Implementers

For each of our Recommendation-track deliverables (e.g. technical specs), we need to establish a list of implementers that will create and deploy implementations that conform to that spec and pass the tests for that spec.

In order to progress from Candidate Recommendation (CR) phase, we need to have multiple implementations of each specification, in products that are significant to the market. The W3C Process Document requires that we "show that a specification is sufficiently clear, complete, and relevant to market needs, to ensure that independent interoperable implementations of each feature of the specification will be realized", and suggests that a group "may wish to develop tests in concert with implementation efforts", with "implementation experience at all levels of the specification's ecosystem (authoring, consuming, publishing…)". In practice, the expectation for the last decade or so has increasingly meant a comprehensive test suite and a set of conforming market-leading implementations.

For each spec, we need to determine what the ecosystem is, and who the market leaders are in that ecosystem. Since our specs are meant to be small interoperable pieces, each implementer may have one or more products that conform to one or more specs.

Annotation Model

The ecosystem (and its market leaders) seem to fall into two types (clients and services) and several classes within each type:

• clients, which allow an author to create annotations, and/or allow a reader to consume those annotations
  • dedicated standalone annotation clients (standalone)
  • eReader ("reading system") apps (ereader)
  • browser-based JavaScript libraries (JS lib), either site-specific or multi-domain
  • native browser implementations (browser)
• services, which serve as a place to publish, store, and/or provide an API for searching and retrieving annotations
  • publishers where authorized users can publish annotations (publisher)
  • providers where clients can retrieve annotations (provider)

Candidates

Annotation Protocol

• client, which sends and retrieves messages from the User Agent
• server, which sends and retrieves messages from the publishing service

Candidates

FindText API

This is purely client-side, and while the target is native browser implementation, there are other environments and polyfills possible

• • dedicated standalone annotation clients (standalone)
  • eReader ("reading system") apps (ereader)
  • browser-based JavaScript libraries (polyfill)
  • native browser implementations (browser)

Candidates