search

cwrc / HuViz

LOD visualization tool for humanities datasets
8 stars 1 forks source link

Subsetting large graphs: include all connections between any two nodes included in a graph #48

Open SusanBrown opened 8 years ago

SusanBrown commented 8 years ago

When constructing subsets, the system should include any edges between any two nodes in the current dataset in order to make the subset more accurate as a relationship of the overall graph.

This could be a switchable option because not all users would necessarily want it, but I think the default should be on, in devising the parameters for a subset.

smurp commented 8 years ago

Interesting point Susan. I'm not sure that there is anything either to switch on or off in this situation though, because if we are creating a subset (ie including triples associated with not just one WRITER but with different WRITERs as subjects) then the only triples which can bridge between these datasets are the ones which would already be in the dataset for the first writer or the dataset for the second writer. As far as I understand we don't yet have ways of generating triples which do not have as the subject one of the WRITERs, so for example we don't yet have a way to directly generate triples between the mother of the Bronte sisters and the father of the Bronte sisters. I hope I understood your meaning correctly!

On Mon, Oct 19, 2015 at 10:15 AM, SusanBrown notifications@github.com wrote:

When constructing subsets, the system should include any edges between any two nodes in the current dataset in order to make the subset more accurate as a relationship of the overall graph.

This could be a switchable option because not all users would necessarily want it, but I think the default should be on, in devising the parameters for a subset.

— Reply to this email directly or view it on GitHub https://github.com/cwrc/HuViz/issues/48.

Shawn Murphy smurp@smurp.com (780) 903-2428 http://www.smurp.com

cmiya commented 7 years ago

Will this be resolved by #37? And should it likewise be deferred until after we've achieved basic functionality for loading from a SPARQL end point?

I'm also wondering if our thinking about this has changed with our move to the web annotation model. Can assertions (e.g. so-and-so - hasReligion - Jewish) have multiple contexts? Or a single context have multiple assertions?