Re:Show HN: Stash, a graph-based cache for Node.js and Redis
by nkohari on Sunday March 04, 2012 @09:31AM
(#3662819)
This was actually my original plan, but I took the easy way out for now. :)
We use MongoDB as our data store, and between the UUIDs used as primary document keys and arrays of UUIDs representing links between documents, we could (reasonably) easily determine the dependencies automatically.
Perhaps instead of explicitly defining dependencies, there could be some way of modeling your cached data as some kind of set of objects which would then understand their own graph. As you can see I haven't quite figured out how it would work but this is how I've been thinking about it lately. Like a simple ORM, in some way.
Something like the following terrible psuedocode: