- From: Robert Sanderson <azaroth42@gmail.com>
- Date: Wed, 15 Oct 2014 16:34:24 -0700
- To: Jacob Jett <jgjett@gmail.com>
- Cc: "Denenberg, Ray" <rden@loc.gov>, Web Annotation <public-annotation@w3.org>
- Message-ID: <CABevsUGM1duQ90DPmCkKgF4ikk5yuyt6zdg8=dudeD13MtzOOA@mail.gmail.com>
I think you're both agreeing :) To re-express the last three axioms: 13. There is a resource which we call a Choice, which consists of some number of other (somehow equivalent) resources, of which one should be chosen. 14. There is a resource which we call a Composite, which consists of some number of other resources. 15. There is a resource which we call a List, which consists of some number of other resources in a specific order. If that's clearer? On Wed, Oct 15, 2014 at 3:56 PM, Jacob Jett <jgjett@gmail.com> wrote: > Hi Ray, > > Going back to my juxtaposition example, the annotation needs the composite > resource because it most faithfully communicates what the body is > annotating -- the aggregation of the two resources. We still need to > communicate what that composite is composed of so that the end user tool > has enough information to actually fetch and render the target resources in > the intended juxtaposition. > > Now there may be other ways of doing this, such as punting the aggregation > part to some other model (like a collection model or whatnot). I'm not sure > if that's on the table for discussion. I've certainly thought that much of > the segmentation work in the model could be spun out because I have the > same needs for arbitrary selectors in non-annotation contexts. > > Regards, > > On Wed, Oct 15, 2014 at 4:19 PM, Denenberg, Ray <rden@loc.gov> wrote: > >> Hi Jacob – >> >> >> >> I’m confused though - if the composite is indeed a resource, why point >> to the multiple components rather than to the single composite resource? >> >> >> >> In 15, a list where order is significant, I can see need. >> >> >> >> But if the semantics of “composite” is that it is simply the union of the >> components, what is the use case that requires pointing to the components >> individually? >> >> >> >> Ray >> >> >> >> *From:* Jacob Jett [mailto:jgjett@gmail.com] >> *Sent:* Wednesday, October 15, 2014 3:50 PM >> *To:* Denenberg, Ray >> *Cc:* Robert Sanderson; Web Annotation >> *Subject:* Re: Maximally Abstract Data Model >> >> >> >> Hi Ray, >> >> >> >> This was one of the original use cases for composite. I think more >> properly we might define it as an aggregate resource composed of multiple >> resources. To steal from chemistry, a suspension would be a good analogy to >> a composite resource. >> >> >> >> Regards, >> >> >> >> Jacob >> >> >> >> On Wed, Oct 15, 2014 at 2:35 PM, Denenberg, Ray <rden@loc.gov> wrote: >> >> 14. A Composite is a set from which all of the resources should be used. >> >> 15. A List is an ordered set of resources, of which all should be used. >> >> >> >> I see “composite” as “composite resource”, in other words, it is itself a >> resource, consisting of the “union” (if you will) of the other resources >> which I would call “component resources”. >> >> >> >> (I don’t necessarily see “list” being modeled similarly.) >> >> >> >> Ray >> >> >> >> >> >> *From:* Robert Sanderson [mailto:azaroth42@gmail.com] >> *Sent:* Wednesday, October 15, 2014 3:20 PM >> *To:* Web Annotation >> *Subject:* Maximally Abstract Data Model >> >> >> >> >> >> All, >> >> >> >> On the call today there was discussion about the data model, versus the >> expression of the model using RDF, and then the serialization of that into >> JSON-LD. >> >> >> >> To try and express the current abstract data model as simple statements... >> >> >> >> Annotation Baseline: >> >> >> >> 1. There is a resource which we call an Annotation, that typically >> represents the linking between other resources. >> >> 2. Annotations have 0..n Body resources. >> >> 3. Annotations have 1..n Target resources. >> >> 4. Body resources are related to Target resources, and are typically >> statements about the Target resources. >> >> 5. As separate resources, Annotations, Bodies and Targets have separate >> properties, typically including provenance and descriptive metadata. >> >> >> >> Anchoring: >> >> >> >> 6. We introduce a type of resource called a SpecificResource that >> identifies a more specific entity (more constrained/specialized) than an >> existing resource which is identified by a URI. >> >> 7. SpecificResources have exactly 1 Source resource, that the >> SpecificResource is more specific than (constrained/specialized from). >> >> 8. The constraints on the SpecificResource are specified in 1..n >> Specifier resources. >> >> 9. A State is a type of Specifier that describes the state of a >> resource, to allow the intended representation to be retrieved. >> >> 10. A Selector is a type of Specifier that describes part of a >> representation of a resource. >> >> 11. A Style is a type of Specifier that describes how the resource should >> be presented to the user. >> >> >> >> Multiplicity: >> >> >> >> 12. We introduce three methods of creating sets of resources. >> >> 13. A Choice is a set from which one resource should be selected for use. >> >> 14. A Composite is a set from which all of the resources should be used. >> >> 15. A List is an ordered set of resources, of which all should be used. >> >> 16. Multiplicity constructs can be used where-ever any resource can be >> used. >> >> >> >> >> >> Additional statements welcome :) >> >> >> >> Rob >> >> >> >> -- >> >> Rob Sanderson >> >> Technology Collaboration Facilitator >> >> Digital Library Systems and Services >> >> Stanford, CA 94305 >> >> >> > > -- Rob Sanderson Technology Collaboration Facilitator Digital Library Systems and Services Stanford, CA 94305
Received on Wednesday, 15 October 2014 23:34:52 UTC