Hi Yaron and others
I've long thought an essential grounding semantic model is based on the concept of a 'story' what with a resource itself only telling half the story about something (ie open wolrd assumption); stories are a good fit with the wiki model as each page is a 'topic' each tells a story. The challenge has been to extract semantic properties that reflect the whats,who,whys,whens,wheres,etc that  the story is or is about.

But I think an equally good grounding model is based on the concept of a 'game' -- many things, even wikis, can be characterized this way to good effect. It models things like teams, players, referees, moves and most particularly, rules. The challenge then is to apply semantics during applications of rules and strategies in response to one or more player moves. Game theory in vitro, ya know.

It would be interesting if a semantic wiki were structured to operate games in general (plus of course its own 'wiki-game'). Over time patterns of games (and of strategies) would surface -- selected by an admin for a wiki's operation, or by users wishing to create or initiate some knid of 'game' (eg shopping) of their own or to join a game in progress. A 'service orchestration' might be involved, bounded by rules. A 'game' wiki does things like create virtual users who participate (or referee or act as agents) during life of the game. It could establish new rules on the fly, maintain scoring, notify participants, manage the audience, collect tickets and so on. It could radically shorten the time to field customized games by organizations and individuals.

'Games' I hope could also bring about intrawikis, that is wikis within a wiki, each game with an ACL. Doing this sort of thing I'd expect to make smw more than just the database application it sometimes appears to be, one whose mission is vulnerable to other semantic database applications I can think of. At some point, SemanticMediaWiki should put semantics INTO mediawiki, not just extract semantic properties from mediawiki pages. We seem prevented though by a common distaste to modify mediawiki core and hence, paralyzed.

On 7/9/2014 7:09 PM, Yaron Koren wrote:
Hi Peter,

On Wed, Jul 9, 2014 at 12:56 PM, Peter Brooks <peter.brooks@kchclinics.com> wrote:
I'm trying to think about this in a more semantic way.

Surely, part of the point of semantics is to allow rules to be governed by them.

If I could create a property, say '[[page_visibility::group_X]]' then
it would make sense for the wiki to understand, from that, that the
rule for this property would be found on the property page
'page_visibility' and it could apply those rules to group_X. Does that
make sense?

I disagree with this - I've stated this opinion before on these mailing lists, but it's important to differentiate between data that applies to the real world, and data that applies to the wiki. The fact that "Ulysses" was written by James Joyce is real-world data; the fact that the page on the wiki about "Ulysses" should only be editable or viewable by some people is wiki-specific data. In my opinion, SMW only has an obligation to handle the former: in some cases, it does handle the latter (like all of the special properties relating to page creation), but I view that strictly as a bonus, not an integral part of the software.

I don't mean to derail this thread - other people may have more of an opinion on the technical specifics - but I just wanted to note that.