On 3/7/12 12:29 PM, David Wood wrote:
Hi Kingsley,On Mar 7, 2012, at 12:09, Kingsley Idehen wrote:On 3/7/12 12:06 PM, David Wood wrote:Callimachus requires a single local RDF store for its own data storage and it connects to that local store via various Sesame SAILs. Our work to connect to AllegroGraph will probably result in more detailed implementation documentation for third parties.Since you connect with Sesame SAILs, and a native provider [1] exists for Virtuoso's graph store, you should be set :-)Perhaps. We make heavy use of an auditing SAIL, which most people haven't implemented. We also need some SPARQL extensions that we use for keyword searches (in lieu of a separate full-text search engine). There are some other bits and pieces. More to the point, though, we don't gain much (if anything) from trying to support every RDF store.
You gain a lot if you exploit a data access abstraction that unleashes the full power of URIs. In the RDBMS world you have ODBC and JDBC providing a generic layer for data access. Of course, as per my posts today, Linked Data (thanks to power of URI abstraction and HTTP based data access) runs rings around ODBC, JDBC, ADO.NET, JDO etc..
You don't need to support every RDF store. That's akin to RDBMS client application developers saying they don't gain much from ODBC or JDBC access en route to database independence where they leverage a data access abstraction layer.
We haven't pursued Mulgara for example, in spite of my personal history with the project and its predecessors. Instead, we are commercially driven to support those stores that help us make income. That's how we ended up with OWLIM and AllegroGraph.
See my comments above.
If the presence of Callimachus helps Virtuoso or vice versa, then we'll do it right away. If not, you are welcome to and we will help, but it is not an immediate priority for us.
I was simply trying to understand what you meant by "agnostic" or RDF store independent etc..
Linked Data is about upping the ante re. Open Data Access, as per my comments above.
Links:1. http://goo.gl/LmFR5 -- getting data directly into Google Spreadsheets via SPARQL protocol URLs (in the RDBMS realm, ODBC moving data into Excel en route dynamic pivot tables was the ultimate value prop. demo)
2. http://goo.gl/ovqPL -- longer post about the journey from Open Database Connectivity to Open Data Access.
We regard the Open Source version of Callimachus to be a community project. Please feel free to join in! Regards, Dave
-- Regards, Kingsley Idehen Founder& CEO OpenLink Software Company Web: http://www.openlinksw.com Personal Weblog: http://www.openlinksw.com/blog/~kidehen Twitter/Identi.ca handle: @kidehen Google+ Profile: https://plus.google.com/112399767740508618350/about LinkedIn Profile: http://www.linkedin.com/in/kidehen
smime.p7s
Description: S/MIME Cryptographic Signature
