I’ve been using Scrivener for a week or so now, and I think you’ve done a fantastic job.
There is one feature I’d like to see added at some point: the addition of a “type” column on the references field. It wouldn’t be a terribly useful feature for Scrivener as it is now; but it would open up a lot of possibilities when/if you add scripting support. For instance:
References between alternative story elements could have types like “conflicts with” and “depends on”. Then a script could recursively trace such references from a selected element, to find the global effect on the rest of the story of using or removing that element.
References between events could have “cause” and “effect” types. A script could then check all the events in a story to make sure they’re in a consistent order.
Characters could reference each other as “mother”, “father”, etc. A script could then use these links to assemble family trees for the characters.
Basically, it would allow scripters to add relational database-like functionality to Scrivener projects without needing particular fields to be hard-coded into the app.
I was thinking of something along these lines earlier today, except I was thinking of expressing type by grouping methods, similar to the Keywords HUD, but restricted to only one level deep, and thus not requiring indentation. Basically the ability to add a “Header” or named divider, beneath which you could sort your references. My thoughts on why this would be better than a column is that it is already pretty cramped over there. Generally you do not need to see the URI, so that column can be minimised, but even so at the default Inspector width, the title column is a bit slim, and third column would make the whole thing nearly unreadable without making the Inspector quite wide.
Though not as powerful in terms of scripting etc., the ability have internal links in the note field would satisfy this need for me since you could annotate a link by type, relation or anything else you liked.
