Implementations fail a CSS1 test suite

I have devised a test suite that mostly deals in CSS level 1, a delayed
response to an friendly challenge [RFT]. The test suite is available at

    http://purl.org/NET/CSS-TEST-EW/

... Among the implementations that fail parts of the suite are Internet
Explorer 6 (for Windows), Internet Explorer 5 Macintosh Edition, Opera 6
(for Mac OS), and iCab 2.9. Some of the bugs uncovered are fundamental,
affecting tokenizing, parsing, and the cascade. These failures strengthen
the case for deprecating CSS1 or for reverting CSS1 to Candidate
Recommendation.

I respectfully ask that the W3C and vendors remove and correct claims of
conformance [WN] [IE] [IEM] [O].

Versions of some of the tests predate and prompted my denial of conformance
to CSS1 on the part of some user agents [NC]. The original tests lacked
clarity, brevity, and consistency; for these reasons and others I delayed
publication. The current tests, although lacking harness pages, do a halfway
decent job of following the Working Group's test suite documentation [TSD]
and guidelines [TSGL]. (Certainly nothing that I'd call "conformant", but
then I'm a stickler about such things.)

At this time I wish to renew my call [CALL]: has anybody produced a
CSS1-conformant user agent?

I wish also to reiterate that no test suite can ever hope to fully cover all
possible cases of valid and invalid CSS [COV-EW] [COV-JM] [COV-IH],
especially in combination with the variations created by document languages.
It follows that conclusive demonstration of conformance is impossible. (Some
proof other than direct demonstration may be possible.)

For what it's worth, I haven't in the least enjoyed staying irritated at
misimplementations over the past year [MIS].

[RFT]
Tantek Çelik (as Tantek Celik).
"Re: conformance (was layout solutions blah blah blah)".
Public correspondence on <mailto:www-style@w3.org> and
<mailto:www-html@w3.org>.
24 January 2002.
<mid:SMARTHOST1hbhSpk6bc0000938f@smarthost.redmond.corp.microsoft.com> /
<http://lists.w3.org/Archives/Public/www-style/2002Jan/0070.html>.
"I'd say that leaves the burden of proof in the naysayers' court."

[WN]
Editor: Bert Bos.
"CSS Browsers" in "Cascading Style Sheets home page".
General information from W3C Cascading Style Sheets Working Group.
5 March 2003.
<http://www.w3.org/Style/CSS/#browsers>.
"Microsoft shipped Internet Explorer 5 for the Macintosh. It apparently
supports full CSS1, the first browser to do so."

[IE]
Microsoft Corporation.
"Features".
Product documentation.
Undated, circa early 2003.
<http://www.microsoft.com/windows/ie/evaluation/features/default.asp#section
5>.
"Provides full support for Cascading Style Sheets, Level 1"...

[IEM]
Microsoft Corporation.
"New features of Internet Explorer 5 for Mac".
Product documentation.
Undated, circa early 2003.
<http://www.microsoft.com/mac/products/ie/ie_main.asp?embFName=ie5features.a
sp>
"In Internet Explorer 5 for Mac, Microsoft has achieved full standards
support for" ... "Cascading Style Sheets (CSS) 1.0"...

[O]
Opera Software ASA.
"Web Specifications Supported in Opera 6".
Product documentation.
Undated, circa early 2003.
<http://www.opera.com/docs/specs/opera6/>.
"Opera 6 supports all of CSS1."

[NC]
Etan Wexler.
Public correspondence on <mailto:www-style@w3.org>.
24 January 2002.
"layout  solutions)ions of CSS for years to come (was RE: CSS multicolumn
layout  solutions)"
<mid:20020125031926.30BA214F7A5@server11.safepages.com> /
<http://lists.w3.org/Archives/Public/www-style/2002Jan/0072.html>.
"I am sorry to bear this message, but neither Internet Explorer 5 Macintosh
Edition nor Internet Explorer 6 conformantly implements CSS1."

[TSD]
Editors: Tantek Çelik, Ian Hickson.
"CSS Test Suite Documentation".
W3C non-normative documentation.
29 January 2003.
<http://www.w3.org/Style/CSS/Test/testsuitedocumentation-20030129.html>.

[TSGL]
Editors: Tantek Çelik, Ian Hickson.
"CSS2.1 Test Case Authoring Guidelines".
W3C non-normative documentation.
Undated, circa January 2003.
<http://www.w3.org/Style/CSS/Test/guidelines.html>.

[CALL]
Etan Wexler.
"solutions)ntations of CSS for years to come (was RE: CSS multicolumn layout
solutions)"
Public correspondence on <mailto:www-style@w3.org> and
<mailto:www-html@w3.org>.
24 January 2002.
<mid:20020124051112.8A3532AD79@server5.safepages.com> /
<http://lists.w3.org/Archives/Public/www-style/2002Jan/0063.html>.

[COV-EW]
Etan Wexler.
"Re: conformance (was layout solutions blah blah blah)".
Public correspondence on <mailto:www-style@w3.org>.
25 January 2002.
<mid:20020125131928.20B416353@server1.safepages.com> /
<http://lists.w3.org/Archives/Public/www-style/2002Jan/0075.html>.
"I decline to confuse conformance with passing an incomplete test suite."

[COV-JM]
Jesse McCarthy.
"Re: conformance (was layout solutions blah blah blah)".
Public correspondence on <mailto:www-style@w3.org>.
25 January 2002.
<mid:MWMail.qaakpjog@host.none> /
<http://lists.w3.org/Archives/Public/www-style/2002Jan/0081.html>.
"Perhaps a more comprehensive test suite is desirable, but it should be
realized that it will never be definitive."

[COV-IH]
Ian Hickson.
"Re: WD-CSS21-20020802 section 8, "Box model", substantive comments"
Public correspondence on <mailto:www-style@w3.org>.
28 February 2003. 
<mid:Pine.LNX.4.50.0302280537580.23128-100000@dhalsim.dreamhost.com> /
<http://lists.w3.org/Archives/Public/www-style/2003Feb/0277.html>.
..."an over-simple test suite (which is utterly inadequate for establishing
overall compliance)"...

[MIS]
Tantek Çelik (as Tantek Celik).
"Re: solutions)ntations of CSS for years to come (was RE: CSS  multicolumn
layout  solutions)".
Public correspondence on <mailto:www-style@w3.org> and
<mailto:www-html@w3.org>.
24 January 2002.
<mid:SMARTHOSTiiZIHd0TlK00007be2@smarthost.redmond.corp.microsoft.com> /
<http://lists.w3.org/Archives/Public/www-style/2002Jan/0069.html>.

Received on Monday, 10 March 2003 05:44:01 UTC