Triplify update vocabulary

Purpose

When RDB data is published on the Web e.g. as Linked Data it is important to keep track of DB (and hence RDF) updates so crawlers know what has changed (after the last crawl) and should be re-retrieved from that endpoint.

To have a centralized registry (such as e.g. implemented by PingTheSemanticWeb service) does not seem to be feasible when Linked Data becomes more popular – think of millions of Linked Data endpoints pinging such a registry each time a small change occurred.

The approach: Linked Data Update Logs

Each Linked Data endpoint provides information about updates performed in a certain timespan as a special/standardized Linked Data source.

Let's assume the Example.com company provides a Linked Data endpoint with information about their products, employees etc. The endpoint is reachable via http://example.com/lod/.

The LOD endpoint contains a special LOD space below http://example.com/lod/updates which contains information about updates.

http://example.com/lod/update for example will return the following RDF:

http://example.com/lod/update/2008 could then return the following RDF:

This nesting could continue until we finally reach an URL, which exposes all updates performed in a certain second in time. For very frequently updated LOD endpoints (e.g. Wikipedia) this interval of one second will be sufficiently small enough, so the related update information can be still easily retrieved. For rarely updated LOD endpoints (e.g. a personal Weblog) links should only point to non-empty Update Collections in order to prevent crawlers from performing unnecessary HTTP requests.

http://example.com/lod/update/2008/Jan/01/17/58/06 then would for example contain RDF links (and additional metadata) to the Linked Data documents updated on Jan 1st, 2008 at 17:58:06, e.g. following triples:

Individual updates are identified by a sequential identifier (i.e. “user123” in the example). Arbitrary meta data can be attached to these updates, such as the time of the update (probably redundant since that can be inferred from the URL) or a certain person who performed the update.

Invocation

Triplify automatically generates all the resources in the update URI space, when the Triplify configuration $triplify['queries'] contains a query named update. This query has to return at least two columns. The first column contains the date when to update occurred, the second column contains the id of the updated resource. An example is given below:

Example

The workings of the Triplify Linked Data Update Logs can be observed with Triplify's own datasource registry:

http://triplify.org/triplify/update/

Classes

update:UpdateCollection

A collection of updates.

update:Update

An atomic update performed on an RDF resource.

update:Deletion

SubClassOf: update:Update

Represents the deletion of an RDF resource.

Properties

update:updatedResource

Type: ObjectProperty
Domain: update:Update

Points to the resource which was updated.

update:updatedAt

Type: DatatypeProperty
Domain: update:Update
Range: xsd:dateTime

Refers to the date and time when a certain update occurred.

update:updatedBy

Type: Property
Domain: update:Update

Points to the user description who performed the update.