Tags in online archives

I’m revisiting the Archive of Our Own (http://archiveofourown.org/) project because I wanted to talk a little bit about tags, and it’s an example of using them that illustrates a larger point. The Archive is still in closed beta, and does not (yet, as far as I can determine) have any FAQs available, so my comments are based strictly on what I can observe from poking around in the site; I know nothing about what is official policy.

From what I can tell, authors upload their stories and input various bits of information about them. A set of symbols identifies some characteristics of the stories: rating, type of relationship, whether there are content warnings, and whether the story is finished. Other information is given in text form, including title, author, fandom, characters, and specific warnings.

The specific warnings (convention in fandom calls for informing readers about certain elements in the story, such as sexual content, violence, death of major characters, etc.) are given in the form of tags, and the tags can be used for conveying other information as well. There is a page (http://archiveofourown.org/en/tags) where all tags that have been used are listed, in a cloud format that shows in larger type those tags more often used. The very most commonly used tags are not only in large type, but in red rather than black font as well; the three most often used are “Angst,” “First Time,” and “Humor.”

It seems clear that the tags are not taken from an established list (or not necessarily), but can be added by individual authors as desired. I deduce this from the variety of formats and meanings of the tags themselves, as shown by copying one line of tags at random:

Catholic | cats | Challenge fic | challenge response | chan | change | Chanukah | character driven | Character Study | character:realisation | Chaya | cheesecake | childhood | chinese take-out

This is a very Web 2.0 approach – the users create the content, and also identify it in ways that they choose, although there are also some standardized ways as well (title, fandom, etc.). It illustrates both the strengths and the weaknesses of such an approach, though. It gives the authors agency and ownership, which is very much in line with the purpose of the site. On the other hand, the total number of tags at this moment is 1085, a number that doubtless increases daily. Someone using the site might have a hard time thinking of what tag might identify the type of story they were hoping to read, especially given that different authors might use different tags to mean essentially the same thing, e.g. “smut” vs. “porn”.

So I wonder, are these tags really useful? That raises larger questions about metadata generally, who should create and maintain metadata, and how. Something worth considering even though I’m not sure there’s an ideal answer, certainly not in this specific case and perhaps not at all.

Choosing metadata schema for digital collections

Kennedy, Marie R. 2008. Nine questions to guide you in choosing a metadata schema. Journal of Digital Information 9 (26). http://journals.tdl.org/jodi/article/view/226/205 (accessed October 26, 2008).

Kennedy assumes that the collection is already in development, with a known rationale for its creation, and that any copyright issues have been considered. She then lays out the following questions as guidelines for choosing a metadata scheme that will be useful for the collection’s users:

1. Who will use the collection?
2. Who will catalog the collection?
3. How much time and money is available?
4. How will the collection be accessed?
5. How does the collection relate to other collections?
6. What is the collection’s scope?
7. Will the metadata be harvested?
8. Will the collection “work with” other collections?
9. How much maintenance and quality control is desired?

Kennedy doesn’t so much promote specific extant metadata schemes (like Dublin Core or others), as she instead encourages developers to pay attention to the needs inherent in their own situation. This is a very useful checklist for someone developing any collection, but particularly for someone working on academic or institutional collections and repositories.