No gumption today (I’m just not a listserv question poster), but in looking through the listserv archives and considering the Fedora documentation some more I am pretty sure that the triplestore that can be set up with Fedora is not something that has to be there for Fedora 4 and Hydra to work together. The triplestore is another way to expose the data, just as a Solr index is a way to expose the data for searching, but Fedora doesn’t use the triplestore. Fedora stores RDF data and takes create, update, and delete commands for objects as SPARQL queries but the messenger piece I talked about yesterday (what’s becoming Apache Camel in this workflow) is what decides if an external triplestore also receives that info (just like it decides if a Solr index is updated based on that command). So the triplestore is just another place where metadata from Fedora can live. And you still have to decide how to handle any descriptive metadata you have in XML with Fedora 4. That’s part of what I am writing about in the article. This toolset has a definite set of choices and they make a difference. I’ll be interested to see if my use cases show these choices are more generalizable beyond the specific Fedora/Hydra software.
I’m not completely certain about the answer to my other question regarding triplestores being able to manage complex hierarchical triples but I don’t know if I actually need to figure that out for this article at this point.
I read more articles today considering the conceptual difference between XML and RDF and I think it helped but I have to stop reading now. I wrote just a little bit more this afternoon and I want to spend all day writing tomorrow. I’m to the point where I need to read what I have in my article from start to finish and compare what I’m trying to say with what it is I’m actually saying. I really want this article to be clear and digestible. It will make it much more useful to me and hopefully help others as well.
And I win for most boring blog post ever! Here is a winking kitten from Flickr: