See also: IRC log
<scribe> Scribe: SteveS
Arnaud: will start with new day and time next week, Nov 20 Thurs @ 2pm ET
... Skipping approval of minutes
... minutes are still not in proper form, working with team as they are backloged after tpac
hopefully by next week
Arnaud: Skipping tracking of actions and issues, have none yet
Arnaud: probably should open an issue for this, will when tracker up, been a long discussion on this
... summarizes some of the positions based on emails and user stories, based on class or direct relationship
arthur: in context of OSLC, or some organization that is defining vocabularies for some domain and org also defining REST APIs so tools can provide a consistent interface for multi-vendor tools to interoperate
... even in a given tool instance, it can host various scoping mechanisms (typically a projects) which have different configuration, based on a standard vocabulary + some customizations (either new properties or additional constraints)
Possible that even each resource could have its own specialization
Having a resource's type (class) linked to the constraint definition had issues as it would typically cause broader impact on resources controlled by different scoping mechanisms (projects), so we took an approach to link a resource or operation to a given shape.
<hknublau> +q
hknublau: sees that if you don't have this linked to the type triple, you need to have something like that..which could be aligned.
... this would point off to a graph that would describe the instance of that resource
<ericP> It would characterize this as using named graphs for truth maintenance.
<arthur> +q
<ericP> It appears to exclude business processes with different constraints that have to talk to two interfaces at once.
<Irene> +q
hknublau: describes approach he outlined in his email
arthur: getting further clearification on usage of owl:import
<Irene> it is really about unions of graphs
hknublau: simple just using it as a way to link to another group, only owl term that is used
<arthur> +q
Irene: the URI used is based on which graph it is contained it and then when it is dereferenced
<arthur> +q
<hknublau> +q
arthur: using http uris provides a single view of a resource, whether named graphs are in volved or not
<arthur> +q
<arthur> +q
(general discussion on various approaches and motivating use cases)
<Zakim> ericP, you wanted to vocalize
<kcoyle> +q
ericP: summarizing some examples given in emaill that highlight some cases where shapes should be decoupled from type
<hknublau> +q
<arthur> +q
<ericP> i believe it's all type in the instance data
kcoyle: hoping to get clarification on what people are talking about when they are talking about type
arthur: type as present in the instance data, which could be more than one type
hknublau: interested in multiple nesting and context aware constraints, have pointed out in latest response to ericP, by using sparql
<solbrig> +q
+q
<hknublau> +q
solbrig: outlining a different use case, involves what information is required in a graph or what is not in it
... looking at data model we typically find in UML and progressively apply constraints, such as labratory test data
may require gender as part of the test who is of type :Person and also recorder who is of type :Person, which would have different constraints
have attempted to use reasoners which help with what the data denotes but not if it contains certain data
SteveS: summarized how LDP has left off the ability to describe what LDPRS can be created ... hope was in the future something such as shapes chould be applied to define these constraints
<kcoyle> so can we gather more use cases?!
<Irene> the request is to have more detailed use case descriptions
hknublau: was hoping we could get to some details to highlight the test cases and we could demonstrate how existing solutions can address tehm
+q
<Irene> with examples like data snippets because words are ambigious
<arthur> +q
Arnaud: wants to focus on consensus of requirements
<arthur> +1 to more examples
SteveS: think it makes sense to work towards a set of test cases (scenarios) and common requirements, we can continue to evaluate existing technologies and if a closer fit to what would motivate the work product of this WG
arthur: examples will help clarify it, will add them and encourage others to do so
Arnaud: continue to flush out user stories and go through items from DC requirements
<Arnaud> trackbot, end meeting