You mean, they can decide whether to be interoperable or to be spec compliant? Please don't put us poor implementers in that position.
-----Original Message----- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Franklin Tse Sent: Saturday, October 20, 2007 5:07 AM To: Anne van Kesteren Cc: atom-syntax; atom-protocol Subject: Re: Atom Bidi Attribute Draft For interoperability, "ltr" and "rtl" in all lowercase should be the best. The specification should not say that implementations are allowed to differ, but it cannot control the real situation. Programmers still can decide how their programs handle the values, even though the spec. said "MUST NOT". -Franklin -------------------------------------------------- From: "Anne van Kesteren" <[EMAIL PROTECTED]> Date: Saturday, 20 October, 2007 19:58 To: "Franklin Tse" <[EMAIL PROTECTED]> Cc: "atom-syntax" <[email protected]>; "atom-protocol" <[EMAIL PROTECTED]> Subject: Re: Atom Bidi Attribute Draft > On Sat, 20 Oct 2007 13:53:06 +0200, Franklin Tse > <[EMAIL PROTECTED]> wrote: >> http://validator.w3.org/check?verbose=1&uri=http%3A%2F%2Fpeaceable_whale.sinacampus.com%2FBidi%2Fbidi-uppercase.xhtml >> >> value of attribute "dir" cannot be "LTR"; must be one of "ltr", "rtl". > > I didn't claim anything about XHTML 1.x. > > >> If the specification says that the value is case-insensitive, then all >> implementations have to do a case insensitive match. If not, >> implementations can choose if they support the case insensitive values >> of "ltr" and "rtl", while the specification allows "ltr" and "rtl" only. > > It doesn't help interoperability if implementation are allowed to differ. > > > -- > Anne van Kesteren > <http://annevankesteren.nl/> > <http://www.opera.com/> >
