Introduction to Web Standards - Web Standards: Mastering HTML5, CSS3, and XML - page 15

Amaya
Unlike other browsers that are typically used exclusively to display web documents, Amaya was a free, open source
web browser and a WYSIWYG web developer environment in one. Managed by W3C, Amaya supported HTML 4.01,
XHTML 1.0, XHTML Basic, XHTML 1.1, HTTP 1.1, MathML 2.0, CSS2, and SVG [68]. Distributions were available for
Windows, Linux, and Mac. While still available, the latest version was released in 2012 when Amaya's development
was discontinued.
Testing the Standard Support of Browsers
Web browsers can be tested from two different aspects: supported standards and user experience through
performance and features. The development of browser tests depends on the approach and aim. From the web
design perspective the first one is more important. The most influential browser tests are platform-independent, and
can compare not only different browsers, but also the same browser running on different platforms. Incomplete or
incorrect standard support is usually tested with comprehensive tests, many of which are public. Most browser tests
require human evaluation to identify bugs with full certainty.
Automatic layout tests are difficult to execute on mobile devices because layout tests generate a screenshot and
compare it to a reference image.
Two different versions of the same web page supposed to result in exactly the same rendering can be compared
using ref-tests [69]. Mozilla also provides an automated testing framework using MochiKit JavaScript libraries [70].
The test applies JavaScript function calls. These browser-specific tests cannot be used for general browser comparisons.
The World Wide Web Consortium provides HTML tests [71], CSS test suites [72], mobile tests [73], a MathML test
suite [74], SVG test suites [75], and internationalization tests [76]. W3C is open to browser test suggestions as well [77].
The web design community develops specialized browser tests such as the HTML5 Test [78] and the CSS3 Test [79].
Standard Compliance Tests
The most well-known browser tests for standard compliance are the Acid tests . The name refers to the acid tests used
for gold assessment. Instead of gold purity, however, these Acid tests provide a fast and easy-to-understand indication
of the standard compliance of rendering engines. In spite of that, the Acid tests have always been criticized for testing
not only the core markup and style sheet compliance, but also a set of rarely used features, along with those without a
finalized specification.
The first version of the Acid test, Acid1 , was written in HTML 4.0 Strict in late 1998 to check interoperability issues
between earlier web browsers. Acid1 tests several features with stress on compliance with the CSS1 specification on a
page against a reference image [80]. According to the document title, Acid1 is a “display/box/float/clear test.”
Acid2 is a test page published by the Web Standards Project in 2005. Again, a reference image is provided that
should be compared to the rendered version. Note that the nose should change to blue when the mouse hovers over
the face [81]. Beyond this hovering effect, Acid2 tests the paint order, the object element, data URIs, alpha transparency
of PNG images, and several CSS features (absolute, relative, and fixed positioning, the CSS box model, CSS table
formatting, CSS generated content, and CSS parsing). Safari was the first among the widely used browsers that passed
Acid2 in late 2005. Others followed two to three years later. For example, Firefox passed the test from version 3.0.
Acid3 has been launched in 2008, and updated in September 2011. The 100 subtests grouped in 6 “buckets”
cover various parts of the following standards: data URI scheme, HTTP 1.1 Protocol, HTTP status codes, Unicode 5.0
UTF-8 and UTF-16, the image/png and text/plain content types, the HTML 4.0 Transitional, HTML 4.01 Strict, and
XHTML 1.0 Strict markup languages, DOM Level 2 (Core, HTML, Events, Style, Views, Traversal, Range), the object
element, ECMAScript (including garbage collection), CSS selectors, SVG 1.1 (including fonts), and SMIL 2.1. Not only
those browsers fail the test that cannot achieve the score 100/100, but also the ones that cannot render the animation
smoothly or render it differently than what is presented in the reference (Figure 1-2 [82]).