On 5/18/05, Graham <[EMAIL PROTECTED]> wrote: > On 18 May 2005, at 9:36 pm, Robert Sayre wrote: > > > atom:entry elements are advised to contain ... a non-empty > > atom:summary element when the entry > > contains no atom:content element > > I'd like us to advise including an atom:summary when atom:content is > binary (or for that matter, any non-text/html/xhtml type)
OK. We can do that by accounting for atom:content elements with media types in @type. ----- Some applications may choose to require a minimum amount of inline text or (X)HTML data to function reliably and predictably. For that reason, atom:entry elements are advised to contain a non-empty atom:title element, a non-empty atom:content element when that element is present, and a non-empty atom:summary element when the entry contains no atom:content element or the atom:content element's type is a MIME media type. The atom:summary element can be omitted if the Atom entry is generated for an application with known requirements that make the inclusion of an atom:summary element impractical (such as limited bandwidth). However, when some general-purpose applications encounter such entries, those entries might be ignored. Regardless of how an associated application will handle entries with no atom:summary element, all Atom Processors MUST NOT fail to process Atom entries simply because they contain no atom:summary or atom:content element. ----- > > > Regardless of how an associated application will handle entries with > > no atom:summary element, all Atom Processors MUST NOT fail to process > > Atom entries simply because they contain no atom:summary or > > atom:content element. > > Do we have a definition of "process" (and/or "fail")? Well, this text assumes something similar to this paragraph in PaceDefineAtomProcessor: This specification describes conformance in terms of two artifacts; Atom Feed Documents and Atom Entry Documents. Additionally, it places some requirements on Atom Processors. An Atom Processor is a software module used to read Atom Documents and provide access to their content and structure. It is assumed that an Atom Processor is doing its work on behalf of another module, called the application. If we had text like that, it would be reasonable to assume that 'process' means 'provide access to their content and structure'. It also be reasonable to assume that 'fail' means 'doesn't provide access to their content and structure'. Obviously, Tim is not a fan of that proposal. I think we should steal from the best. Robert Sayre
