Client and Server Implementation Guidelines for Common HTTP Use cases

TREEcg / specification

RDF vocabulary and hypermedia specification to publish your Linked Data using search trees

29 stars 13 forks source link

Client and Server Implementation Guidelines for Common HTTP Use cases #92

Closed kkostov closed 1 month ago

kkostov commented 1 year ago

When creating a production implementation of the spec (both client and server-side), it may be beneficial to have access to a set of guidelines on how to tackle common technical implementation details:

Handling of redirects along the graph
Handling upgrade of HTTP to HTTPS
CORS
The need to pass custom headers in case of scenarios like authentication or API limits?
...

pietercolpaert commented 1 year ago

Found an interesting example for this: https://iiif.io/api/discovery/1.0/#4-network-considerations

kkostov commented 1 year ago

@pietercolpaert this is an excellent example, I like that the guidelines also appear as a SHOULD.

sandervd commented 1 year ago

A server should define a default Accept header (e.g. if no accept is defined, return content type Turtle)

pietercolpaert commented 2 months ago

Related to #111 and #112 , I think it makes sense to do this in specifications that use the TREE specification that are directed towards server implementers. TREE should not do this as it is directed towards clients.