This week, you might choose to comment on how suitable Drupal might be for your collection. Begin to develop some criteria you would use to judge how well an application such as Drupal meets the needs of your collection and its users. We will expand on this problem over the semester.
We have been reading about the need for humanities scholars to be able to use a digital collection with a degree of confidence about the nature and authority of the relations between objects, yet having the structure of those relations clear so that the information added is objective rather than subjective. What I would really like to make is a database or collection or semantic web of all the texts (with attached full-text) that George Eliot read or interacted with, with some degree of confidence added in about how influential those texts were. One could argue that there is a sort of taxonomy to how much she interacted with a text, in ascending order from hearing it read aloud, to reading it in translation, to reading it herself in the original language, to reviewing it, to editing it, to translating it from another language into English. These are all types of relations with a text. One can also argue that reading it more than once, or attesting to its influence in letters or in research notebooks, is also a measure of influence. I was reading about RDF, and that seems exactly the sort of inferential structure I want to be able to capture, starting with the simplest: What she read, with some sort of statement about her relation to the text, and a documentary page showing the authority for that relation. One can infer the direction of influence between texts according to who read what and when.
Because eventually I would want this to be part of a larger database of “Literary intelocutors,” I’m having trouble figuring out if the key entity in this collection is texts or a person. The way I envision the normalized tables in a database would be a table of persons, a table of texts, and a table of links between the two, in the form of “GE read Rousseau’s Les Confessions, in French, in 1834, according to these authorities, and here is a link to that edition of Les Confessions in French (or perhaps a digital image), plus a searchable English translation.” I have been thinking that I needed to include all the standard metadata for each text in each entry, but that seems a waste of space. The new and useful information to be collected is the table of links, so all I really need to capture is what I have underlined; Each underlined phrase is a field in my collection.
Any content management system I use for my collection will need to be able to search and manage large attached text files in a variety of formats, to query the collection of these files with a full-text search, and have a faceted search that narrows the query results by type of relation, by subject, by language, by year, or type of text file. I also want to be able to widen the search if necessary, though, across subjects, dates, etc. The idea is to be able to use this collection to specify a group of texts to search, and to be able to document the relationships and direction of influence between them. I would love to be able to actually graph the connected nodes in some sort of network display and to assess the degree of influence.