Management of Linked Data
In order to handle a larger amount of linked data a comprehensive and consistence strategy for managing it is inevitable. This includes the naming of resources, the naming of graphs, strategy for storing, interlinking and cleansing.
Resources are identified uniquely by HTTP URIs.
The base URI is http://finder.tuwien.ac.at/. The machine-readable description can be found at the URL base/<id> and is at the same time the unique identifier for the resource.The human-readable page of the resource can be found atbase/page/<id>. The id is a path that identifies the desired resource uniquely.
| URI | Description |
|---|---|
/spatial |
References to the graph that contain the complete data set of spatial information |
/spatial/building/id/<buildingID> |
References to the building with the given ID |
/spatial/buildingtract/id/<buildingtractID> |
References to the building tract with the given ID |
/spatial/floor/id/<floorID> |
References to the floor with the given ID (floor code) |
/spatial/room/id/<roomID> |
References to the room with the given ID (room code) |
| URI | Description |
|---|---|
/datasets |
Gets a document listing a summary of all local data sets (graphs) |
/spatial |
References to the graph that contain the complete data set of spatial information |
| URI | Description |
|---|---|
/vocab |
Gets a document listing a summary of all local vocabularies. |
/vocab/spatial# |
References to the spatial vocabulary. |
As mentioned above resources are uniquely identified by URIs of the form base/<id> that are consequently used by the resolving process to get the description of the corresponding resource. The desired content type of the description can be specified in the HTTP header. This is a quite useful tool of HTTP called content negotiation. The following table maps the format to the content type of HTTP.
| Format type | Readable for | Content type (HTTP header) |
|---|---|---|
| HTML | human | text/html |
| RDF | machine | application/rdf+xml |
| Turtle | machine | text/turtle |
| JSON-LD | machine | application/ld+json |
For web browsers text/html is the default content type.
If the client wants to see a machine-readable description, the server simply returns the description in the desired format (RDF/XML,JSON-LD, Turtle). Otherwise if the client wants to see the HTML page of the resource, the server redirects the client to the corresponding URI (base/page/<id>) and then the corresponding HTML page is returned. As a result an additional HTTP request is required for clients functioning as a HTML browser. This approach is called 303 URIs by cooluris. The mechanism is visualized in the following graphic.
