Multiple Selection Metadata Lists and Dates.

When I went to the new metadata controls (Scrivener 3), I immediately wanted to produce a list of topics salient to my book (2nd Law, least action,shortest path, prediction,learning,fitness,information theory, computation theory,topology,arbitrary boundaries, causality,locality,domain independence, latency,emergence) and then I want to select multiple items as befitting the section or chapter to which I’m attaching that metadata.

Same for characters list (Thelma, Louise, Jimmy, J.D., Hal) or in my case, (Einstein, Darwin, Bohr, Boltzmann, Shannon, Susskind, Carol, Aaronson, Schrödinger, Carnot, Turing, Gödel, Dawkins, Hawking, Brooks, Whiley)… obviously i want to be able to attach multiple characters as metadata to any particular document. Once a user has set up metadata, Scrivener could use that data for text hinting (completing words and phrases while the user types), and to perform automatic semantic analysis of the user’s manuscript (for presentation to the user as timelines, plot lines, character frequency graphs across the arch of the manuscript, maps with travel lines, topic graphs, etc.), and to aid the user in whole manuscript and document and document collection searches (where in draft are Thelma and Louise not present together? Where in draft are the topic’s “thermodynamics” and “computation theory” present within the 200 word spans? Where in Manuscript is Billy Bibitt not on the psych ward? Graph the occurrence of Billy Bibbit and Randle McMurphy across the manuscript. Show map of all locations within project “One Flew Over Cuckoo’s Nest”. One can imagine many intuitive ways in which this metadata could be used graphically to produce visual search results. One can also imagine many ways in which semantic analysis (simple word occurrence statistics could be used to auto-populate meta-data lists for standard who what when where why records ).

Same for dates! There are often a whole bunch of relevant dates associated with any one document in my draft. If the dates metadata is to be useful, it needs to allow attaching several dates to one metadata item.

Same with places or locations. And salient to any author’s writing process is the key “events” that transpire in a book. Events are a special case where characters, places, and dates are linked. Would be great if such data was definable as metadata. As school kids, we all learned that stories are about who/what/where/when/why descriptions . It would seem reasonable to expect that a writing environment would know of and allow the writer to build formal definitions of such things. Once project specific characters, things, places, dates, and motives) were defined by the writer, Scrivener could allow the formal combination of these parts into events (Event: Characters-Thelma, Louise, J.D./Place-Roadside Motel/Date-May 10,1996) and plot arcs (Plot Arc: Characters-Thelma, J.D.-Fall in Love). Once combined, these events and plot arcs could populate metadata lists.

Sure, one can do some of this with keywords but a simple misspelling or alias or synonym will really screw up the works.

In most applications that make use of this affordance, there is the user option to restrict selection to one item (radio button), or to allow many, or to restrict the number of allowable selections to a user indicated number.

Also, and correct me if I am wrong, but it doesn’t seem that metadata applies to text selections within a document? Does it? It should if it doesn’t.

Thank you, Randall

Depending on how you set your document structure up, it is entirely possible to have multiple Scrivener documents compile together into a single section within your final document – for example, each paragraph being a separate Scrivener document at the same level, with the containing folder being the scene.

By allowing (but not requiring) you to break things down by this level of granularity, you can achieve the same result – specific metadata applying to portions of the text, not others. Metadata is always tied to the document level, though, because (as I understand it) there is no easy or reliable mechanism for tying the metadata to the content except by Scrivener associating it with the corresponding raw RTF document. If I had multiple paragraphs within a single Scrivener document, I would have a single corresponding raw RTF document in my Scrivener project, and there’s no good way to break out the different sets of metadata to reference cleanly back to the paragraphs within that RTF document.

In v3 for Mac, you can use the Outliner mode to show the relevant metadata columns you are looking for and then export that view to a CSV file. Once there, you can import it into whatever analysis tool you most prefer. I do not know if the v3 beta for Windows has this functionality yet, but my understanding is that it is expected to. You can then set up multiple types of custom meta-data and combine them as needed in your analysis tool.

Thank you, Will give it a go.