rest/json
Revision as of 20:35, 5 September 2008 by DrErnie (talk | contribs) (→Implementations: Grassy Knoll)
RESTful JSON
Charter
Given that:
- REST is a powerful architecture for scalable web services
- JSON is a lightweight container for exchanging data
- AtomPub, while RESTful, requires XML documents with strict metadata requirements
The goal of the RESTful JSON project is to develop a series of conventions for:
- URLs
- HTTP methods
- HTTP headers
- JSON fields
that:
- is maximally compatible with existing (RESTful, generic) clients
- enables partial updates
- allows paging and/or partial returns of large datasets
- standardizes linking to related resources
- use of hypermedia (links + context) to manage application state
- does NOT become a full-fledged RPC solution
Resources
- Original proposal by Joe Gregorio
- Standardizing RESTful JSON by Kris Zyp
- Google Group
- AtomPub (for comparison/inspiration)
Proposals
Issues
- Is the top-level entity a simple array (Persevere, Dojo) or a full-fledged object (CouchDB, ActiveRecord?)
Implementations
- CouchDB
- Jester (same URLs as ActiveResource)
- Dojo REST store
- DOM Resource
- Persevere
- SproutCore
- Grassy Knoll (Python)