabura (PTC): If we do decide to extend in this area can we be sure that we won't break existing implementations?

[08:19] Martin Sarabura (PTC): Jim: Add normative clause "servers may support pre-fill post to selection dialog and you get uRI to temporary dialog"

[08:20] Martin Sarabura (PTC): Server can interpret the information however it wishes to

[08:20] Martin Sarabura (PTC): We don't specify the content of the iFrame or how generated, nor do we need to specify the contents of pre-fill

[08:21] Martin Sarabura (PTC): Information given to pre-fill may be constrained by shape

[08:24] Martin Sarabura (PTC): Ian: Selection and creation both have associated shape - selection shape describes the resources you could select

[08:25] Martin Sarabura (PTC): Pre-fill content intended to inform client how it can constrain the results

[08:26] Martin Sarabura (PTC): Jim to put clause in and request review

[08:26] Martin Sarabura (PTC): Should we remove 6.4.6? MCS note: Should this be 6.4.7?

[08:27] Martin Sarabura (PTC): Client could very well cache prefilled dialogs; if no longer valid it could get rejected by server

[08:28] Martin Sarabura (PTC): However probably best practices

[08:28] Martin Sarabura (PTC): Ian: These are resources just like any other - they could get stale

[08:28] Martin Sarabura (PTC): David: Couldn't we just use no-cache or etag?

[08:28] Martin Sarabura (PTC): Ian: Yes

[08:29] Martin Sarabura (PTC): Jim will remove clause

[08:30] Martin Sarabura (PTC): Jim: Useful as a non-normative descriptive section

[08:30] Martin Sarabura (PTC): Dialog label still outstanding

[08:31] Martin Sarabura (PTC): Out of time, meeting adjourned

[08:31] Martin Sarabura (PTC): Next call core vocabulary

oslc-core message