Schema Information and Controlled Vocabulary - jdwerst91/world-of-witcher GitHub Wiki
The World of Witcher collection presents some interesting problems and usage scenarios for metadata. It aims to collect both real world and in-universe metadata. This means that for an object like a sword that appears in the games, it would have real world metadata of being created by a designer at CD Projekt Red, or first described by Andrzej Sapkowski, in addition to having metadata for who forged it in the Witcher universe and what sort of physical attributes it has therein. Some objects, like the physical books that the universe takes place in, are real world items and have primarily real world metadata, but we included in universe information on fictional time period they cover. To further complicate matters, the collection has a wide range of both real world and in-universe items; some of these items were better suited to particular metadata schema and were a poor fit for others. In addition, the Witcher universe contains some items that share names, such as Last Wish. Not only is it the first English translated story in the series, and a short story within that novel, but it is also a quest within one of the Witcher videogames, and also a story and song existing within the universe.
With all those things in mind, we came up with a metadata schema that utilizes Dublin Core, MODS, VRA, and a few of our own elements. They are split up into “real world” and “in-universe” selections. We found that this structure allows us to use schema that fits certain parts of each object. Taking a look at the sword example again, we treat the “in-universe” metadata as you would a real sword; it has a creator (blacksmith), physical dimensions, date created, and background information that fit well with VRA. When we look at the “real world” information, however, we found the information of who put this object into the game, or who wrote it into the book, was better served with Dublin Core or MODS elements. We utilized aspects of all three schemas depending on the item itself and whether it captured real world or in-universe information. For more information on how we used those elements from their respective schema, please refer to our application profile.
In addition to Dublin Core, MODS and VRA, we had a need for controlled vocabulary to provide more organization to the collection. Each metadata record is set up with a general section, a real world, and an in-universe selection. The general section contains each object’s unique ID, a MODS title field, a Boolean check, and controlled vocabulary for category. We choose to use MODS for title as it allows for the most options in terms of title and alternate title, which is important in dealing with a collection that includes multiple languages both real and fictional. Having the title outside of “real world” or “in-universe” also avoids any confusion from having multiple title fields appear in the record for a single object. The Boolean check asks if an object physically appears in the real world or only in the Witcher universe. This helps to distinguish multiple objects with the same name and aids in our functional requirements.
Lastly, we utilize controlled vocabulary to place each object in the collection into one of eight categories: Media (trailers, video games, songs, movies, etc.), Art (paintings, sculptures, etc.), Printed Materials (books, maps, etc.), Gear (swords, armor, clothing, etc.), Collectibles (plants, trophies and all other non-gear items), Beasts (animals, monsters, all non-human creatures), People (characters, NPCs, etc.), and Places (geographic and physical locations). These categories are broad enough to allow every object to in the collection to fall into one, yet narrow enough to avoid confusion between each other. The categories help alleviate the problem of multiple objects having the same name as mentioned before, and serve an important role in the functional requirements of the collection.
Word version of this document: Schema Information and Controlled Vocabulary
