* Eric Scheid <[EMAIL PROTECTED]> [2005-05-20 20:10]: > On 21/5/05 3:41 AM, "A. Pagaltzis" <[EMAIL PROTECTED]> wrote: > > However, it does pose a problem of default semantics. Do we > > define particular categories in the spec? If we donÂt, do we > > define a default category to be assumed in absence of > > explicit category elements in the document? If we do, do we > > define such a default? > > The simplest thing that could possibly work is to say that if > there is no <category> element inside the <contributor> then > assume a default of @term="author" with unspecified @scheme and > unspecified @label. > > Covers 99% of use cases, I should think. > > No need to explain what the string "author" means, surely?
Sounds fine; but you did not directly address the question of whether we define any default semantics. The absence of atom:category and <category term="author" /> mean the same thing per your proposal. You did effectively specify a term âauthorâ with particular semantics, if only implicitly. My question is: do we define even as much? Background: we could say something like âThe given contributor is to be assumed to be an author of the entry in absence of an atom:category stating otherwise.â which avoids defining any terms at all, even implicitly. To get to the point: if we do define one term, do we define more of them as well? Such as âeditor,â âcorrespondentâ or whatever? This is the only reason Iâm at all wary of the proposition. The infrastructure it supplies is sound and very elegant, but the infrastructure per se is hollow scaffolding without the semantics it is supposed to carry, and I feel really uncomfortable about the idea of getting into that semantics game. Particular at this so very late stage. If we can find an approach that avoids getting into that can of worms, Iâm definitely in support of the idea. If we cannot stay away from it, then allowing multiple atom:author elements and leaving any additional complexities to extension elements would be the simpler thing to do. Regards, -- Aristotle