SIOC/OntologyIssues - SW Cluster Wiki

This page contains issues related to to the [http://www.rdfs.org/sioc/ SIOC Ontology]. Including proposed additions / deletions / updates to the ontology.

Divide User->sioc:name into 2 parts

To reflect the fact that many systems keep first name(s) and surname separate, let's also have separate properties. It is easier to join them when displaying that extracting from a sioc:name property.
Looking at [http://rdfweb.org/topic/NamesInFoaf NamesInFoaf] (RDFWeb), there is no agreement on this and actually the simplest proposal is to keep 'name' property and add more elaborate way to describe sequence of names.
Proposal: keep sioc:name and introduce 2 new properties - for first name / last name OR for given name / surname.BR Which of those 2 to choose?

Need to Identify Individuals

There is a need to identify individuals, to link together their submissions from different community sites. Proposal: add sioc:mbox and sioc:mbox_sha1sum as IFP (inverse functional properties) for sioc:User. There would nicely map to same properties for foaf:Person.
Also IFPs in FOAF: aimChatID, homepage, icqChatID, jabberID, msnChatID, weblog, yahooChatIDBR We can also add skypeChatID (but how long will we continue a list and add new IDs when new services appear?)
Options: Andreas mentioned ways to identify individuals other than IFPs.BR What are they?

Groups of Individuals vs. Types of Individuals

Should we represent [special] groups of authors as:BR

   a) separate instances of sioc:Administrator (subclass of sioc:User) w/o further grouping?[[BR]]
   -OR-
   b) [special] subclasses of sioc:Group like sioc:Administrators, ...

The latter would allow us to list all the administrators and we can keep the existing subclasses of sioc:User and be able to infer (?x :member_of :Administrators) => (?x rdf:type sioc:Administrator). Proposed additions: for each subclass of sioc:User add a relevant subclass of sioc:Group

Roles of some subclasses sioc:Person

At least one of subclasses of sioc:Person - sioc:Moderator has a badly defined semantics. E.g., if we know he is a sioc:Moderator, what do we know about him. We may deduce that there exists a forum :f for which this person :x is a moderator. But that is not a site-wide status - he might be a moderator on one forum, but just a regular user on other. Proposal: clarify this issue.

Name of property sioc:created is ambiguous

This property contains "When this was created, in ISO 8601 format".BR Yet, the property name does not indicate it is a date/time. Proposal: change the name to 'date_created' or similar.

WordPress SIOC export

This is the summary of assumptions used in creating WP SIOC export.

User information: {{{# each sioc:User gets a rdf:nodeID="u-3" assigned to it, where 3 is user id in the system # sioc:name contains user's prefered displayed name as determined by blog settings # sioc:description = sioc:firstName . ' ' . sioc:surname # properties sioc:firstName and sioc:surname are introduced # properties sioc:mbox and sioc:mbox_sha1sum are introduced}}}

Forum information: {{{# a blog has one sioc:Forum with rdf:nodeID="f-1" # forum has a sioc:name = "Main blog at $weblog_name" # @@issue - we need to add a property for forum (site ?) url # @@issue - also information about RSS feeds. (but do we need to point to RSS feeds if we export the same info in SIOC? what about the apps that do not grok SIOC?)}}}

Post information: {{{# property sioc:created is named ambiguously # @@issue - we need a property to indicate post title - proposals are 'title' or 'name # @@issue - also might need properties for the url and the rss feed # assuming that we indicate sioc:has_creator property with rdf:nameID="u-3" (where 3 is the user's ID), do we also link post to its creator via sioc:mbox or mbox_sha1sum properties? if so, then how do these co-exist? # RSS 1.0 has property description for a plain text contents (possibly abbreviated) and content:encoded for full (HTML) content. SIOC currently has only sioc:content. @@issue - desicion needed what type of content do we put in sioc:content and if we need 2 separate properties - one for plain-text and the other for rich-text.}}}

Groups and sioc:knows relation:

John: do we have sioc:knows relation for Groups? (similar to foaf:knows for persons) Question: what is the purpose of sioc:knows between sioc:Groups?

New/missing properties: {{{# paper describes sioc:closed property, but it is in the ontology.

# sioc:knows is in the paper, but not in the ontology }}}

Event information: {{{# event was changed from the class to a property of Post. start and end dates also added. # but 'event' propery name does not clearly define what is a sioc:event of sioc:Post. # from modelling point of view event is best displayed as a subclass of sioc:Post, possibly in a separate Events module.

# solution: remove event, starts_at, finishes_at and go through it again}}}

== This is may be interesting ==