W3C home > Mailing lists > Public > public-declarative-apps@w3.org > April 2015

Hypermedia acid test

From: Martynas Jusevičius <martynas@graphity.org>
Date: Tue, 21 Apr 2015 00:45:56 +0200
Message-ID: <CAE35VmxCwuZ7jP98GtXzT9-hh_i+bMOJsv1=GisDBk2+JpjMHg@mail.gmail.com>
To: Hydra <public-hydra@w3.org>, public-declarative-apps@w3.org
Hey everyone,

I had this idea for a while of some kind of test for hypermedia agents.

Consider 2 components:

1. a server S that serves data from a dataset D

The server implements HATEOAS using vocabulary V to serve all possible
states for data output and input.
Vocabulary V defines the meaning of transitions between states.

2. client C that renders server responses as a user interface UI. The
UI renders all states and transitions served by S.

The client has 2 modes:
- local, where it uses responses of an embedded server with dataset DL
- remote, where it uses responses of a remote server with dataset DR,
communicating over HTTP (the concept of a Linked Data browser)

As a result, C can access its own dataset over HTTP by working
simultaneously in modes a and b.

My "thesis" is: With the conditions that
- the only shared knowledge between S and C is V
- datasets DL and DR are equal
the S/C architecture is only really HATEOS if the UI is exactly the
same for every application state in local client mode and remote
client mode.

Does this make any sense? :) It could most likely be simplified, but
probably becomes harder to demonstrate.

This requires a client that can render and edit local and remote data
in exactly the same way. Formats such as JSON-LD as well as RDF/XML
should be useful for this.

Received on Monday, 20 April 2015 22:46:23 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 19:30:57 UTC