[Application-profiles-ig] "Binding" profile entities - a simpler way

Mon Sep 16 15:02:20 BST 2019

On Mon, Sep 16, 2019 at 10:37:10AM +0200, Tom Baker wrote: >The entities in a profile (whatever they are called) must>somehow be "bound", or "mapped", to the appropriate parts>of instance data. I propose the following rules:>>1. An Entity Description (in the data) is bound to a>Entity Description Template (in the profile) on the>类型声明的基础。例子:一个实体>Description with a statement that the entity is of>type "foaf:Person" would be bound to a Entity>Description Template with a corresponding entity type>constraint.Since posting this morning I have discussed this with Eric, who points out that if one Entity Description Template were designated as the starting point, there would be no need to add explicit types to the other EDTs simply for the purpose of associating EDTs with entities in the data. It would suffice to put a type onto the starting-point EDT (for the purpose of identifying that starting point in the data). After that, parts of the data would be associated with the starting-point entity by means of properties. (This assumes that the descriptions in a description set are somehow related (as in a metadata record), and that one of the descriptions is somehow central, which fits nicely with the original motivation for DSP. Of course, this may not always be the case, but would it be good enough for our purposes to make this assumption?) Example: 1. A book should be the starting point of a profile because the descriptions of authors and publishers are all related to the book. 2. The Entity Description Template "book" is matched up to descriptions (in the data) that have an explicit type of ex:Book. 3. The book has a dc:creator, the value of which is defined in the Entity Description Template "author". 4. The book has a dc:publisher, the value of which is defined in the Entity Description Template "publisher". Only one thing would need to have an explicit type -- the Book. Eric notes out that starting points are used in XML Schema in order to ensure that documents are processed completely. Tom -- Tom Baker <tom at tombaker.org>