NOTE: THIS IS A PRELIMINARY DRAFT, AND MAY BE OBSOLETED OR REVISED AT ANY TIME
Tests below evaluate the ability of browsers to render CSS1 information according to accepted behavior. The tests below are given as examples. The CSS1 specification is covered to identify all testable statements. There is one test case included for each testable statement of the CSS1 specification. Not included are informative units, or units for which there were no testable statements. It is attempted to use the same style sheets within each section, to limit what is being tested. The purpose of these tests is to compare CSS1 implementations against the CSS1 Recommendation, or to demonstrate interoperability of CSS1 implementations from different vendors.
Each test has one or more style sheets, a testable statement, a heading describing the test, and a pass criterion (including one or more reference snapshots for comparison). The "pass criterion" portion covers the style sheet implementation; the rest of the page is just HTML. This is again to limit what is being tested for each test and to control the test environment. The tests assume that (where appropriate) no personal stylesheets are being used, and User Agent default information cannot be tested here; in addition, other factors may affect whether or not the tests are passed. Various assumptions are made, including support for appropriate CSS1 properties. The tests should be able to stand alone, or be used in combination. Also, it is possible to test more than one testable statement in a single test case, although this has not been done here. One should know by looking directly at the browser's rendition and the comparison snapshots whether a test has been passed, although it is possible to look at the source code also. NOTE: Individual comparison snapshots have yet to be done. Also, since certain tests have at this point similar tests for different rules, these may be combined in the future, so that there may just be one test for each set of these "redundant" tests, with all rules the page tests included. Finally, where reader rules are to be tested, the reader may need to "manually" insert reader rules into the test as appropriate and directed.
You have the ability to enter your test information, view previous test results,comment on particular tests, or propose new tests. Results of the tests are available to any visitor to the site. In the future, a facility may be added where you can dynamically test your customized CSS code interactively.
Running your browser through the test suite is a way for you to contribute to the collective knowledge of the Web development community. When you participate, you will be offered a series of tests, each of which depicts a particular CSS1 fragment. You will be asked to indicate whether your browser seems to render the CSS1 according to acceptable behavior.
Help information is included below, in the sense that if one wants to test a particular function, there is a mapping of that function to tests for it below. In addition, there is a reference parser covering CSS1. It is not intended to duplicate that capability here, but to complement it.
You may take one, some, or all of the tests. There are some 550 tests included here under development. The syntax of the test name is "xxx-y.htm", where "xxx"refers to the CSS1 structural unit, and "y" refers to the position of the testable statement within that unit. To take a test, just click on the test name, and a rendition should appear on your screen.
Comments are appreciated. This is a preliminary draft ( a strawman), being made available for public review and comment. Please send comments to: frederick.boland@nist.gov, or to the www-style mailing list. These tests are not complete. NOTE: The reference images displayed in each test are placeholders only.
The CSS1 Specification used as a reference for these test cases is the W3C Recommendation REC-CSS1-961217, dated 17 December 1996. To access the CSS1 Recommendation, click here. To find out more about the World Wide Web Consortium, click here.
If you wish, you can assign your browser a score by comparing your browser rendition against the reference images and descriptions of how the rendering should appear. Your score may range from 1 (poor-totally dissimilar) to 10 (excellent-exactly the same). Things to consider include degree of similarity in color, positioning, and general appearance.
NIST assumes no liability for dissemination of this information, but is making it available in the public interest.
Thank you very much.
Click on the appropriate unit number of the CSS1 specification to retrieve the test(s) for that unit, or look at the function(s) desired to determine what tests to run.
NOTE: A CSS2 Specification is under development. Tests for CSS2 are available. All tests for CSS1 also apply to CSS2. Draft tests for CSS2 may be accessed by clicking here