Reification?

[Claude Warren <cl...@xenei.com>]

What is the new thinking on Reification?  I see that the Reifier has
been removed from Graph.  But I suspect that there is still
reification in the model.  How do we cause reification to occur and
how do we remove reified statements?

-- 
I like: Like Like - The likeliest place on the web
Identity: https://www.identify.nu/user.php?claude@xenei.com
LinkedIn: http://www.linkedin.com/in/claudewarren

Re: Reification?

[Andy Seaborne <an...@apache.org>]

On 26/12/12 09:33, Claude Warren wrote:
> What is the new thinking on Reification?  I see that the Reifier has
> been removed from Graph.  But I suspect that there is still
> reification in the model.  How do we cause reification to occur and
> how do we remove reified statements?

There is still support for reification in the API in Model - it's only 
for reification style "standard" (which has always been the default ). 
"Minimal" and "convenient"

"Standard" can be done with just code - no additional state management 
is needed in the storage.  The code is in ReifierStd.  It should provide 
the same functionality to Model and the Model-level reification 
operations call ReifierStd functions.

"Minimal" and "convenient" (and IMO the names were always arguable!) 
both need additional state management outside the storage of RDF 
triples.  As we found, this was a lot of the complexity of the old RDB 
database layer, and it had a simple DB layout.  SDB and TDB have only 
ever provided "standard" as pure code on top of storage.

In "Standard", reification occurs if the 4 triples of the reification 
occur in the graph, if there are 3 or 5 (duplicate of a 
subject/predicate/object triple with the same subject), it's not 
considered a reificiation.  The triples are always visible by just 
looking in the graph.

Model.createReifiedStatement generates the 4 triples for the reified 
form of the statement, Removal is deletion of all wellformed groups of 4 
triples which form reifications of the statement.

	Andy