W3C home > Mailing lists > Public > public-css-testsuite@w3.org > February 2012

Re: Requirements for (level >=3) tests

From: Aryeh Gregor <ayg@aryeh.name>
Date: Thu, 16 Feb 2012 13:10:46 -0500
Message-ID: <CAKA+AxkCqX6Nvoh8R0r4MMd_EpmaPr_Tm4ZkpUxaDGvo5pKxQA@mail.gmail.com>
To: Øyvind Stenhaug <oyvinds@opera.com>
Cc: CSS-testsuite <public-css-testsuite@w3.org>
On Thu, Feb 16, 2012 at 11:29 AM, Øyvind Stenhaug <oyvinds@opera.com> wrote:
> I have assumed that the guidelines at
> http://wiki.csswg.org/test/css2.1/contribute etc also apply to non-CSS2.1
> tests, even though the URLs, breadcrumbs and content tend to mention 2.1
> specifically. Whether this is true or not, I am wondering which parts are
> absolutely *required* for submitted tests. Maybe that could be stated more
> clearly.
> For instance, is the XHTML format mandatory for other testsuites? (It seems
> not all of them use it.) Are #help <link>s always necessary for building?
> Of course, if I write a new testsuite I will try to keep all of these
> guidelines in mind. However, for existing and more ad-hoc tests, and tests
> written by others, the higher the bar the harder it will be to find the
> required time to do all the work for converting and submitting.

To date, I've been ignoring the CSS 2.1 guidelines.  Just as a
motivating example, I currently have a ref file at
contributors/aryehgregor/incoming/ref-3d/green-rect.html, which
consists of

<!doctype html>
<div style="width: 100px; height: 100px; background: lime"></div>

If I read the guidelines correctly, this should really be

<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN"
<html xmlns="http://www.w3.org/1999/xhtml">
  <title>CSS Reftest Reference</title>
  <link rel="author" title="Aryeh Gregor" href="mailto:ayg@aryeh.name"/>
  <style type="text/css"><![CDATA[
   div { width: 100px; height: 100px; background: lime }

I think I'm not being unreasonable when I suggest this requirement is
onerous, given that the two files render identically in all browsers
and contain essentially the same information.  Requiring XHTML 1.0
Strict may have made sense in the early days of the CSS 2.1 test
suite, but it doesn't make sense now.

I really hope no one brings up the issue that HTML5 is still in WD.  I
believe my HTML file here will work interoperably in all browsers
since probably 1998 at least, even though it happens not to have been
endorsed by any kind of specification until maybe 2005 or 2006.  XHTML
1.0 was officially standardized in 2000, but has always been and will
always be ignored by browsers for the vast majority of content on the
web, viz., files served as text/html.  Suggesting that XHTML 1.0 is
preferable because of its official standardization status rather than
paying attention to what browsers actually implement would indicate a
severely problematic attitude toward testing and interoperability.
We're here to get browsers interoperable, not adhere to process for
its own sake, right?
Received on Thursday, 16 February 2012 18:11:58 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:13:24 UTC