On 5/4/14 6:44 PM, Sarven Capadisli wrote:
On 2014-04-25 18:44, Kingsley Idehen wrote:All, Circa. 2014, we shouldn't be depending on any piece of platform specific technology to create a graphical representation of datasets across the entire LOD cloud, or specific sub segments of said cloud. We should be able to generate graphical representations that include the following fundamental features: 1. graph visualization 2. live hyperlinks of URIs that denote the cloud data sets. Why? You end up with yet another beachhead for Linked Data follow-your-nose pattern. Whenever we produce output with non existent or blurry access to HTTP URIs that denote entities (documents, agents, or other entity types) with inadvertently missed a powerful opportunity to showcase the basic Linked Open Data principles value proposition. Also what applies to this specific scenario also applies to any other resource published on the LOD banner. Shortcut: Find a visualization tool that can process CSV ouptput. Feed it SPARQL URL for a solution where the output format is CSV. Done.I think what you are thinking of is more along the lines of http://lodlive.it/Perhaps void-graph's aim to produce the LOD Cloud diagram is a bit ambitious at this time. I'm not sure if that itself is still a worthwhile exercise any way anymore. There are efforts like http://lod4all.net/ which goes in that direction, but the datasets that are listed feel curated.However, what might be more interesting is creating a more realistic LOD cloud that's perhaps derived from the crawls like LDspider.I'm not convinced on the shortcut idea you propose. Just before it you were talking about tapping on the power of true LOD, then you want to process the information via CSV.
What I saying, for additional clarity, boils down to: 1. querying dataset information using SPARQL 2. returning the results using CSV -- instead of XML.In regards to visualization, the benefit of a CSV based SPARQL solution lies in the fact that the HTTP URIs that denote entities, are simply a part of the payload that don't distract clients that are focused on the data representation format.
End result is that you have common data visualization tools working with 5-Star Linked Data, without knowingly make a single commitment to Linked Data principles or RDF relation semantics.
You end up with a LOD cloud bubble using tools that know zilch about Linked Data or RDF :-)
IMO, lodlive and void-graph are both in the right direction. void-graph in and of itself is a useful tool, at least to create a simple diagram for the 270a.info dataspaces.
It's about simply adding data visualization tools that only consume CSV to the mix, That's it. Basically, we don't depend on one or two tools, we end up having many tools at our disposal for these data visualization endeavors.
-Sarven http://csarven.ca/#i
-- Regards, Kingsley Idehen Founder & CEO OpenLink Software Company Web: http://www.openlinksw.com Personal Weblog: http://www.openlinksw.com/blog/~kidehen Twitter Profile: https://twitter.com/kidehen Google+ Profile: https://plus.google.com/+KingsleyIdehen/about LinkedIn Profile: http://www.linkedin.com/in/kidehen
smime.p7s
Description: S/MIME Cryptographic Signature
