Talk:Best practices for stable URIs

From pro-iBiosphere
Jump to: navigation, search

303 vs 303-redirect

The following is not right:

The advantage of the hash method over the alternative 303-redirect method is that it is twice as fast as the 303 method. It returns the desired information resource in a single http service call, whereas the 303-redirect requires two http calls.

Yes, 303 method requires two calls, but that's very minimal overhead, and doesn't mean the hash method is "twice as fast". For most real-world responses, the bulk of the call will be in transferring the content of the document, not the overhead of the HTTP call. Depending on implementation, the 303 method is often faster because the documents can be customized to the requested resource, thus smaller. Chris Maloney (talk) 05:32, 23 August 2013 (CEST)

Thanks, Chris! --Daniel Mietchen (talk) 07:41, 23 August 2013 (CEST)
Yes and No. In many situations with little data to transfer the dominant problem may be server latency. This doubles with 303 (2 independently queued request). Also, the hash methods returns can equally customized to the requested resource, this is equal (in practice the hash method is used often to return entire ontologies, which is a good choice if most consumers need a large portion of the terms in an ontology). -- Gregor Hagedorn (talk) 21:19, 23 August 2013 (CEST)