- From: Marko A. Rodriguez <marko@lanl.gov>
- Date: Tue, 17 Mar 2009 15:18:15 -0600 (MDT)
- To: "Frederick Giasson" <fred@fgiasson.com>
- Cc: public-lod@w3.org
Hi Frederick, Thank you for your thorough response. Yes--it looks like an LOD voID instance is the thing to have. I originally created my csv of the LOD cloud by painstakingly going from the OmniGraffle PNG to a csv file by hand (see a visualization of that csv file in [1]). If I had the original OmniGraffle source, I could ensure accuracy in this manual method by removing links in the Ommnigraffle representation as I created them in the csv representation. If someone could provide me that Omnigraffle source I could create a clean LOD csv file and provide the group with it for their interests much like Frederick did with the UMBEL work. [1] http://markorodriguez.com/Blarko/Entries/2009/2/28_The_Linked_Data_Cloud_is_Growing_files/lod-graph.pdf Thanks so much, Marko A. Rodriguez http://markorodriguez.com > Hi Marko! > >> I'm finishing up the final version/draft of the LOD March 2004 analysis >> (see the first draft at [1]). I was wondering if there is a easily >> parseable format for getting the graph structure of the Linked Data >> cloud. >> That is, something analogous to how [2] provides [3]. Hopefully >> Omnigraffle isn't the "raw" data. :) >> >> [1] http://arxiv.org/abs/0903.0194 >> [2] http://www.umbel.org/lod_constellation.html >> [3] >> http://www.umbel.org/lod_constellation/class_level_lod_constellation.csv >> > > I think the LOD graph has only been created and maintained by hands so > far. So I cannot really talk about it. > > However, I can talk about UMBEL and how we created the graph. Basically, > we have a script that query a triple store where all ontologies > descriptions are indexed. From these queries, we built-up the CSV file > you mentioned above. We use RDFS & OWL properties used to describe these > ontologies to build-up the graph of relationship between all the classes > of each ontology that is linked to UMBEL. > > Once we generated the CSV file of relationships between classes linked > to UMBEL, we import it in Cytoscape to create the actual graph that you > also mentioned above. > > So, what is needed in order to to the same for the LOD cloud is a > mixture of things. First, people have to define all the criterias that > lead to a link between two datasets (for example, is only a link, with > any properties, between two resources of two datasets enough to create a > link between the two dataset in that cloud? (I doubt), Etc.). Once you > have these criterias, you have to query each dataset, given these > criterias, to build-up the linkage between each dataset bubble. > > I think Void should be used to describe these datasets, and to describe > the linkage; however I am not sure if the linkage criterias are in > place. Having the Void representation of the linkage of all dataset in > the LOD cloud would be helpful to make the query against that data > instead of each dataset to create the actual CSV file to import in > Cytoscape. > > Hope this is helpful, > > > Take care, > > > Fred >
Received on Tuesday, 17 March 2009 21:19:40 UTC