[ShelleyPowers] Define author? Do we mean a person? An organization? Or can an author be multiple people, writing as a single entity with one single level. Before you all decide to use dc:creator and drill down into the use of RDF for this, may I suggest that you define 'author' first?
[JoeGregorio]: How about the definition of Author being any entity that can be assigned a unique e-mail address? (I'm not trying to be smarmy here, but this would cover the case of individuals, organizations, companies, etc.)
-
[ShelleyPowers] Joe, that's not bad. However, there's nothing that says a weblogger has to have an email address. And if we roll comments into this mix, anonymous posters don't have an email address. I'm not trying to drag this into data concept navel gazing, but I see us defining an entity, which could be a syndication excerpt, a comment, or a trackback, as much as it is a weblog entry. And that changes the concept of 'author'.
[JoeGregorio]: Ooops, didn't mean to tightly intertwingle the concepts. How about this for a re-wording.
-
Author: Any entity that could be assigned a unique e-mail address
I didn't mean to say that you had to have an e-mail address, just using the ability to assign an entity an e-mail address as an marker of potential author-ship. So, for example, an anonymous poster or commentor may not have an e-mail address, but being a person, they could have an e-mail address and thus fall into the category of entities that can be authors.
-
[MarkCidade] I'm not sure how solid the e-mail address definition is. An e-mail address for what? Hypothetically, anything can have its own e-mail adress for almost any reason.
[JoeMadia] How about allowing Author to be any one of the following:
-
Explicitly Anonymous
-
Simple Identity (email plus title or something similar)
-
Strong Identity as defined (eventually) in Identity
I'm not trying to add complexity here but it seems that layers of increasing structure/rules would work well in the weblog world since there are such a wide variety of usage scenarios. Some authoring tools will never do anything other than simple identity (comment systems) but it seems likely that some will go to great lengths to ensure reliable identity (journalism, larger content management systems). Providing standard answers at each layer (aka level of service) seems much more feasible than finding a one size fits all answer. Just my two cents... any thoughts?
JoeMadia has a good idea. Treat enhanced author concepts as modules.
[TimBray] Defining the notion of author is a hard and possibly open-ended problem. I suggest we just say authors are identified by URI (note, mailto:tim@example.com is a URI, so is http://www.tbray.org/ongoing). Then let the community evolve best practices within the (very large) URI sandbox.
-
[JonathanSmith RefactorOk] Tim, is there any correspondence between this discussion and the InternetArchitectureBoard, sp.:
-
Requirements for a Distributed Authoring and Versioning Protocol for the World Wide Web (RFC 2291)
-
HTTP Extensions for Distributed Authoring -- WEBDAV (RFC 2518)
[TimothyAppnel] +1 to the suggestion by "the other Tim." and the use of an optional module of contributor for multi-author situations.
[JoeGregorio] I like that idea, and along with AaronSw's observations about journalism I propose:
-
There is always one, and exactly one, Author of an Entry, and that Author is identified by a name and an optional URI.
[ClayShirky] AaronSw misread journalistic practice -- its more varied than he imagines, and in any case journalism is just one pattern that can be supported by weblogs. See my comments at NumberOfAuthorsDiscussion. (In fact, since part of the definition is the 0,1,N question, I think these two pages shoudl be merged under DefinitionOfAuthorDiscussion)
See also Authors, AuthorElement, DefinitionOfAuthorDiscussion