- From: Joseph Reagle <reagle@w3.org>
- Date: Fri, 4 Apr 2003 14:21:26 -0500
- To: "Hallam-Baker, Phillip" <pbaker@verisign.com>
- Cc: www-xkms@w3.org
On Friday 04 April 2003 13:40, Hallam-Baker, Phillip wrote: > I have been looking at the issue of QNames in the examples, they appear > to me to be correct. My mistake, the in-scope default namespace is declared for the value space, and the lexical value does not require the prefix: it's optional. Regarding the prefix in the running text, in XENC I used the convention that the xenc element/attribute types were unprefixed and the external types (i.e., dsig) were prefixed, both of which corresponded to the examples. XENC and DSIG identifiers are URIs, so that's what the examples showed though in the running text I'd use an entity (e.g., &dsig;KeyInfo). Those conventions work well for element/attribute types but can be confusing for QNames that are used as identifires. The SOAP primer simply uses prefixes for everything; WSDL is specified mostly as an Infoset description, but in its table and examples it uses the prefixes [1]. Which ever way we go, we should strive for some consistency such that one can search for and see these things consistently in the spec. So that would mean either all XKMS types are prefixed or not. In tables of identifiers, I like the prefixes (or it should be made clear all those types in the table have certain NS declarations in scope) but in the examples I don't like to see the "xkms:" prefix anywhere... I suppose I prefer the former to the latter ... ? [1] http://www.w3.org/TR/wsdl12/#language-extensibility
Received on Friday, 4 April 2003 14:21:33 UTC