Wednesday, January 12, 2005, 10:51:58 PM, you wrote:
>> The root element of a Structured Extension construct MAY have >> attributes, it MAY contain well-formed XML content, or it MAY be >> empty. > It took me a minute to realize that the content of a structured > extension element could be a text node--that it needn't have element > children. > The name "structured extension" lends itself to this > misperception. I considered calling them Custom Extensions - maybe that would have been clearer? > Perhaps this paragraph could have a sentence or phrase > added to it to ensure that that is clear. For example: > The root element of a Structured Extension construct MAY have > attributes, it MAY contain well-formed XML content (including > unstructured character data), or it MAY be empty. Yes, that probably is clearer. >> The presence of an Extension construct MUST NOT redefine the meaning >> of another Extension construct or core element. > Does this mean that an Extension construct cannot redefine the meaning > of another Extension construct from the same namespace, or does it only > restrict interaction across namespaces? The intention is to assist in the interpretation of a known extension amongst unknown extensions. The proposal doesn't give any special status to elements belonging to the same namespace. The isolation applies accross instances of Extension constructs. For Structured Extension construct, because the interpretation of its contents is determined by the extension, it is possible for attributes, and sub-elements to overide each other within a construct, but not accross constructs. >> The meaning of the namespaced-qualified XML element used as the >> property name SHOULD be defined by an Atom extension specification, >> or by a compatible vocabulary. > I'm not sure I have any idea what this means. I haven't seen any discussion about whether there will be any registration procedure for extensions, so I haven't elaborated much on this section. > Does it mean that extensions should have an Atom-specific document > describing them? Simple Extension constructs may be defined for more general use than just within Atom - the PRISM vocabulary for example. Extensions designed specifically for Atom need to be specified. I didn't actually use the word "document", but documenting them would be a good idea. > Is there more to it than that? No, not really. It basically just means: "somebody needs to define an extension before it can be used". Were you having trouble with the "Atom extension specification" bit in just 9.1.1, or for 9.1.2 aswell? Would it help if I made another bulleted list for what should be defined for 9.1.1? -- Dave
