- From: Jeremy Carroll <jjc@hpl.hp.com>
- Date: Tue, 24 Jan 2006 15:19:45 +0000
- To: Michael Kay <mhk@mhk.me.uk>
- CC: "'Bjoern Hoehrmann'" <derhoermi@gmx.net>, "'Martin Duerst'" <duerst@it.aoyama.ac.jp>, www-international@w3.org
Michael Kay wrote: >> http://www.w3.org/TR/charmod/#C073 has "Publicly interchanged content >> SHOULD NOT use codepoints in the private use area." Reference to and >> use of are equivalent concepts here. You might well argue that SHOULD >> NOT is not MUST NOT, conformance checkers are nevertheless going to >> point out violations of the SHOULD NOT requirement > > I fail to see how a conformance checker is supposed to know whether the > content is being publicly interchanged. I would think that most conformance > checkers are likely to give the user the benefit of the doubt, or at least > to provide options. > > Michael Kay > http://www.saxonica.com/ > > > Personally, I am tending to increasingly strict reading of SHOULDs and SHOULD NOTs. I feel that generic software should implement these as MUST and MUST NOT, with a user configurable switch that allows the user to assert that "the full implications [have been] understood and carefully weighed before choosing a different course." (RFC 2119) In say, a web form based conformance checker, this looks like a check box, whose default value is MUST. For any document based on a Web standard, it is a reasonable assumption that the document is intended for public interchange ... Quite how far an XSLT conformance checker should go in checking conformance with any other standard on which it depends is a matter of taste and judgment. I have removed the two comments lists from the headers on this message, since I do not wish to make any formal comment about either group's taste and judgment. Jeremy
Received on Tuesday, 24 January 2006 15:32:57 UTC