Known errors or unintended results?

Several of the lisitngs under "Known Errors" in the CSS2 errata
(<http://www.w3.org/Style/css2-updates/REC-CSS2-19980512-errata.html>)
change passages that were not inconsistent with the rest of the
specification.  These passages cannot be considered errors per se; rather,
they are the result of an error in conveying an intention.  Example of such
changes include allowing underscore in unescaped in identifiers; allowing
CRLF as a numeric escape terminator; eliminating the double sign for
<length>, <percentage>, and <angle> terms; using 96 dots per inch as a
reference for 'px' units; allowing bare zero as an <angle>,  <time>, and
<frequency>; and allowing the 'transparent' keyword as a value for the
elemental border color properties.

I propose that each "known error" erratum cite one or more passages which
the error contradicts.  The cited passages must be at least as normative as
the passage in which the error is found.   If no contradictory passages are
to be found, the ostensible known error is, in fact, not an error, and the
erratum must be filed under "Changes".

--
Etan Wexler <mailto:ewexler@stickdog.com>

Received on Thursday, 14 November 2002 07:10:50 UTC