- From: Wing C Yung <wingyung@us.ibm.com>
- Date: Fri, 1 Dec 2006 08:31:21 -0500
- To: Steve Cassidy <Steve.Cassidy@mq.edu.au>
- Cc: semantic-web@w3.org
Hi Steve, Thanks very much for your interest. We think that being able to revert to a particular version of a named graph is a very powerful idea. It gives you the ability to keep an audit trail to see what changes from version to version. We also have all of the data necessary to be able to reconstruct the state of the store at any point in the past, but that has not been exposed through an API yet. Revision tracking also makes client replication easier from a development viewpoint. The requirement originally came from some work we have done with the life sciences community. One very specific application of this comes from our work with CViT [1] (For more information, see the presentations page [2]). We are building a research tool to manage digital information; in particular, everything revolving around in silico experiments. We want to store everything: workflow for running computational simultations and rendering images based on the output data, linked to the input values (and the literature references that they come from), output data and images at every stage of the process, and code for a particular implementation of an algorithm. As things change (algorithm source code, input values, etc), new versions are generated. So when a scientist goes off to a conference several months after the paper is published (and grad students have been hacking away madly, trying new things, replacing old chunks of code), and a colleague asks how a particular data set was generated, everything necessary to re-run the simulation will be contained in an old version of the graph or graphs representing the whole process. While this example is in the life sciences domain, it's pretty easy to apply it to others as well. [1] http://www.cvit.org/ [2] http://www.cvit.org/node/69 Wing Yung Internet Technology wingyung@us.ibm.com 617.693.3763 Steve Cassidy <Steve.Cassidy@mq .edu.au> To Wing C Yung/Cambridge/IBM@IBMUS 11/30/2006 10:43 cc PM semantic-web@w3.org Subject Re: [ANNOUNCE] IBM Semantic Layered Research Platform Release Wing C Yung wrote: > Our group at IBM Cambridge has open-sourced our Semantic Web projects. > Check it out: IBM Semantic Layered Research Platform [1] (with > documentation [2] and downloads [3]). Lee Feigenbaum has written about our > motivations [4] and roadmap [5]. > > Boca, our enterprise-ready RDF store, is the only component that has been > officially released; others are still works in development. Matthew Roy's > post [6] covers its most important features, including named graph support, > client replication, security, revision history, and real-time notification. > I wonder if you have any more information on your thoughts on revision histories for RDF triple stores? I find some notes on how your implementation works on the wiki but I'm interested in what you think the motivations and potential applications are for this. Any pointers welcomed. Steve Cassidy
Received on Friday, 1 December 2006 13:31:41 UTC