This is the homepage for EARL, the Evaluation And Report Language being developed by the W3C WAI ER Group.
We expect that EARL will be used more widely as part of the new QA effort at W3C, as the base format for storing results of test runs. EARL is also related to the W3C's Semantic Web Activity.
This document is circulated for discussion about EARL only. Version 0.95 is considered to be a stable version of the language, but note that it is only a beta version for people to start implementing. No guarantee of the efficacy of the language is made at this time, or of relationships to future versions.
We hope that people will start integrating EARL concepts and syntax in their tools, and give us feedback on the ERT mailing list (archives) about functionalities they would like to see added or removed from this language.
The Evaluation And Report Language is an RDF based framework for recording, transferring and processing data about automatic and manual evaluations of resources. The purpose of this is to provide a framework for generic evaluation description formats that can be used in generic evaluation and report tools.
The basic statement we want to express is that:-
In some contextSome resource (e.g. a Web page, a browser)Evaluates-toSome result (e.g. fails to meet a checkpoint, correctly implements CSS 2 display property)
EARL itself comprises of a core model, and extensible vocabulary, of RDF terms. It is our hope that, in the future, other EARL based vocabularies could be made. This modularization means that people aren't constrained to using the properties in the main vocabulary, and that the EARL vocabulary can consist of core terms only.
It should be noted though that at the present time, the EARL vocabulary should be adequate for expressing 95% of evaluations.
EARL would initially be a means for expressing in a machine readable form ...
- Results of evaluating web pages and web sites against the Web Accessibility Content Guidelines
- Storing results of running a online test-suite (e.g. SVG)
- Results of syntax validation of Web resources (e.g. HTML validator)
The language would be extensible to:-
- Evaluations of languages, e.g. as defined by their schemas
- Evaluations of authoring tools
- Repair of above
EARL expresses evaluations, and assertions, about all sorts of languages and tools. We make the distinction that we have a "context" for something - so for example "John says that...", and an assertion is the object of this, so for example "MyPage passes X". The evaluation is simply the context and the object(s) put together. EARL, at its base, is a really simple langauge. For example, the following says that "Bob" asserts that his page passes all the priority 1 checkpoints in WCAG 1.0 (and some other information - all in stripped down Notation3 syntax):-
@prefix : <http://example.org/net/2001/05/bob#> .
@prefix earl: <http://www.w3.org/2001/03/earl/0.95#> .
@prefix rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#> .
:Bob earl:asserts
[ rdf:subject :MyPage;
rdf:predicate earl:passes;
rdf:object :WCAG10P1 ];
earl:email <mailto:bob@example.org>;
earl:name "Bob B. Bobbington" .
:MyPage
a earl:WebContent;
earl:testSubject <http://example.org/net/bob/>;
earl:lastModified "2001-05-07" .
:WCAG10P1
a earl:TestCase;
earl:testMode earl:Manual;
earl:testCriteria
[ earl:suite <http://www.w3.org/TR/WCAG10/>;
earl:level <http://www.w3.org/TR/WCAG10/#wc-priority-1>;
earl:note """All Priority 1 checkpoints conformed to means that
a document is WCAG Level Single-A compliant""" ] .
Note that because EARL is RDF based, it can only make assertions about things that have URIs. There has been some work on using the BNF of languages to transform them into XML syntax so that they may be processed as XML. In this way, it will be possible for EARL to describe and repair all sorts of languages, from XHTML, through CSS, to ECMAScript, and beyond; due to XPath/XPointer.
EARL is an application of RDF (Resource Description Format). Part of the reason behind choosing RDF is set out in "Why RDF is not the XML Model". Hence, EARL follows the RDF model, but we make no recommendation as to the syntax of EARL - as long as the (EARL) model is followed. [N.B. As of Dec. 2001, the WG is leaning towards defining XML RDF as the canonical serialization for EARL 1.0.]
For example, RDF syntax is not limited to XML RDF for it can take on certain (non-normative, as far as the W3C is currently concerned) plain text forms as well. It may be also be that an EARL evaluation processor (EP) can only support a subset of EARL designed for its own particular task.This is fine as long as the EP does not introduce any features outside of RDF Model and Syntax, or the associated Schemas ontologies and logic.
Notation3 (N3 for short) was chosen as the primary syntax to be used in discussions about EARL on the ERT list, which sped up the developmental process somewhat, due to its readability. See the Notation3 specification and N3 Primer for more information.
EARL 0.95 is the latest version of EARL, incorporating revisions from EARL 0.9 (see list of changes). EARL 0.9 is now deprecated - please use 0.95 instead (we've made available a guide with rules files to help you make the conversion).
Also available are some EARL 0.95 implementation notes.
An RDF schema for EARL 0.95 is available (also in Notation3). The persistent namespace for this version of the language is:-
Please note the hash (#) on the end (see also: Cool URIs Don't Change). The EARL datatypes (of which there is only one - "Date") are declared in a separate file.
We have produced three sets of three examples of EARL 0.95 (in Notation3), one set using contexts, one reification, and the other one converted from the old 0.9 examples.
The ER Working Group is currently working on version 1.0 of the language. A non WG endorsed work-in-progress EARL 1.0 specification is available.
See EARL Background.
Comments are most welcome; please send any questions and comments about the content of this document to the ERT mailing list at w3c-wai-er-ig@w3.org.
$Id: Overview.html,v 1.54 2001/12/09 23:04:36 spalmer Exp $